Artificial intelligence atmospheric water generating system control

Abstract

To provide use of artificial intelligence for controlling and improving production of an atmospheric water generating system.SOLUTION: Since an amount of water output by an atmospheric water generating system depends on energy consumption and environmental conditions at a location of the system. To optimize performance of the atmospheric water generating system, the system includes a controller and a control system configured to initiate one or more atmospheric water generating operations based on time-based energy predictions for the system. The time-based energy predictions are generated using a machine learning model that learns correlations between optimization inputs aggregated from a plurality of different sources and performance of the atmospheric water generating system.SELECTED DRAWING: None

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This patent application is related to U.S. patent application Ser. No. 17 / 552,173, filed December 15, 2021, which claims priority from Provisional Application No. 63 / 126,860, filed December 17, 2020, which in turn is related to U.S. patent application Ser. No. 16 / 782,808, filed February 5, 2020, which is a continuation of U.S. patent application Ser. No. 15 / 850,870, filed December 21, 2017, which claims priority from Provisional Application No. 62 / 437,471, filed December 21, 2016, Provisional Application No. 62 / 459,462, filed February 15, 2017, and Provisional Application No. 62 / 459,478, filed February 15, 2017, all of which are incorporated herein by reference in their entireties.

[0002] Additionally, this patent application is related to U.S. Patent Application No. 18 / 179,750, filed March 7, 2023, U.S. Patent Application No. 18 / 199,427, filed May 19, 2023, and U.S. Patent Application No. 18 / 538,832, filed December 13, 2023, all of which are incorporated by reference in their entirety herein. [Background technology]

[0003] The amount of freshwater available for human consumption, plant irrigation, livestock and flock nourishment, commercial and / or industrial use, and other purposes generally exceeds the amount of freshwater required for such purposes. Maintaining sufficient quantities of water for human and / or animal consumption and use has become increasingly expensive in recent years, particularly in arid climates characterized by minimal annual precipitation and no access to other freshwater sources. Processes such as desalination, water filtration and / or purification, groundwater (e.g., aquifer) exploitation, and other processes are often used in combination to supply freshwater to various geographic regions, depending on the relative availability and cost of each water-sourcing process.

[0004] Water shortages in certain geographic regions are also at least partly responsible for food shortages in certain areas of the world. When water is not readily available for irrigating crops and hydrating livestock, basic nutritious foods may be difficult to grow and may be difficult or expensive to obtain on the open market.

[0005] Atmospheric water generation (AWG) systems can help expand freshwater availability, particularly in arid geographic areas and / or areas without access to standing or groundwater, or where water sources are contaminated. However, such systems require significant energy resources to operate, and the ratio of water produced to energy required varies depending on several factors (e.g., time of day, environmental conditions, and / or the like). Conventional AWG systems are generally controlled through the use of programmable logic controllers, microcontrollers, or microprocessors, which may take inputs from local sensors and manual buttons on the system. However, these inputs traditionally consider only a subset of parameters that affect the system's energy consumption and may not account for real-time or anticipated changes in the environment, such as ambient / forecasted weather and / or the like, which can significantly affect system use. This can lead to processing inefficiencies that reduce the effectiveness of the AWG system and, in some cases, can lead to erratic and unreliable freshwater sources, such as in dry geographic areas with low humidity. Therefore, a need exists for automated control of AWG systems to optimize the performance, reliability, and energy usage of such systems. Summary of the Invention

[0006] Certain embodiments of the present disclosure provide for the use of artificial intelligence to control and improve production of atmospheric water generating systems. Various embodiments are directed to an AWG system including a control system, such as a standalone control system or edge device connected to a centralized server, and one or more controllers for automatically implementing optimized atmospheric water generation operations based at least in part on real-time, historical, and forecasted parameters for the environment. In doing so, an AWG system is provided that utilizes artificial intelligence to process inputs from multiple sources, including local and external sensors, user input, external information sources, and / or the like, and identify optimal operation (e.g., timing and utilization) of the AWG system in real time or near real time. This, in turn, enables control instructions to automatically control one or more subsystems of the AWG system according to optimal system operation.

[0007] In some embodiments, a computer-implemented method includes receiving, by one or more processors, one or more optimization inputs for times and locations associated with operation of an atmospheric water generating system; generating, by the one or more processors and using an optimization machine learning model, one or more time-based energy forecasts for the atmospheric water generating system based at least in part on the one or more optimization inputs; and communicating, by the one or more processors, one or more control instructions to one or more controllers of the atmospheric water generating system based at least in part on the one or more time-based energy forecasts to initiate one or more atmospheric water generation operations.

[0008] In various embodiments, the optimization machine learning model is pre-trained using one or more supervised training techniques based at least in part on a training dataset including a plurality of labeled optimization training entries, each of the plurality of labeled optimization training entries including a set of historical optimization inputs and historical performance data corresponding to the set of historical optimization inputs, the historical performance data indicating ground truth water output from the atmospheric water generating system based at least in part on one or more historical atmospheric water generating operations.

[0009] In various embodiments, the computer-implemented method further includes receiving, by the one or more processors, energy usage data and performance data corresponding to the one or more atmospheric water production operations, and storing, by the one or more processors, the one or more optimization inputs, the energy usage data, and the performance data as labeled optimization training entries in a training dataset.

[0010] In various embodiments, the computer-implemented method further includes retraining, by the one or more processors, the optimized machine learning model based at least in part on the labeled optimized training entries.

[0011] In various embodiments, the computer-implemented method further includes generating, by the one or more processors, an optimized energy output for the atmospheric water generating system based at least in part on one or more time-based energy forecasts and water usage data for a location corresponding to the atmospheric water generating system, and generating, by the one or more processors, one or more control instructions based at least in part on the optimized energy output.

[0012] In various embodiments, the water usage data is based at least in part on user input or sensor data from the atmospheric water generating system, and the one or more optimization inputs include sensor data from the atmospheric water generating system or current or future weather data from one or more external sources.

[0013] In various embodiments, the atmospheric water generating system is associated with a cluster of a plurality of connected atmospheric water generating systems, each of the plurality of connected atmospheric water generating systems being associated with a different location, and the one or more optimization inputs include remote sensor data from each of the plurality of connected atmospheric water generating systems.

[0014] In various embodiments, communicating the one or more control instructions includes providing, by the one or more processors, the one or more control instructions to an edge device that is (i) physically disposed on the atmospheric water generating system and (ii) electrically connected to at least one of the one or more controllers of the atmospheric water generating system.

[0015] In some embodiments, a computing system comprises a memory and one or more processors communicatively coupled to the memory, wherein the one or more processors are configured to receive one or more optimization inputs for times and locations associated with operation of the atmospheric water generating system, use an optimization machine learning model to generate one or more time-based energy forecasts for the atmospheric water generating system based at least in part on the one or more optimization inputs, and communicate one or more control instructions to one or more controllers of the atmospheric water generating system to initiate one or more atmospheric water generation operations based at least in part on the one or more time-based energy forecasts.

[0016] In various embodiments, the optimization machine learning model is pre-trained using one or more supervised training techniques based at least in part on a training dataset that includes a plurality of labeled optimization training entries, each of the plurality of labeled optimization training entries including a set of historical optimization inputs and performance data corresponding to the set of historical optimization inputs.

[0017] In various embodiments, the performance data indicates ground truth water output from the atmospheric water generating system based at least in part on one or more historical atmospheric water generating operations.

[0018] In various embodiments, the one or more processors are further configured to generate an optimized energy output for the atmospheric water generating system based at least in part on one or more time-based energy forecasts and water usage data for a location corresponding to the atmospheric water generating system, and to generate one or more control instructions based at least in part on the optimized energy output.

[0019] In various embodiments, the water usage data is based at least in part on user input or sensor data from the atmospheric water generating system. In various embodiments, the one or more optimization inputs include sensor data from the atmospheric water generating system. In various embodiments, the one or more optimization inputs include current or future weather data from one or more external sources.

[0020] In some embodiments, the atmospheric water generating system comprises a control system comprising one or more controllers electrically connected to one or more subsystems of the atmospheric water generating system, a memory, and one or more processors communicatively coupled to the memory, wherein the one or more processors are configured to receive one or more optimization inputs for times and locations associated with operation of the atmospheric water generating system, use an optimization machine learning model to generate one or more time-based energy forecasts for the atmospheric water generating system based at least in part on the one or more optimization inputs, and communicate one or more control instructions to the one or more controllers to initiate one or more atmospheric water generation operations based at least in part on the one or more time-based energy forecasts.

[0021] In various embodiments, the atmospheric water generating system is associated with a cluster of a plurality of connected atmospheric water generating systems, each of the plurality of connected atmospheric water generating systems being associated with a different location, and the one or more optimization inputs include remote sensor data from each of the plurality of connected atmospheric water generating systems.

[0022] Reference is now made to the accompanying drawings, which are not necessarily drawn to scale. [Brief explanation of the drawings]

[0023] [Figure 1A] 1 shows a schematic diagram of an exemplary atmospheric water generating system according to one embodiment. [Figure 1B] 1 shows a schematic diagram of an exemplary atmospheric water generating system according to one embodiment. [Figure 2A] 1 illustrates an exemplary configuration of a membrane-based water extraction device according to various embodiments. [Figure 2B] 1 illustrates an exemplary configuration of a membrane-based water extraction device according to various embodiments. [Figure 2C] 1 illustrates an exemplary configuration of a membrane-based water extraction device according to various embodiments. [Figure 2D] 1 illustrates an exemplary configuration of a membrane-based water extraction device according to various embodiments. [Figure 2E] 1 illustrates an exemplary configuration of a membrane-based water extraction device according to various embodiments. [Figure 3A] 1 illustrates exemplary packing components within an absorber according to certain embodiments. [Figure 3B] 1 illustrates exemplary packing components within an absorber according to certain embodiments. [Figure 4] 1 illustrates an exemplary growth habitat in accordance with certain embodiments. [Figure 5] 1 illustrates an exemplary growth habitat in accordance with certain embodiments. [Figure 6] 1 illustrates an exemplary grow habitat configuration according to one embodiment. [Figure 7]1 illustrates an exemplary grow habitat having a solar canopy exhibiting different levels of opacity in different areas of the grow habitat, according to certain embodiments. [Figure 8] 1 illustrates a schematic diagram of an exemplary AWG computing ecosystem according to one embodiment. [Figure 9] 1 illustrates a schematic diagram of an exemplary AWG computing ecosystem in accordance with various embodiments. [Figure 10] 1 illustrates a schematic diagram of an exemplary AWG computing ecosystem in accordance with various embodiments. [Figure 11] 1 illustrates a schematic diagram of an exemplary AWG computing ecosystem in accordance with various embodiments. [Figure 12] 1 illustrates a schematic diagram of an exemplary AWG computing ecosystem in accordance with various embodiments. [Figure 13] 1 is a flowchart illustrating an example training process for generating an optimized machine learning model, according to some embodiments discussed herein. [Figure 14] 1 is a flowchart illustrating an example control process for optimally controlling an atmospheric water generating system using a trained optimization machine learning model, according to some embodiments discussed herein. DETAILED DESCRIPTION OF THE INVENTION

[0024] The present disclosure will more fully describe various embodiments with reference to the accompanying drawings. It should be understood that some, but not all, embodiments are shown and described herein. Indeed, embodiments may take many different forms, and thus, this disclosure should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. Like numbers refer to like elements throughout.

[0025] Overview Some embodiments provide an AWG system that extracts water from atmospheric air. Even with low humidity atmosphere (air having humidity greater than zero), a certain amount of water can be extracted from the air by contacting the air with a desiccant-rich solution under controlled conditions that lead to, at least in part, mass transfer of water from the air (where water is present in vapor form) to a diluent of the desiccant solution (where water is present in liquid form).

[0026] One exemplary AWG system may utilize a membrane separation module to separate water from a desiccant solution. The membrane separation module defines two flow paths separated by a permeable membrane. The desiccant solution (diluted desiccant solution after absorbing water from the atmosphere) flows on a first side of the permeable membrane, and permeate (e.g., in liquid and / or vapor form) flows on a second side of the permeable membrane. Water (e.g., water vapor) can permeate the permeable membrane from the first side to the second side of the permeable membrane. The desiccant solution remains on the first side of the membrane, and water remains on the second side of the membrane. The membrane separation module may additionally allow for control of environmental conditions within the membrane separation module to facilitate bulk water flow from the desiccant solution on the first side of the membrane across the permeable membrane to the flow of water (liquid water and / or water vapor) on the second side of the membrane. For example, the temperature and pressure may be increased on a first side of the membrane (e.g., by increasing the temperature and / or pressure of the desiccant solution entering the membrane separation module) and the temperature and pressure may be decreased on a second side of the membrane (e.g., by decreasing the temperature and / or pressure of the water stream on the second side of the membrane). Certain embodiments may include multiple membrane separation modules operating in series or parallel and / or may include additional systems for extracting water from the desiccant solution.

[0027] Other examples of AWG systems may not include membranes or may include alternatives to the membranes described herein. For example, embodiments of the present disclosure may be applied to membrane and / or membraneless AWG systems that utilize other water extraction processes, such as electrostatic nucleation, photosensitive adsorbent materials, and / or the like.

[0028] In certain embodiments, the AWG system may be integrated with one or more carbon dioxide filtration / capture modules, one or more greenhouse modules, one or more power generation modules, one or more control systems, and / or the like. For example, source intake air to the AWG system may be routed through a carbon dioxide capture system (after extracting water vapor from the atmosphere) before discharging dry, dehumidified air to the ambient environment. The captured carbon dioxide may be stored in a tank for later processing or may be released (e.g., in monitored amounts) into one or more greenhouse modules to increase the carbon dioxide concentration within the greenhouse, thereby increasing crop growth efficiency.

[0029] Additionally, a power generation module, which may include one or more renewable energy power generation systems, such as solar / photovoltaic, geothermal, and / or the like, or hydrocarbon fuel-based power generation systems, may be integrated with the AWG system to provide the electrical and / or thermal energy input required for the AWG process. If such a power generation module produces carbon dioxide or other exhaust gases, the power generation module exhaust gases may be routed through a carbon dioxide capture module to reduce the carbon dioxide production of the integrated system.

[0030] In certain embodiments, the AWG system may be integrated with one or more variable sun canopies. The variable sun canopies may be integrated into greenhouse modules (whereby the sun canopies define the exterior surface of the greenhouse). The variable sun canopies are configured to allow variable amounts of solar radiation to pass through the sun canopy to the interior of the greenhouse, thereby providing control of the growing environment (including the amount of light received) within the greenhouse environment. The variable sun canopies may provide location-specific adjustability of the opacity of the sun canopy to allow location-specific customization of the amount of light passing through the sun canopy. The opacity level of the sun canopy may be adjusted by applying an external signal, such as an electrical signal, to the sun canopy using one or more controllers.

[0031] The AWG system may be integrated with a greenhouse or any other water-consuming environment, including, by way of example, a biosphere environment, a residential structure, a recreational facility, and / or the like. In various embodiments, the AWG system may be configurable based at least in part on the water usage requirements of a particular environment. For example, the AWG system may generate water output based at least in part on input data such as demographic information (e.g., age, gender, weight, height, etc.) of the environment's occupants, water usage information (e.g., current and / or historical water usage at different times, etc.), water-dependent facilities (e.g., number of sinks, toilets, showers, sprinklers, etc.), water storage capacity (e.g., storage tank size, etc.), and / or any other data reflecting the end-user's water requirements. Due to the power requirements of the AWG system, the system may be optimized by selectively powering the system to meet the water requirements for a particular environment using the lowest possible input energy.

[0032] In certain embodiments, the AWG system may be integrated with a control system (e.g., located on the AWG system) configured to selectively power, run, initiate, schedule, and / or the like, one or more components of the AWG system. In other embodiments, the control system may communicate with the AWG system via the Internet. The control system may include one or more controllers coupled (e.g., via one or more wired and / or wireless connections, etc.) with and / or communicate with (e.g., via one or more wired and / or wireless connections, etc.) one or more subsystems of the AWG system, such as, for example, a humidity augmentation system, an air preconditioning system, a power generation module, one or more valves (e.g., electromechanical mixing valves, etc.), motors, actuators, and / or the like. As described in further detail herein, the control system may be configured to selectively control operation of the AWG system (e.g., via one or more control commands to one or more controllers, etc.) based at least in part on one or more optimization inputs. In some examples, the control system may autonomously operate the AWG system using an optimization module configured to meet water requirements for a particular environment using the lowest possible input energy.

[0033] In various embodiments, the optimization module may include computer-readable instructions configured to generate one or more energy outputs based at least in part on one or more optimization inputs. The computer-readable instructions may define any other type of model configured to map multiple optimization inputs to an energy-based output, such as, for example, a machine learning, rule-based, and / or time-based energy prediction and / or optimized energy output. The optimized energy output may include, for example, one or more real-time and / or scheduled operation instructions configured to initiate operation of one or more components of the AWG system.

[0034] In various embodiments, the optimization module may include a machine learning model, such as one or more supervised, unsupervised, semi-supervised, reinforcement, and / or similar learning models, configured to generate one or more energy-based outputs based at least in part on one or more optimization inputs. The machine learning model may be pre-trained, for example, using labeled and / or unlabeled training data reflecting multiple training optimization inputs.

[0035] In some examples, the training data may include historical performance data (e.g., water output, etc.) and / or energy usage data from the AWG system paired with multiple corresponding historical optimization inputs. For example, the historical performance and / or energy usage data may include ground truth labels for multiple corresponding historical optimization inputs. In some examples, the machine learning model may be continuously trained by recording real-time performance and / or energy usage data, matching the real-time data with corresponding optimization inputs to generate training pairs, and retraining the machine learning model with the new training pairs. This may advantageously allow for dynamic consideration of new and / or modified optimization inputs over time.

[0036] In certain embodiments, the control system may include a memory configured to store the optimization module. Additionally or alternatively, the control system may include one or more edge devices communicatively connected via one or more wireless networks to a centralized system configured to store, train, and / or implement the optimization module and / or its one or more machine learning models. By way of example, the centralized system may include a connected cloud platform, including a distributed serverless cloud environment, a dedicated cloud server environment, and / or the like. In some examples, the centralized system may train a machine learning model and provide the trained model to the control system. In such cases, the control system may locally implement the machine learning model to locally generate one or more optimized energy outputs. Additionally or alternatively, the control system may provide one or more optimization inputs to the centralized system and receive one or more energy-based outputs from the centralized system based at least in part on the optimization inputs.

[0037] In certain embodiments, the control system and / or centralized system may use an optimization module to generate one or more energy-based outputs based at least in part on multiple optimization inputs received from multiple information sources. The multiple information sources may include, for example, one or more weather forecast services, global time services, heat forecast services, water usage tracking services, adjacent AWG systems, and / or the like. In this manner, the control system may intelligently control the AWG system in real time by considering multiple inputs across an array of diverse information sources. Using some of the techniques of this disclosure, the control system may leverage artificial intelligence to continuously learn and adapt to different conditions over time.

[0038] atmospheric water resources The atmosphere is about 3,100 cubic miles (mi 3 ) or 12,900 cubic kilometers (km 3) of water. This amount, by volume, is roughly equal to the total amount of water held by the Great Lakes. As a natural resource, water vapor is constantly replenished by the natural closed-loop water cycle, which provides a nearly unlimited supply of water that can be extracted from the air without adverse environmental effects.

[0039] Atmospheric Water Generation The AWG process includes systems and methods for extracting water vapor from atmospheric source air, at least in part, by condensing the water vapor and recovering the condensed liquid water (e.g., by absorbing the liquid water into a desiccant solution). Certain embodiments can be combined with carbon dioxide capture systems as discussed herein. Certain embodiments include preconditioning and / or compressing raw source air (e.g., air at atmospheric conditions) to facilitate the water extraction process, and / or condensing water vapor trapped in the raw source air (e.g., by increasing the humidity of at least a portion of the raw source air) to maximize the amount of water vapor that can be extracted from a given unit volume of source air. As discussed herein, treated source air may be compressed, compacted, and / or otherwise manipulated, for example, through one or more processes to facilitate the water extraction process.

[0040] Finally, various embodiments of the AWG process include a condensing mechanism that can direct source air (untreated source air and / or treated source air as discussed herein) across one or more condensing surfaces, each having a surface temperature below the dew point of the source air. As the source air flows across and / or around the condensing surfaces, the temperature of the source air adjacent the condensing surfaces drops (e.g., through convective heat transfer), water vapor in the source air condenses on the condensing surfaces, and the condensed liquid water flows into a storage vessel (e.g., a collection tank) and / or to one or more associated modules (e.g., greenhouse modules) for immediate use.

[0041] Air preconditioning As noted above, raw source air can be preconditioned to facilitate a water extraction process that ultimately condenses the water vapor into usable liquid water. In certain embodiments, the preconditioning process can include steps to compress the air to increase the vapor pressure of the air (thereby biasing a larger volume of water toward the liquid state rather than the vapor state) and / or to reduce the temperature of the source air to a temperature closer to the dew point. In certain embodiments, the air preconditioning systems described herein can be utilized before and / or after a humidity augmentation system, such as a desiccant-based humidity augmentation system as described herein. Moreover, the air preconditioning system can be utilized before and / or after a carbon dioxide capture system as discussed herein.

[0042] By way of example only, the air preconditioning process may include a series of compressors / pumps, venturi valves, vortex valves, manifolds, and / or the like collectively configured to reduce the temperature of the source air closer to the air dew point and / or increase the pressure of the air before removing water vapor from the air (e.g., through condensation or absorption by a desiccant). For example, raw source air may be drawn into the air preconditioning system via a vacuum created at an inlet through a compressor 101 (e.g., a turbine / blower compressor having multiple stators or variable-pitch turbine blades controllable via a servo motor) and / or a centrifugal fan configured to increase the pressure of the raw air entering the air preconditioning system. In certain embodiments, the compressor and / or centrifugal fan may be rotated via one or more electric motors (which may receive input power from one or more power systems in communication with the air preconditioning system) mechanically connected to the compressor and / or centrifugal fan via gear transmissions, belt drives, chain drives, and / or the like.

[0043] In embodiments including a centrifugal fan, particles, dust, and other heavy air contaminants are rotated toward the outermost edge of the centrifugal fan, where they are removed from the air stream and expelled from the air pre-conditioning system.

[0044] In certain embodiments, the filtered air can be directed into a carbon dioxide capture column where it is passed over a fixed absorption bed configured to absorb carbon dioxide from the air, as discussed in more detail herein. The carbon dioxide can be separated via a compressor and directed away from the air stream.

[0045] In certain embodiments, the filtered air (with reduced carbon dioxide content) can be further directed through an air preconditioning system into a primary manifold where the air is split in a ratio selected by a variable plenum / valve. From the primary manifold, the first air stream continues along with the bulk air stream, and the second air stream is directed to a vortex tube manifold as discussed herein.

[0046] The bulk air stream can proceed through one or more venturi valves, each configured to reduce the pressure and temperature of the bulk air stream (the volume and amount of air remains constant across each venturi valve while the pressure is reduced, thereby reducing the temperature of the air stream proportionally), and / or through a precooler (e.g., a heat exchanger through which a cooling fluid is passed). After processing through the one or more venturi valves and / or precooler, the bulk air stream can proceed to a temperature measurement section, where the temperature (e.g., dry-bulb and wet-bulb temperatures) of the bulk air stream are measured by one or more temperature measurement devices (e.g., thermometers) to determine the dew point of the bulk air stream. Output from the temperature measurement devices can be utilized by a controller to mix the bulk air stream with at least a portion of the vortex-cooled air stream to lower the temperature of the bulk air to closer to the air dew point. For example, the controller can be in electrical communication with an electromechanical mixing valve that can be selectably opened or closed to vary the amount of vortex-cooled air introduced into the bulk air stream. Based at least in part on the determined dry-bulb and / or wet-bulb temperatures (as monitored by the controller), the controller can send a signal to the motor to move the electromechanical valve to a desired position to obtain a desired mixture of the vortex-cooled air with the bulk air stream.

[0047] The vortex-cooled air begins as a second stream of air exiting the primary manifold. The second stream of air exits the primary manifold and proceeds to a vortex tube manifold, where it is pressurized (e.g., via a compressor) to a pressure sufficient to achieve a temperature drop of approximately 70 to 150 degrees Fahrenheit for the air traveling through one or more vortex tubes 106. For example, the air may be pressurized to at least approximately 70 to 120 PSI before being directed into one or more vortex tubes. Each vortex tube includes an inlet port that directs the air stream tangentially into the interior spin chamber. As the air enters the spin chamber, it gains some angular momentum, causing dense warm air to migrate toward the outer periphery of the spin chamber and exit through the exhaust valve. In certain embodiments, the warm air can be utilized to heat a carbon dioxide capture column. The remaining vortex-cooled air migrates toward the center of the spin chamber and exits through the vortex outlet. As noted above, the vortex-cooled air can mix with the bulk air stream to lower the temperature of the bulk air stream closer to the dew point. As yet another alternative, the vortex-cooled air can be utilized to cool a precooler through which the bulk air passes.

[0048] In certain embodiments, the mixed and cooled bulk air stream is then directed into a condensation chamber, where water vapor in the air condenses into liquid water. By way of example only, the bulk air stream can be directed across a series of condensation surfaces (e.g., cooling plates, screens, tubes, and / or the like) configured to reduce the local temperature of the air at the condensation surfaces below the air dew point, thereby causing water vapor to condense on the condensation surfaces. The condensed water can then be routed from the condensation surfaces into a holding chamber for collection and later use. However, it should be understood that any of a variety of condensation mechanisms may be used. For example, as discussed herein, one or more desiccant-based condensation mechanisms may be utilized to more effectively remove water vapor from the bulk air stream. Moreover, in certain embodiments, an air preconditioning system may be omitted, and untreated air may be filtered and / or directly directed into the condensation chamber. Such embodiments may have lower input power requirements and therefore reduce the amount of power needed to generate water.

[0049] It should also be understood that certain preconditioning system embodiments may include, instead of or in addition to the vortex and venturi valve mechanisms discussed herein, one or more filters (e.g., woven-based air filters, nonwoven-based air filters, and / or the like), one or more refrigeration systems (e.g., warm air passed through a heat exchanger to lower the temperature of the air closer to the dew point), and / or the like.

[0050] Use of desiccants in AWG systems As noted above, certain embodiments include one or more subsystems configured to increase the humidity of a portion of the source air to increase the amount of water that can be extracted from the source air. Specifically, water vapor can be extracted from a first larger volume of source air and reintroduced into a second, smaller volume of source air, thereby increasing the humidity of the second volume of source air before the water vapor in the source air compacts and the water vapor in the second volume of source air condenses into liquid water.

[0051] An AWG system as discussed herein includes at least one air scrubber comprising a column for contacting atmospheric air (e.g., after increasing the humidity of the air) with a desiccant. In certain embodiments, the desiccant solution may be a fluid, gel, and / or the like within the typical operating temperature range discussed herein. The desiccant may be selected from any of a variety of ionic solutions capable of absorbing water, such as lithium chloride (LiCl), lithium bromide (LiBr), calcium chloride (CaCl), triethylene glycol, and / or the like. Other unlisted compounds with hygroscopic properties may be provided for use as the desiccant solution in certain embodiments. The hygroscopic fluid may also comprise a surfactant and / or a nanofluid. In certain embodiments, the desiccant solution may include a mixture of multiple ionic solutions, such as a mixture of LiCl and CaCl solutions. The desiccant may be dissolved in water to provide a concentrated desiccant solution that can be pumped (e.g., via a liquid pump) through at least one desiccant column. Other materials, including gels, aerogels, desiccant particles capable of flowing under the influence of particle flow principles, and / or the like, may be utilized in place of the desiccant solution in certain embodiments.

[0052] Moreover, the amount of water vapor that can be absorbed by (and / or released by) a desiccant depends on the vapor pressure and temperature of the closed system containing the desiccant. Thus, various embodiments are configured to absorb water from the air into the desiccant while the vapor pressure and temperature within the closed system are high and low, and those same embodiments are configured to extract water from the desiccant while the vapor pressure and temperature are low and high.

[0053] Water absorption from the atmosphere Water can be extracted from the air via one or more absorbent modules. An AWG system (one example of which is shown as system 600 in FIGS. 1A-1B) may incorporate a single absorbent module or multiple absorbent modules. The absorbent modules are connected to additional modules in the AWG system, including a water extraction module, discussed in more detail below, allowing desiccant to flow between the water absorbent module and the additional modules in the AWG system. The flow path between the absorbent modules in the overall AWG system may be configurable (e.g., via a valve between an open and a closed configuration) so that the absorbent module can operate as a closed module in a batch operation mode (with desiccant flowing only within and / or between multiple absorbent modules) before passing the desiccant through the water extraction module. In other embodiments (e.g., when the valve is in the open configuration or in an embodiment without a valve), the desiccant flows freely between the absorbent module and the water extraction module, such as in a continuous flow operation configuration.

[0054] Additionally, as mentioned, air entering the absorption module (e.g., entering absorber 610 via air flow path 601 and exiting the absorber as dry air via flow paths 602-604 shown in FIG. 1A) may flow from a pre-conditioning module configured to increase the humidity of the air prior to extracting water from the air within absorber 610. In other embodiments, air may flow directly into the absorption module from the ambient atmosphere external to the AWG system.

[0055] An absorption module according to certain embodiments includes an absorber 610 configured to contact ambient air / atmospheric air (e.g., after preconditioning) with a desiccant (flowing through absorber 610 along flow paths 621-622) to absorb water from the atmosphere into the desiccant. Absorber 610 may be embodied as a vessel in which the desiccant flows between an inlet (via flow path 621) and an outlet (via flow path 622), and the ambient air / atmospheric air flows between an air inlet (flow path 601) and an air outlet (flow paths 602-604, including blower 603). Within the vessel, the ambient air contacts the desiccant, enabling bulk transfer of water vapor from the ambient air to the desiccant. In certain embodiments, one or more baffles, flow interrupters, or packing (e.g., structured or random packing) may be positioned within the absorber to increase the surface area of the desiccant and / or increase the contact time between the ambient air and the desiccant.

[0056] In certain embodiments, absorber 610 is configured in a counterflow configuration, with ambient air entering absorber 610 near the bottom of absorber 610. Dry ambient air then exits absorber 610 near the top of absorber 610 via specific flow paths (with pump / blower 603 utilized to move the air through absorber 610). Concentrated desiccant (e.g., desiccant fluid) enters absorber 610 near the top via inlet flow path 621 and flows downward through the interior of absorber 610 by gravity. Water is absorbed into the desiccant from the ambient air such that diluted desiccant exits the absorber near the bottom of absorber 610 along outlet flow path 622. In certain embodiments, flow modifiers, such as barriers, mesh, packing components, and / or bends in the outlet piping from absorber 610, may be used to reduce desiccant carryover at the air outlet of absorber 610. In certain embodiments, the flow modifier may be positioned inside the absorber 610 adjacent to the top of the absorber 610 (e.g., at the mouth of the exhaust port for dry ambient air to exit the absorber via the exhaust air flow path).

[0057] In certain embodiments, absorber 610 operates in a cross-flow configuration, where air enters the absorber from one side of absorber 610 and crosses (e.g., at least substantially horizontally) to the other side. Concentrated desiccant (e.g., desiccant fluid) enters the top of absorber 610 and absorbs water from the ambient air as it flows to a lower point within absorber 610, where diluted desiccant exits near the bottom. In this manner, the air flow is at least substantially perpendicular to the flow of desiccant within absorber 610. In this configuration, the air inlet and air outlet are at approximately the same height above absorber 610.

[0058] In a particular configuration, the absorber 610 is operated in a cross-counterflow configuration, where air enters the absorber 610 from one side and crosses to the other side. Concentrated desiccant (e.g., desiccant fluid) enters near the top of the absorber 610 and absorbs water from the ambient air as it flows to a low point within the absorber 610, where diluted desiccant exits near the bottom of the absorber 610. In this configuration, the air inlet and air outlet are horizontally offset from one another. In this configuration, the air inlet can be on a side of the absorber 610 near the top of the absorber 610, and the air outlet is on a side of the absorber 610 near the bottom. In another orientation of this configuration, the air inlet is positioned on a side of the absorber 610 near the bottom of the absorber 610, and the air outlet is positioned on the opposite side of the absorber 610 near the top of the absorber 610. The air travels along a path that extends through the absorber 610 at an angle, either from the top to the bottom of the absorber 610 or from the bottom to the top of the absorber 610 .

[0059] In one particular embodiment, the interior of absorber 610 comprises multiple packing elements around which concentrated desiccant flows as it absorbs water extracted from humid ambient air. The packing elements are provided to increase the surface area of concentrated desiccant flowing within absorber 610 and to provide a highly tortuous flow path for ambient air flowing through absorber 610 so that the air has turbulent flow through the interior of absorber 610. Absorber 610 may be embodied as a counterflow vessel as described above, where concentrated desiccant (e.g., desiccant fluid) enters absorber 610 at a desiccant vessel inlet located at or near the top of absorber 610 and an ambient air inlet is located at the bottom of absorber 610. Ambient air flows upward to a dry air outlet located at or near the top of absorber 610, and desiccant fluid flows downward across the packing elements to a diluted desiccant outlet of absorber 610. By way of example, the packing elements may include individual blocks, balls, trays, baffles, and / or any other shape defining multiple baffles, slits, holes, mesh, and / or other flow-modifying elements that may be positioned within the absorber 610 to collectively define a highly tortuous path for the ambient air and desiccant fluid to pass through the absorber 610. The packing elements may include (or be formed from) materials that are not reactive with the desiccant fluid. Exemplary packing elements are illustrated in FIGS. 3A-3B. In certain embodiments, multiple packing elements (such as the unstructured packing element shown in FIG. 3B) may be positioned within the absorber 610 without physically connecting the packing elements to one another. In other embodiments, a single packing element (such as the structured packing element shown in FIG. 3A) specifically sized and shaped to fit the interior of the absorber 610 may be provided and positioned within the absorber 610.

[0060] The packing components of certain embodiments may be arranged in a structured configuration to define channels, with or without holes, set at different angles to one another that collectively define structured flow paths for the ambient air and desiccant fluid flowing through the absorber 610. To provide a structured packing configuration, the packing components are stacked sequentially and arranged within the absorber 610. The packing components may also be randomly positioned, with multiple geometrically shaped components randomly arranged within the absorber 610 to increase surface area. While discussed as a packing-based absorber, it should be understood that the desiccant fluid may pass through the absorber 610 through other configurations, such as by atomizing the liquid desiccant fluid by spraying the desiccant fluid into the absorber and / or the like.

[0061] In use of the overall AWG system, ambient air (at ambient temperature and humidity) is directed into absorber 610 (although not shown, a blower may be implemented at the inlet to absorber 610 to increase the volumetric flow rate of ambient air entering absorber 610). In absorber 610, the ambient air contacts a concentrated desiccant (e.g., desiccant fluid) that is provided to absorber 610 at a low temperature, increasing the vapor pressure within the absorber and encouraging water vapor in the humid ambient air to condense and be absorbed by the desiccant fluid while the humid air is in contact with the concentrated desiccant. As the ambient air and desiccant fluid flow through absorber 610, moisture in the air condenses and / or is otherwise absorbed into the desiccant fluid, diluting the desiccant fluid and drying the air. The dry air then exits the absorber and returns to the atmosphere, as shown at 602. As shown, blower 603 may be incorporated into the ambient air outlet of absorber 610 to increase the volumetric flow rate of air passing through absorber 610. Blower 603 may be provided in addition to, or as an alternative to, the above-mentioned blower located at the ambient air inlet of absorber 610. Moreover, as the desiccant fluid passes through absorber 610, diluted but still cool desiccant fluid exits absorber 610, as shown at 622.

[0062] According to certain embodiments, the diluted desiccant (e.g., desiccant fluid) exits absorber 610 and is directed to pump 623. In certain operations, the absorber module may be operated in a batch configuration, in which a series of valves may be configured to isolate the absorber module from the rest of the AWG system, thereby recirculating the diluted desiccant along a recirculation flow path (while preventing additional concentrated desiccant fluid from entering the closed loop via flow path 637 while appropriate valves remain closed), and through a heat exchanger 625 (e.g., a shell-and-tube heat exchanger, a plate heat exchanger, and / or the like) prior to the absorber to cool the diluted desiccant (on the other side of the heat exchanger is cooling water collected from the overall system, as discussed in more detail herein) before returning it to the top of absorber 610 as shown at 621. In this manner, the amount of water absorbed into the desiccant fluid may be increased (thereby increasing the level of dilution of the desiccant) before the desiccant is directed to the evaporative portion of the overall system.

[0063] In certain embodiments, the absorption module includes a pre-absorber heat exchanger 625 that is cooled by use of a chiller that uses a cooling medium (e.g., water, glycol, and / or the like) to cool the concentrated desiccant that flows into the pre-absorber heat exchanger via flow path 624 before the concentrated desiccant enters the absorber 610 (e.g., the cooling medium is separate from the desiccant, such as on the opposite side of the heat exchanger).

[0064] In one particular operation, the absorber module may be operated in a serial configuration, with the valves configured to recirculate a constant amount of dilute desiccant fluid along the recirculation flow path, and a constant amount of dilute desiccant fluid flowing toward the water extraction module through a separate flow path (flow path 626) connecting the absorber module and the water extraction module. In this configuration, a constant amount of dilute desiccant fluid flows along the recirculation flow path through a heat exchanger 625 (e.g., a shell-and-tube heat exchanger, a plate heat exchanger, and / or the like) before the absorber to cool the dilute desiccant fluid (the other side of the heat exchanger 625 is cooled using water collected from the entire system) before returning it to the top of the absorber 610, as shown at 621. The dilute desiccant fluid simultaneously traverses along separate flow path 626 to the water extraction module. The cooled fluid passes through flow paths 661-667, which contain the fluid recirculation loop, as well as a cooling system 662 and a pump 664, which ensures sufficient fluid flow through the heat exchanger 625 before the absorber. A portion of the fluid returns to the fluid reservoir via flow path 667 .

[0065] In certain embodiments, absorber 610 is configured so that the concentrated desiccant is not cooled in a heat exchanger. In this embodiment, cooling of the desiccant fluid may be provided via conductive heat exchange with ambient air through conductive piping along certain flow paths leading to the absorber intake. Fluid cooling may be provided within absorber 610 as sensible heat is exchanged with the atmosphere, provided the air temperature is lower than the incoming desiccant fluid temperature.

[0066] By way of example only, the concentrated desiccant fluid traveling toward the absorber 610 may be directed through a series of geothermal tubes that have heat transfer characteristics with the surrounding ground below the AWG system. The concentrated desiccant fluid may pass directly through the series of geothermal tubes, or the concentrated desiccant fluid may pass through a dual-fluid heat exchanger against a cooling fluid that is maintained at a desired low temperature via geothermal cooling. As yet another example, the desiccant may pass through a heat exchanger (e.g., a shell-and-tube heat exchanger) to cool the desiccant fluid. The heat exchanger may be cooled via a cooling solution that passes through a refrigeration circuit and / or other fluid cooling device to absorb heat from the desiccant fluid before it enters the absorber 610.

[0067] As yet another example, a single-stage hydraulic consolidation system may be positioned adjacent to a high-pressure gas well, such as adjacent to a natural gas well, an oil well (where natural gas is co-extracted with oil), and / or the like. The high-pressure gas may be directed through one or more expansion valves to regulate and / or reduce the pressure of the incoming gas, which experiences a rapid temperature decrease due to the Joule-Thompson effect (according to the gas law equation, the pressure of the gas drops rapidly across the valve, while the volume and amount of gas remains substantially constant, thereby causing a proportionally rapid temperature decrease across the expansion valve). The expanded subcooled gas may pass through a heat exchanger against a concentrated desiccant fluid, thereby absorbing heat from the concentrated desiccant fluid and reducing the temperature of the desiccant fluid before entering the absorber 610. The expanded gas may then be directed away from the AWG system, where it may be collected for future use, flared, utilized to generate electricity (e.g., via a steam turbine), and / or utilized to heat desiccant fluid entering the water extraction module, as discussed herein.

[0068] In certain embodiments, absorber 610 is configured such that diluted desiccant fluid exits absorber 610 and is sent to the water extraction module without a recirculation path. The concentrated desiccant fluid returning from the water extraction module may or may not be cooled in a heat exchanger and / or chiller and / or geothermal cooling before entering absorber 610. For example, as discussed above, the diluted desiccant exiting absorber 610 is pumped along a flow path extending between the absorber module and the water extraction module when appropriate valves are closed to prevent the diluted desiccant from recirculating into absorber 610.

[0069] In certain embodiments, the absorber 610 may be embodied as a membrane-separated absorber, having a desiccant flow path on a first side of a porous membrane and an air flow path on a second, opposite side of the porous membrane. Separating the air flow path from the desiccant solution flow path may prevent undesirable bulk flow of the desiccant salt (e.g., aqueous desiccant salt) itself into the air flow path and ultimately out of the AWG system. Water may be absorbed from the air by the desiccant based at least in part on osmotic water flow through the membrane from the air to the desiccant solution. Water vapor may condense on the second side of the membrane, travel through the membrane pores by capillary action, and be absorbed by the high-salt concentrated desiccant solution. In certain embodiments, the membrane is embodied as a porous membrane having pores of sufficient size (e.g., average pore size, maximum pore size, and / or the like) to allow water molecules to pass through the membrane, but prevent the desiccant salt from passing through the membrane. By way of example, the porous membrane may be a nonwoven material such as polytetrafluoroethylene (PTFE), expanded polytetrafluoroethylene (ePTFE), polypropylene (PP), polyvinylidene fluoride (PVDF), and / or the like. In certain embodiments, other materials may additionally be used, including nylon and / or other synthetic materials. In various embodiments, synthetic and / or natural materials may be utilized. In certain embodiments, the porous membrane may be at least partially woven. In certain embodiments, the membrane may be organic, inorganic, polymeric, mesoporous, ceramic, and / or the like. In certain embodiments, the membrane may comprise metal-organic frameworks, carbon nanotubes, and / or combinations thereof. The membrane may be hydrophilic and / or hydrophobic, or may be treated (e.g., with a coating) to enable the membrane to be hydrophilic and / or hydrophobic. Exemplary membrane geometries include spiral wound, flat plate and frame style, or tubular. As the desiccant solution and water flow across opposite sides of the membrane, water molecules migrate (via capillary action) through the membrane from the desiccant solution to the permeate stream (e.g., water vapor and / or liquid water).In certain embodiments, the membrane may be supported within the frame via spacers such as gratings, grids, and / or the like to provide mechanical support to the membrane so as to maintain a desired orientation of the membrane within the housing.

[0070] Various embodiments of the absorption module include a multi-stage absorber configuration that includes multiple absorbers arranged in series such that the desiccant flows from a first absorber, through a second absorber, and sequentially through additional absorbers in the multi-stage absorber configuration. Alternatively, the multi-stage absorber configuration encompasses multiple absorbers arranged in parallel such that the desiccant is divided and flows in parallel through the multiple absorbers.

[0071] Additionally, multiple absorbers may be arranged in series within the air flow path, such that source air is drawn from the environment and passed through multiple absorbers in series before being discharged back to the environment as dry air. For example, source air may first pass through a low-concentration absorber to absorb a first amount of water from the air, and then through a high-concentration absorber to absorb a second amount of water from the air. Because initial absorption requires less energy (and does not require a low vapor pressure between the air and the liquid desiccant), initial absorption using a lower concentration of desiccant fluid allows for the absorption of the first amount of water from the air. After the initial, low-energy absorption process is complete, the air (still containing water vapor) passes through a second absorber with a higher concentration of desiccant fluid, such that a second amount of water is absorbed from the air. The now dry (e.g., low humidity) air may then be discharged from the system to the environment.

[0072] On the desiccant side, once the diluted desiccant exits the low concentration absorber, the desiccant fluid passes through a water extraction module as discussed herein to extract water therefrom (and consequently reconcentrate the desiccant).

[0073] In certain embodiments, each absorber may be in fluid communication with a corresponding water extraction module, such that each absorber is associated with a separate and independent desiccant flow path loop. For example, a first amount of desiccant fluid may flow between the first absorber and the first water extraction module, and a second amount of desiccant fluid may flow between the second absorber and the second water extraction module, with the first amount of desiccant fluid immiscible with the second amount of desiccant fluid. In certain embodiments, the first amount of desiccant fluid may include a first desiccant (e.g., LiCl), and the second amount of desiccant fluid may include a second desiccant (e.g., CaCl).

[0074] Moreover, in embodiments including multiple independent desiccant streams, each desiccant stream can have a different concentration range. For example, a first desiccant stream (e.g., corresponding to a first absorber through which source air passes) can have a first concentration range measured between a high concentration value at the outlet of the water extraction module and a low concentration value at the outlet of the absorber, and a second desiccant stream can have a second concentration range. When source air is directed through successive absorption columns, the air can be directed first through the absorption column with the low concentration range and second through the absorption column with the high concentration range.

[0075] Membrane-based water extraction from desiccants The water extraction module is provided to remove water from the diluted desiccant (e.g., from the diluted desiccant fluid) for storage and use as drinking water or other clean liquid water uses. The water extraction module comprises one or more membrane-based water extraction devices (e.g., connected in series or parallel) and is connected to the absorption module via various flow paths, allowing the desiccant (e.g., diluted desiccant and concentrated desiccant) to flow between the absorption module and the water extraction module. The membrane-based water extraction device (also called a fluid separation device) also acts to concentrate the desiccant solution into a concentrated desiccant retentate solution that can be reprocessed through the AWG system.

[0076] In some embodiments, the water extraction module can include at least one membrane-based water extraction device 633 defining two flow paths separated by a permeable membrane 633a. On a first side of the permeable membrane 633a, the desiccant flows along the desiccant flow path (between flow paths 632 and 634), and on the opposite second side of the permeable membrane, the permeated fluid (e.g., water vapor) is collected and then transferred (in a portion 633c located within the water extraction device between flow paths 650 and 651). Both the desiccant (on the first side 633b of the permeable membrane 633a) and the collected permeate fluid (on the second side 633c of the permeable membrane 633a) are in contact with both sides of the permeable membrane 633a as they flow through the permeable membrane 633a. The membrane 633a can thereby separate the desiccant flow path from the recovered permeable (water vapor) flow path, which includes water vapor collected during the water collection / fluid separation process (e.g., water vapor migrating through the porous membrane 633a) and condenses into a liquid permeable water stream (e.g., downstream of the membrane-based water extraction device). In certain embodiments, water vapor mass transport across the membrane 633a can be driven by increasing the vapor pressure of the desiccant. This can be achieved by heating the desiccant (on the desiccant side 633b of the membrane) before contacting the membrane 633a. Reducing the pressure on the permeate side 633c of the membrane 633a by lowering the pressure (e.g., by using a vacuum mechanism) also triggers water vapor mass transport across the membrane. As the desiccant flows through the porous membrane 633a (on the first desiccant side 633b of the membrane), water within the desiccant begins to permeate across the membrane 633a in a vapor state and exits the membrane 633a (on the second side) in a vapor state. The water vapor is then cooled (e.g., at least in part, by utilizing a heat exchanger (e.g., condenser 627) and / or by contact with a cooler fluid (e.g., condensed water)) along the water flow path and condenses as it flows.

[0077] In certain embodiments, the desiccant side 633b may be heated by a heating fluid (e.g., heated oil, steam, glycol, and / or the like) separated from the desiccant flow path within the membrane-based water extraction device via a thermally conductive layer (e.g., a thermally conductive, non-porous film, metal plate, and / or the like). In certain embodiments, the desiccant side 633b may be embodied as a shell-and-tube heat exchanger in which the heating fluid flows through the tubes and the desiccant flows through the shell, certain walls of which are embodied as the porous membrane 633a. In certain embodiments, the heating fluid may be a product of the AWG system, thereby utilizing sensible heat transfer to heat the desiccant fluid and promote water vapor migration across the membrane. In other embodiments, the heating fluid may be a product (e.g., a final product, a waste product, or an intermediate product) of a spatially proximate process such as mining, gas extraction, power generation, and / or the like.

[0078] As discussed in more detail herein, the permeate side 633c may similarly comprise a heat exchanger configuration to reduce the temperature of the permeate fluid to promote its condensation. For example, a cooling fluid (e.g., liquid water extracted from the liquid water storage tank 655, a refrigerant (e.g., glycol), and / or the like) may be separated from the permeate flow path by a thermally conductive layer (e.g., a thermally conductive, non-porous film, a metal plate, and / or the like). The cooling fluid cools the thermally conductive layer, thereby providing a surface within the permeate fluid stream upon which water vapor may condense. In certain embodiments, the permeate side 633c may be embodied as a shell-and-tube heat exchanger, in which the cooling fluid flows through the tubes and the permeate fluid flows through a shell, certain walls of the shell being embodied as a porous membrane 633a. In such embodiments, the outer surfaces of the tubes provide a condensation surface for water vapor. In certain embodiments, the cooling fluid can be a product of the AWG system (e.g., liquid water), thereby utilizing sensible heat transfer to cool the permeate, promote water vapor transport across the membrane, and promote condensation of the water vapor.

[0079] The membrane 633a may include a hydrophobic porous membrane, such as a nonwoven membrane with small pore sizes. By way of example only, the membrane may include PTFE, ePTFE, PP, PVDF, and / or the like, which are hydrophobic by design. In certain embodiments, the membrane may be organic, inorganic, polymeric, mesoporous, ceramic, and / or the like. In certain embodiments, the membrane may include metal-organic frameworks, carbon nanotubes, and / or combinations thereof, such as by stacking layers of materials. The membrane may be hydrophilic or hydrophobic, or may be treated (e.g., with a coating) to enable the membrane to be hydrophilic and / or hydrophobic. The use of a hydrophobic material (or a material with a hydrophobic coating) promotes the selective passage of water vapor through the membrane and retention on the permeate side 633c of the membrane. Applying a layer of hydrophobic material to the membrane on the desiccant side 633b allows only volatile vapors to pass through, while the liquid fluid is retained on the desiccant side 633b of the membrane. As examples of membrane geometries, the membrane can be spiral wound, flat plate and frame style, or hollow tubular. As the desiccant solution flows across the desiccant side 633b of the membrane, water molecules migrate (via capillary action) from the desiccant solution through the membrane in vapor form to the permeate side 633c, leaving a concentrated desiccant solution on the desiccant side 633b of the membrane 633a.

[0080] The membrane-based water extraction device 633 is embodied as a housing having a desiccant inlet and a desiccant outlet on the desiccant side 633b of the membrane 633a and a permeate inlet and a permeate outlet on the permeate side 633c of the membrane 633a. In other exemplary embodiments, the membrane-based water extraction device 633 may utilize gravity to remove permeate fluid from the membrane-based water extraction device 633; in such embodiments, the permeate flow path need not include an inlet (whereby gravity alone is sufficient to move the permeate fluid through the outlet of the device). For example, as the desiccant fluid flows from the inlet to the outlet of the membrane-based water extraction device, water vapor passes through the membrane and transitions to vapor form on the permeate side of the membrane. The permeate side of the membrane may be cooled (e.g., using a cooling fluid separated from the permeate side of the membrane by a thermally conductive film), and water may condense within the permeate side of the membrane-based water extraction device or fall (under gravity) through an outlet port located at the bottom end of the permeate side of the membrane.

[0081] The desiccant flow channel extends between the desiccant inlet and the desiccant outlet. The permeate (water) flow channel extends between the permeate inlet and the permeate outlet. In certain embodiments, the permeate outlet can be located below the permeate inlet (and at the opposite end of the membrane-based water extraction device) to utilize gravity to facilitate permeate flow out of the membrane-based water extraction device. As noted above, the desiccant flow channel and the permeate flow channel are each bounded on opposite sides by a porous membrane 633a. By separating the desiccant fluid from the permeate stream, the membrane 633a prevents the bulk transport of dissolved solids from the desiccant into the resulting permeate stream, thereby preserving the desiccant for continued use and maintaining the purity of the recovered water. Similarly, the membrane prevents migration of gas (e.g., sweep gas) on the permeate side of the membrane 633a from penetrating into the desiccant stream.

[0082] In certain embodiments, the housing defines two parallel flow paths that contact opposite sides of the porous membrane 633a with a countercross flow, as illustrated in FIG. 2A , where the first flow path is a desiccant flow path 633b and the second flow path is a permeate flow path 633c, and the porous membrane 633a is embodied as a planar membrane (e.g., defined in a frame) that separates the desiccant flow path 633b from the permeate flow path 633c. In certain embodiments, the housing is configured to allow access to the porous membrane for maintenance purposes, such as replacing the membrane as needed. In other embodiments, the first flow path (e.g., the desiccant flow path) may flow horizontally across the surface of the membrane 633a, and the second flow path (e.g., the permeate flow path) may flow vertically across the opposing surface of the membrane, such that the permeate flow outlet is below the permeate flow inlet.

[0083] In other embodiments, the desiccant flow path can be defined by a desiccant inlet that directs the desiccant toward the membrane (e.g., at a perpendicular angle or at least substantially perpendicular to the first side of the membrane), and the desiccant is directed toward an outlet after it contacts the membrane. This embodiment is a dead-end flow, where the inlet desiccant contacts the porous membrane in a direction at least substantially normal (perpendicular) to the permeate flow path on the opposite side of the membrane. In this embodiment, the permeate flow path flows parallel to the porous membrane. An example of this configuration is illustrated in Figure 2B.

[0084] In yet other embodiments, such as illustrated in FIG. 2C , the housing may define luminal flow paths, with the first flow path being at least substantially concentric with the second flow path. The wall of the first flow path (separating the first and second flow paths) may be at least partially defined by a porous membrane. In certain embodiments, the porous membrane may extend partially around the first flow path. In other embodiments, the porous membrane may extend entirely around the first flow path. By way of example only, the first inner flow path may be a desiccant flow path and the second outer flow path may be a permeate flow path, whereby permeate migrates from the inner flow path to the outer flow path through the porous membrane. As another example, the first inner flow path may be a permeate flow path and the second outer flow path may be a desiccant flow path, whereby permeate migrates from the outer flow path to the inner flow path through the porous membrane.

[0085] In luminal flow membrane-based water separation devices, the permeate and desiccant channels can extend in the same cocurrent or countercurrent (opposite flow) direction. In certain embodiments, the flow can be horizontal or vertical. In a vertical orientation, the permeate side can flow downward, thereby utilizing gravity to facilitate the flow of water (after condensation) out of the membrane-based water separation device.

[0086] To achieve separation of water from the desiccant solution, a chemical potential difference is introduced. This can be realized in a temperature gradient across the membrane, a pressure gradient across the membrane, and / or a concentration gradient across the membrane between the desiccant channel 633b, the membrane 633a, and the permeate channel 633c. The temperature gradient across the membrane can be achieved by heating the desiccant fluid channel 633b and / or cooling the permeate liquid water channel 633c. The temperature of the desiccant fluid and / or liquid water can be manipulated through the use of any heating / cooling source, including, but not limited to, heat exchangers, heating elements, waste heat, geothermal heat, solar heating, geothermal cooling, refrigeration, cooling ponds, cooling streams, sweep gas, and / or the like. For example, a membrane-based water extraction device can incorporate a heat exchange arrangement across a non-porous film on the desiccant side of the membrane (e.g., a heating fluid can heat the desiccant fluid as it flows across the surface of the membrane), and / or a cooling arrangement can be incorporated on the permeate side of the membrane (e.g., a cooling fluid separated from the permeate fluid across a non-porous film can cool the permeate fluid). A pressure gradient across the membrane can be achieved by a high pressure on the desiccant side of the membrane and / or a low pressure on the liquid water side of the membrane. The pressure difference between the desiccant side of the membrane and the permeate side of the membrane can be introduced by the following mechanisms, but is not limited to: a high-pressure pump, a pump, a blower, a compressor, a vacuum pump, a Venturi vacuum-inducing mechanism, and / or the like. Additionally, one or more pumps and / or agitators can be utilized on the desiccant side or permeate side to ensure uniformity of the desiccant and / or liquid water properties. Finally, the pressure and / or temperature differential may provide a vapor pressure differential on either side of the membrane to encourage water vapor to migrate through the membrane from the desiccant side to the permeate / vapor side of the membrane. Specifically, the vapor pressure on the permeate side of the membrane may be lower than the vapor pressure on the desiccant side of the membrane (e.g., to drive water permeation through the membrane using an induced vacuum).

[0087] In certain embodiments, the membrane-based water extraction device is configured to drive the chemical potential using the mechanisms described above. In one example incorporating vacuum membrane distillation (VMD), the permeate outlet pathway 651 is pressurized to a vacuum pressure (e.g., utilizing one or more vacuum pumps, such as compressor 652, located downstream of the membrane-based water extraction device 633 along the permeate flow path) to induce a pressure gradient across the membrane 633. By way of example only, the vacuum pump may be located in the permeate storage tank 655 and / or along the gas vent flow path 659 from the water storage tank 655. In this embodiment, the absolute pressure on the desiccant side 633b is greater than the absolute pressure on the permeate side 633c. The interior of the membrane-based water extraction device may be similar to that shown in FIGS. 2A-2C. In certain embodiments, additional components, such as heating and / or cooling fluid streams (as discussed below), may be additionally incorporated into the VMD configuration of the membrane-based water extraction device to further improve the efficiency of the water extraction process.

[0088] Another exemplary embodiment utilizes an air gap membrane distillation (AGMD) process to drive the permeation of water vapor across membrane 633a to permeate side 633c of membrane 633a. Exemplary AGMD configurations are shown in Figures 2D-2E, each of which provides for the permeate side of membrane 633c to carry a cooling fluid along a cooling fluid flow path separated from permeate flow path 633c by a thermally conductive, nonporous film (e.g., cooling water directed from storage tank 655 along flow path 658 to flow path 650 and entering membrane-based water extraction device 633 at permeate side 633c of membrane 633a). Cooling fluid flowing along flow path 633d sandwiches an air gap between nonporous membrane 633e (which separates cooling fluid flow path 633d from the permeate side 633c of the membrane) and porous membrane 633a. The created air gap is formed by nonporous membrane 633e, which promotes condensation via heat transfer. The void serves as a pathway for the condensate as it separates from the desiccant through the porous membrane. The cooling fluid is at a lower temperature than the heated desiccant fluid, thereby creating a temperature gradient that begins at the heated desiccant side of module 633b and ends at the low-temperature, non-porous membrane that holds the cooling fluid. The cooling side also condenses the transported vapor to a liquid state. A higher temperature gradient (greater difference between the heated desiccant fluid and the cooling fluid) creates a greater driving force for separating the water molecules within the condensed desiccant across the porous membrane, thus achieving higher separation performance for the membrane-based water extraction device. As discussed above, the water vapor within the void may be directed out of the housing of the membrane-based water extraction device based at least in part on a vacuum created within the void, at least in part on a sweep gas flowing through the void, and / or at least in part on gravity, which forces the condensed water to flow downward to and out of the outlet located at the bottom end of the void. In one particular embodiment, the housing defines three outlets: a desiccant outlet (to flow path 634), a permeate outlet (to flow path 651), and a cooling fluid outlet (for recirculating the cooling fluid along the cooling fluid flow path).In certain embodiments, the housing defines at least two inlets (if no inlet is required for the permeate fluid, such as when gravity is used to direct the permeate fluid out of the void), including a desiccant inlet (from flow channel 632) and a cooling fluid inlet (e.g., from flow channel 650). In other embodiments, the housing defines at least three inlets (if a permeate inlet is required, such as when vacuum pressure or a sweep gas is used to direct the permeate out of the housing), including a desiccant fluid inlet (from flow channel 632), a permeate fluid inlet (e.g., to allow gas flow through the permeate side 633c of the membrane), and a cooling fluid inlet (e.g., from flow channel 650).

[0089] The permeate side of membrane 633a is illustrated as a flat, planar, non-porous membrane 633d, but it should be understood that it may be embodied as a shell-and-tube heat exchanger configuration (where the walls of the tubes embody non-porous film 633e and the cooling fluid flows within the interior of the tubes).

[0090] Moreover, certain embodiments may provide a similar configuration on the desiccant side of the membrane by utilizing a heating fluid separated from the desiccant by a nonporous, thermally conductive film, as shown in FIG. 2E. In such embodiments, a heating fluid (e.g., heating oil, steam, heated refrigerant, and / or the like) passes through the nonporous film 633g, opposite the desiccant fluid 633b, along a flow path through the heating fluid stream 633f. The heating fluid transfers heat to the desiccant flowing through the desiccant fluid 633b, which promotes water vapor transfer across the porous membrane 633a. In particular, when the heating fluid stream 633f configuration is incorporated together with the cooling fluid stream 633d configuration, a large chemical potential difference is introduced between the desiccant fluid 633b and the permeate stream 633c, promoting water migration across the membrane 633a. As discussed above, water vapor within the voids can be directed out of the housing of the membrane-based water extraction device based at least in part on a vacuum created within the voids, on at least in part on a sweep gas flowing through the voids, and / or on at least in part on gravity, which forces the condensed water to flow downward to and out of an outlet located at the bottom end of the voids. In certain embodiments, the housing defines three outlets: a desiccant outlet (to flow path 634), a permeate outlet (to flow path 651), a heating fluid outlet (e.g., for recirculating the heating fluid along the heating fluid flow path), and a cooling fluid outlet (for recirculating the cooling fluid along the cooling fluid flow path). In certain embodiments, the housing defines at least three inlets (if an inlet is not required for the permeate fluid, such as when gravity is used to direct the permeate fluid out of the voids). In other embodiments, the housing defines at least four inlets (if a permeate inlet is required, such as when vacuum pressure or a sweep gas is used to direct permeate out of the housing), including a desiccant fluid inlet (from flow channel 632), a permeate fluid inlet (e.g., to allow gas flow through the permeate side 633c of the membrane), a heating fluid inlet, and a cooling fluid inlet (e.g., from flow channel 650).

[0091] As illustrated in any of Figures 2D-2E, separating the permeate stream 633c from the cooling fluid 633d can increase the efficiency of water permeation through the membrane due to the low vapor pressure on the permeate side 633c of the membrane. Vacuum pressure, sweep gas, and / or other flow-enhancing mechanisms may be implemented in the permeate stream 633c to encourage the water vapor on the permeate side 633c of the membrane to exit through the permeate outlet of the membrane-based water extraction device 633 and be directed to one or more compressors (to incorporate the beneficial features of mechanical vapor compression, similar to that described in co-pending U.S. patent application Ser. No. 17 / 552,173, filed December 15, 2021, the contents of which are incorporated herein by reference in their entirety). The water vapor exiting the permeate outlet of the membrane-based water extraction device may flow to one or more heat exchangers and / or one or more condensers 627 to condense the water vapor into liquid water before being stored in the storage tank 655.

[0092] Another embodiment implements a direct contact membrane distillation (DCMD) process to drive the chemical potential across the membrane 633a. In a DCMD configuration, the permeate side 633c of the membrane 633a carries a cooling fluid (e.g., cooling water) that directly contacts the porous membrane 633a. In such a configuration, water is directed directly to the permeate stream 633c from a storage tank 655 to maintain a low temperature on the permeate side 633c of the membrane 633a. Because the cooling fluid is cooler than the desiccant fluid, the cooling fluid creates a temperature gradient across the membrane 633a, starting at the heated desiccant side 633b and ending with the cooling fluid in direct contact with the permeate side 633c of the membrane 633a. The cooling fluid can be removed from the storage tank 655 and directed into the membrane-based water extraction device 633 along flow paths 658 through 650 and 633c (and out of the membrane-based water extraction device via flow path 651). In certain embodiments, a portion of the water exiting storage tank 655 along flow path 658 can be directed to an external system for use as liquid water. A cooling fluid (e.g., liquid water from storage tank 655) is fed to permeate side 633c of membrane 633a, which condenses the water vapor that permeated through membrane 633a and transports the separated water vapor away from the concentrated desiccant. A higher temperature gradient (greater difference between the heated desiccant fluid and the cooling fluid) creates a greater driving force for separating water molecules within the concentrated desiccant across the porous membrane, thus achieving higher separation performance of the membrane-based water extraction device.

[0093] As yet another example, a sweep gas membrane distillation (SGMD) process can be utilized to create a chemical potential across the membrane 633a. SGMD configurations can be combined with AGMD and / or DCMD to further drive the efficiency of such configurations. According to an SGMD configuration, the permeate side 633c of the membrane 633a carries a sweep gas (e.g., nitrogen gas, humid air, an inert gas, and / or the like) that transports the permeate fluid (e.g., water vapor) away from the porous membrane 633a after permeating through the porous membrane. In certain embodiments, the sweep gas can carry the permeate water vapor to the compressor 652 and / or the condenser 627, where the water vapor condenses to liquid water. The sweep gas can be separated from the permeate water vapor by bringing the mixture to the dew point of water, and finally, the sweep gas can be directed away from the liquid water along flow path 659 and exit the storage tank 655. The sweep gas can be directed back to the storage tank and recycled into the system along flow path 660, as discussed above. Condensing the permeate water vapor to a liquid allows the water to be stored in product water tank 655. In certain embodiments, SGMD configurations utilize an inert gas-state fluid as a sweep gas that does not contaminate the product water. In certain SGMD embodiments, the gas flow to permeate flow path 633c is heated to maintain the physical state (vapor state) of the permeate. To help maintain the vapor state, a heat tracing element can be utilized to prevent the fluid pipe walls from cooling the fluid to a liquid state. The water vapor can then be transported to condenser 627 downstream of membrane-based water extraction device 633 to condense the water, which is then stored in water storage tank 655. Additionally, in SGMD embodiments, the sweep gas can be introduced into the membrane-based water extraction device along flow paths 660 and 650. The sweep gas can be directed away from the liquid water after condensation by directing the sweep gas out of the water storage tank, such as along flow path 659, as shown in FIG. 1B.

[0094] In certain embodiments, the water extraction module includes one or more heating mechanisms along the flow path that directs the desiccant fluid to the inlet of the membrane-based water extraction device (on the desiccant side 633b of the membrane 633a). In one exemplary embodiment, the desiccant fluid is heated to a temperature of approximately 40-80°C. This process can result in sufficient separation even at low temperatures (approximately 40°C). In certain embodiments, separation efficiency can be increased at higher temperatures (e.g., approximately 60-80°C).

[0095] For example, the diluted desiccant leaving the absorber module passes through one or more heating subsystems between the absorber module and the membrane-based water extraction device. These heating subsystems are provided as part of the water extraction module. The one or more heating subsystems may include one or more of a condenser 627, a pre-extraction heat exchanger 629, and / or a heater 631. It should be understood that the one or more heating subsystems may be provided in any order relative to the desiccant flow. In one example, the diluted desiccant remains cool after passing through the absorber 610 and passes through a condenser 627 that utilizes the generally cool temperature of the diluted desiccant to promote condensation of water vapor from the water vapor flowing along the permeate flow paths 651-654 (e.g., the water flow paths encompassing the water flow paths on the second side of the porous membrane). In one particular embodiment, the condenser 627 is a shell-and-tube heat exchanger, where the diluted desiccant (upstream of the membrane-based water extraction device 633) passes through the tubes and the water vapor condenses on the exterior of the tubes within the shell of the heat exchanger. In another embodiment, the condenser 627 is a plate-and-frame heat exchanger, where the diluted desiccant fluid passes through one set of plates and the water vapor passes through the other set of plates, condensing within the heat exchanger as it heats the diluted desiccant as it travels through the heat exchanger. In another embodiment, the condenser 627 is a dual-pipe heat exchanger, where the diluted desiccant passes through an inner pipe and the water vapor passes through an outer pipe, allowing the water vapor to condense on the outer surface of the inner pipe. In certain embodiments, the condenser 627 can have a counter-flow configuration (diluted desiccant flowing in the opposite direction to the water vapor). In other embodiments, the condenser 627 can have a parallel, co-flow configuration, where the diluted desiccant and water vapor flow in the same direction through the condenser 627.

[0096] The diluted desiccant exiting the condenser 627 via the flow path represented by 628 increases in temperature due to a certain amount of heat transferred from the water vapor to the diluted desiccant in the condenser 627. The diluted desiccant then passes through a pre-evaporator heat exchanger 629 (e.g., a shell-and-tube heat exchanger, a plate-and-frame heat exchanger, a dual-tube heat exchanger (with concentric tubes), and / or the like) and / or a heater 631 (e.g., an externally powered heater such as an electric heater, a natural gas heater, a solar heater, and / or the like) to raise the temperature of the diluted desiccant to near the evaporation temperature. In certain embodiments, the heater may be an in-line electric heater with a bundle of heating elements to heat the diluted desiccant fluid. The heater may have an orientation that reduces the possibility of fluid spurts over the elements. The heater 631 may be positioned within the housing of the membrane-based water extraction device. In certain embodiments, the heater is a heat exchanger (e.g., a shell-and-tube heat exchanger, a plate-and-frame heat exchanger, and / or the like). In certain embodiments, heater 631 is a solar heater that utilizes photovoltaic panels to generate electrical energy to drive an electric heater element (e.g., a resistive heater element). In certain embodiments, heater 631 is a Fresnel lens heater that utilizes solar energy to generate thermal energy in the form of heat. In certain embodiments, heater 631 comprises a geothermal heater arrangement comprising a series of pipes extending into the earth and utilizing geothermal energy to heat the diluted desiccant fluid. In other embodiments, heater 631 is a thermal heater that utilizes a hydrocarbon fuel source (e.g., natural gas, oil, wood, biomass, and / or the like) combined with oxygen (supplied from ambient air) to create heat from combustion.

[0097] In embodiments including both a pre-extraction heat exchanger 629 and a heater 631, the dilute desiccant fluid first exits the pre-extraction heat exchanger 629 via a flow path designated 630 before entering the heater 631. Additionally, as discussed in more detail herein, the opposite side of the pre-extraction heat exchanger 629 is provided with heated concentrated desiccant fluid exiting the membrane-based water extraction device.

[0098] As the desiccant exits the heater 631 (if used), it flows along flow path 632 into the housing, specifically the desiccant inlet of the membrane-based water extraction device. In a membrane-based water extraction device, the desiccant solution is in direct contact with the membrane. Using temperature, pressure, and / or concentration differences, water migrates through the membrane and is collected on the second side. The water may migrate through the membrane as a liquid and / or vapor. The water may then be distributed for use and / or collected in a storage tank 655. In certain embodiments, the vapor on the liquid water side of the membrane may be collected in cooled liquid water or another cooling fluid to condense the vapor. The vapor on the liquid water side of the membrane can be swept away from the membrane using a sweep gas flowing through the membrane-based water extraction device on the liquid water side of the membrane. The vapor may then be condensed in a further process using a heat transfer process, such as a heat exchanger or other embodiment of a heat transfer process, to condense the water vapor into liquid water (the opposite side of the heater exchanger and / or condenser may define a portion of the flow path for a cooling fluid, such as a dilute desiccant, before entering the desiccant side of the membrane-based water extraction device). In certain embodiments, after exiting the membrane-based extraction device 633, the vapor enters the heat transfer process directly on the permeate side of the membrane. The heat transfer process may be a heat exchanger 629 or other embodiment of a heat transfer process to condense the water vapor into liquid water. In other embodiments, a vacuum may be induced on the permeate side of the membrane to achieve a pressure differential between the desiccant and permeate sides of the membrane. Water enters the membrane as a liquid and / or vapor and exits the membrane as vapor on the permeate side. The water flow path directs water (via flow paths 651-654, including pumps, heat exchanger 627, compressor 652, and / or other flow assist devices) into storage tank 655. The water flow path also circulates from the storage tank via flow path 650 (not shown, but which in certain embodiments includes one or more pumps) to the membrane-based water extraction device 633. In certain embodiments, a sweep gas (e.g., humid air) is blown through the water flow path (e.g., via a blower in-line with the water flow path), pushing water vapor from the membrane-based extraction device 633 into the compressor 652 and ultimately into the condenser 627.

[0099] In certain embodiments, the permeate side vapor is compressed by mechanical and / or thermal means to a higher pressure (e.g., via compressor 652). The vapor then passes through heat exchanger 627, allowing the latent heat to be used to heat the desiccant solution. Through this heat exchanger, the vapor is also condensed into a liquid.

[0100] The membrane-based water extraction device 633 may be configured for batch operation, in which the desiccant circulates in a closed loop, repeatedly contacting the desiccant fluid with the membrane (without directing the desiccant to the absorption module) until a desired amount of water is separated. The valves in the water extraction module may be configured to provide a closed-loop flow of the desiccant solution. After the desiccant solution reaches a desired concentration, the desiccant solution is then sent back to the absorption module by reconfiguring the valves to allow desiccant flow from the water extraction module to the absorption module. The membrane-based water extraction device 633 may include a system of parallel subunits, in which multiple membranes are contained in a single apparatus and / or multiple membrane-based water extraction devices operate in parallel (each device operates to separate water from a portion of the desiccant solution). In certain embodiments, the membrane separation subunits may also be configured in a series configuration. In this configuration, the desiccant solution contacts a first membrane, and a certain amount of water migrates through the first membrane. The residual desiccant solution from the first membrane is fed to the second membrane as a desiccant solution, where an additional amount of water is separated from the desiccant solution. This process is repeated for the number of membrane separation subunits in the series configuration. In certain embodiments, a combination of parallel and series units may be utilized.

[0101] In certain embodiments, the membrane-based water extraction device 633 is configured for continuous operation, where a stream of desiccant flows in direct contact with the membrane (e.g., perpendicular, parallel, or tangential to the membrane plane) and water is continuously separated from the desiccant solution through the membrane. Continuous mode operation may utilize series and / or parallel subunits as described in the sections above. During continuous operation, the desiccant flows in a continuous loop from the absorption module to the water extraction module and back to the absorption module.

[0102] In certain embodiments, the membrane-based water extraction device may be provided in combination with an evaporation-based water extraction device (e.g., an evaporation vessel for evaporating water from a desiccant and a condenser for condensing the evaporated water into potable liquid water). For example, the membrane-based water extraction device may be provided upstream (along the desiccant flow path) from an evaporation vessel, such as the evaporation vessel described in co-pending U.S. patent application Ser. No. 17 / 552,173, filed December 15, 2021, the contents of which are incorporated herein by reference in their entirety. Alternatively (or additionally), the membrane-based water extraction device may be positioned downstream (along the desiccant flow path) of the evaporation vessel, as discussed above.

[0103] Carbon dioxide capture Treated air (which may include air exiting a water consolidation system, as discussed herein) may be passed through a carbon dioxide capture system before being discharged to the atmosphere. Carbon dioxide may be captured from the air for filtration and / or disposal (e.g., through one or more chemical processes to convert the carbon dioxide into water, oxygen, and / or solid or liquid compositions that can be disposed of, through capture of the carbon dioxide in a filtration medium, and / or the like).

[0104] The carbon dioxide capture system may include a carbon dioxide capture column having a fixed bed of carbon dioxide absorbent material (e.g., sodium hydroxide solution). As air passes over the carbon dioxide absorbent material, carbon dioxide is absorbed by the material. Additionally, as shown in Figure 1, the carbon dioxide capture column may be heated (e.g., with a hot fluid jacket) to facilitate increased carbon dioxide absorption by the absorbent material.

[0105] As yet another example, the carbon dioxide capture material may be configured to reversibly absorb carbon dioxide such that the captured carbon dioxide can be compressed and stored as a gas for later use.

[0106] In certain embodiments, the recovered carbon dioxide gas may be directed to a greenhouse to optimize the interior greenhouse environment for plant growth. As discussed herein, the greenhouse may be supplied with water produced by the AWG system discussed herein.

[0107] solar canopy The AWG system may be associated with a greenhouse or other agricultural system for facilitating plant growth (e.g., growing consumable plants). The greenhouse or other agricultural system for facilitating plant growth may include one or more features for optimizing conditions within the greenhouse to promote plant growth. As an example, a greenhouse may incorporate one or more features that enable sunlight control by varying the direction, intensity, and color of light to facilitate achieving ideal plant growth conditions.

[0108] In various embodiments, a solar canopy may be provided as a surface covering for a plant growth habitat (e.g., a greenhouse), as discussed in more detail herein. The solar canopy may be configured to be secured to at least a portion of the support structure of the plant growth habitat. The support structure of the plant growth habitat may include a plurality of frame members connected together to form a support frame of a desired shape for receiving the solar canopy and / or other materials (e.g., structural panels, and / or the like) of the plant growth habitat. As a specific example, as shown in FIG. 4 , the support structure may include a plurality of vertical frame members 20A (e.g., beams, columns, and / or the like), horizontal frame members 20B, and diagonal frame members 20C, which are connected together and spaced apart to form a support frame 20 that defines the perimeter of the plant growth habitat (e.g., a greenhouse) and the interior of the plant growth habitat. The frame 20 may be made from one or more of a variety of materials (e.g., metal, wood, plastic, and / or any suitable material). In some embodiments, the frame 20 may be a rigid frame. In some embodiments, frame 20 may be flexible. In some embodiments, frame 20 may be collapsible (e.g., foldable).

[0109] 6 , multiple solar canopies 100 may be integrated or otherwise attached together to define a portion of a surface covering or an entire surface covering, in which case the opacity of each solar canopy 100 may be independently controllable such that the opacity across different areas of a plant growth habitat (e.g., a greenhouse) having a solar canopy covering may be independently varied. In certain embodiments, the opacity of each solar canopy 100 of a surface covering may be independently controllable via one or more environmental control systems 1004. In certain embodiments, the one or more environmental control systems 1004 may be in electronic communication with a voltage source / power source (e.g., a battery, an electric field generator, and / or the like), in which case the one or more environmental control systems 1004 may be configured to independently control the application of an external signal (e.g., an electric field signal) from the voltage source / power source to the liquid crystal fluid of the solar canopy 100. In certain embodiments, each solar canopy 100 may be associated with an identifier, where each identifier may correspond to a location in the greenhouse such that the environmental control system 1004 may independently control each solar canopy 100 (and thus a particular area of the greenhouse) based at least in part on the identifier.

[0110] In various embodiments, the solar canopy 100 may be utilized in a plant growth habitat (e.g., a greenhouse) to control the amount or intensity and directionality of light to different areas of the plant growth habitat. As an example, for optimal plant growth, different plants in a greenhouse may require different light intensities and directionality at different times. As shown in FIG. 7, a greenhouse utilizing a solar canopy 100 as discussed herein may control the opacity of the solar canopy 100 so that different areas have different opacity levels.

[0111] In various embodiments, a method of using the solar canopy 100 may include determining an identifier associated with the solar canopy 100 or an area of the solar canopy 100 (and thus a particular area of the greenhouse) and applying an appropriate signal (e.g., an electric field) to the corresponding solar canopy 100 and / or solar canopy area to change the opacity of the corresponding greenhouse area.

[0112] Automated Usage The AWG system may be utilized to generate water and / or power supplied to an agriculture module, which may include a greenhouse, plant growth habitat, and / or other structure that may be utilized to promote plant growth within controlled atmospheric conditions. FIGS. 4-5 illustrate various embodiments of an agriculture module 1000 associated with an AWG system 110 housed within a shipping container, according to one embodiment. As shown in the figures, the agriculture module 1000 may define a plant growth habitat having at least a substantially rectangular shape, or a shape with multiple distinct lobes (e.g., forming a star shape as shown in FIG. 5). In embodiments including distinct lobes, the volume within each lobe may be separated from the remainder of the growth habitat such that each lobe may be provided with a unique growth environment (e.g., different temperatures, carbon dioxide levels, humidity levels, and / or the like) to foster the growth of different agricultural products.

[0113] Exemplary Automated Usage The AWG system may be integrated with a greenhouse or any other water-consuming environment, including, by way of example, a biosphere environment, a residential structure, a recreational facility, and / or the like. By way of example, the AWG system may be utilized to generate water and / or power supplied to an agriculture module, which may include a greenhouse, a plant growth habitat, and / or other structure that may be utilized to promote plant growth within controlled atmospheric conditions. FIGS. 4-5 illustrate various embodiments of an agriculture module 1000 associated with an AWG system 110 housed within a shipping container, according to one embodiment. As shown in the figures, the agriculture module 1000 may define a plant growth habitat having at least a substantially rectangular shape, or a shape with multiple distinct lobes (e.g., forming a star shape as shown in FIG. 5). In embodiments including distinct lobes, the volume within each lobe may be separated from the remainder of the growth habitat such that each lobe may be provided with a unique growth environment (e.g., different temperatures, carbon dioxide levels, humidity levels, and / or the like) to foster the growth of different agricultural products.

[0114] FIG. 6 shows a schematic detailed view of a portion of a growing habitat of an agricultural module 1000, according to one embodiment. The growing habitat of the agricultural module 1000 may include one or more stackable structures 1001, each having one or more base portions 1002 configured to support a growing medium (e.g., soil, hydroponic supports, and / or the like), one or more sidewalls, and a ceiling. The stackable structures 1001 may be suspended from a support frame of the growing habitat, stacked such that the supports of the upper structure are supported by the lower structure, and / or the like. The one or more sidewalls and ceiling are configured to contain controlled atmospheric conditions within the structure (e.g., ambient air with controlled oxygen and carbon dioxide levels, controlled temperature, controlled humidity, and / or the like). The one or more sidewalls and ceiling may include a covering material, such as a flexible covering material, a rigid covering material, and / or the like. In certain embodiments, the covering material may include integrated grow lamps (e.g., light-emitting diode grow lamps) and / or integrated electrical circuitry and / or may be configured to allow natural sunlight to pass through the covering material into the enclosed environment. In certain embodiments, the integrated grow lamps may be spaced at regular intervals throughout the flexible covering material and may be electrically connected to each other and / or to one or more power sources via electrical circuitry. For example, in the illustrated embodiment of FIG. 6, the covering material comprises a solar canopy (e.g., solar canopy 100) as discussed herein, with integrated LEDs 12 spaced across the surface of the solar canopy.

[0115] In certain embodiments, the LEDs 12 may be directed through the inside of the solar canopy 100, opposite the outside of the solar canopy 100, such that the LEDs 12 emit light through the inside of the solar canopy 100 (e.g., through a second flexible protective sheet, such as to the interior of the greenhouse in an embodiment in which the solar canopy surrounds a greenhouse). In certain embodiments, the LEDs 12 may be aligned with one or more photovoltaic elements and configured to emit light toward the back side of the photovoltaic elements, such that the light reflects off the back side of the photovoltaic elements and reflects through the inside of the solar canopy 100. The LEDs 12 may additionally be connected to one or more conductors (which may be in series with the photovoltaic elements, in parallel with the photovoltaic elements, or provided in a separate circuit from the photovoltaic elements). In certain embodiments, the LEDs 12 may receive power from multiple photovoltaic elements. Additionally or alternatively, in certain embodiments, various LEDs 12 may be suspended from the solar canopy 100 (e.g., inside the greenhouse). For example, LEDs may be suspended within a growing habitat of an agricultural module 1000 (as shown in Figures 4-6) surrounded by one or more solar canopies 100, so that the LEDs provide additional light to plants growing therein from additional angles (e.g., in proximity to the growing medium in which the plants are growing).

[0116] In embodiments including a flexible cover material, the agriculture module 1000 may include one or more rigid supports that collectively form a rigid support frame for the flexible cover material.

[0117] In certain embodiments, the agriculture module 1000 may be embodied as a portable system configured to be quickly set up at a desired agricultural site. The agriculture module 1000 may additionally include one or more sensors 1003 that may be provided within the growing medium of the growing habitat. These sensors may be embodied as part of a flexible bundle of electrical circuits including conductors, sensors, and / or the like that may be quickly deployed within the growing habitat by deploying the bundle on a supporting surface of the growing habitat prior to providing the growing medium therein. In certain embodiments, the various sensors may be electrically connected to each other, to the environmental control system 1004, and / or to a power source via one or more conductors (e.g., flexible conductors). The various sensors may include a moisture sensor, a temperature sensor, a carbon dioxide content sensor, an oxygen sensor, a humidity sensor, and / or the like. It should be understood that certain of the described sensors may be configured for wireless data transmission to a controlling computing system via one or more wireless communication technologies, such as Wi-Fi, Bluetooth, Internet of Things (IoT) technologies, and / or the like.

[0118] In certain embodiments, sensor outputs (e.g., indicative of measured aspects of the environment within the growth habitat) may be utilized by environmental control system 1004 (and / or the control systems shown in FIGS. 8-12) to regulate environmental conditions within the growth habitat. For example, environmental control system 1004 may include data indicative of one or more target environmental conditions, such as a target temperature, a target carbon dioxide content, and / or the like. Based at least in part on monitored data output from various sensors 1003 within the growth habitat, environmental control system 1004 is configured to compare the monitored data output to the target environmental conditions and may be configured to adjust the flow of water, the flow of carbon dioxide, and / or the like from AWG system 110 to the growth habitat. For example, the environmental control system 1004 may be configured to automatically activate a sprinkler (or drip irrigation) system (which may be incorporated into the stackable structure 1001) within the growth habitat to water plants within the growth habitat in response to predetermined conditions, increase and / or decrease the amount of carbon dioxide flowing into the growth habitat from the carbon dioxide capture system of the AWG system 110, and / or the like.

[0119] Additionally, the growing habitat may include one or more automated planting and harvesting mechanisms configured to autonomously plant seeds for new plants and / or automatically harvest fruits and / or vegetables grown within the growing habitat (this includes the use of agricultural robots and drones).

[0120] For example, seed planting / management may be provided via a planting probe 1010 operable to move along a grid / track system 1011 elevated above a support surface of the growing habitat. In certain embodiments, the grid / track system 1011 may be raised and / or lowered via a support mechanism (e.g., a pneumatic and / or hydraulic support mechanism). The planting probe 1010 may be operable in response to signals received from an environmental control system 1004 including data indicative of an internal mapping of the planting bed and / or base portion 1002 within the growing habitat. The environmental control system 1004 may additionally include data indicative of desired crops for planting within the growing habitat, crop spacing, and / or the like, and may provide movement signals to the planting probe 1010 to insert seeds into the planting bed according to the desired planting plan.

[0121] The planting probe 1010 itself may include a hopper 1012 configured to hold a quantity of seeds and an insertion probe 1013 (e.g., a wedge-shaped insertion probe) configured to inject the seeds at an appropriate depth within the planting cultivation area (as determined by the environmental control system 1004). The planting probe 1010 additionally includes a movement mechanism (e.g., one or more motors) configured to move the planting probe 1010 along the track / grid to plant the seeds within the planting cultivation area. Moreover, the planting probe 1010 may be configured to periodically return to a refilling position within the growing habitat to retrieve additional seeds within the contained hopper 1012. The refilling position may be positioned within the growing habitat in proximity to a filling chute containing additional seeds that may be selectively provided to the planting probe 1010 as needed. In certain embodiments, the filling chute may be embodied as a container supported (e.g., suspended) above the planting probe's path of travel such that the planting probe 1010 may move below the filling chute to be refilled by gravity, which moves seeds from the filling chute into the planting probe 1010. Moreover, in certain embodiments, the filling chute may include an actuatable feed door (e.g., a servo-actuated feed door) configured to open in response to a signal received from the environmental control system 1004 and allow seed flow from the feed chute. Thus, when the planting probe 1010 is positioned below the feed chute, the environmental control system 1004 may be configured to open the feed door and allow seeds to flow from the feed chute to the planting probe 1010. Once an appropriate amount of seeds has been provided to the planting probe 1010, the environmental control system 1004 may send a second signal to close the feed door.

[0122] The planting probe 1010 may additionally include a harvesting mechanism that may be removably secured to the moveable planting probe 1010. The harvesting mechanism may include a mechanically moveable cutting / picking arm 1014 and a holding basket / tray 1015. When the planting probe 1010 receives a signal from the environmental control system 1004 to initiate the harvesting process, the planting probe 1010 may process the harvested produce / plants by harvesting and / or cutting them from various plants within the growing habitat and depositing the cut produce / plants into the holding basket / tray 1015. When the holding basket / tray 1015 is full, the planting probe 1010 may return to a docking position, where the holding basket / tray 1015 may deposit the harvested items into a holding crate from which they can be removed from the growing habitat. Moreover, in certain embodiments, the holding crate may include one or more level sensors configured to monitor the amount of harvested items in the holding crate to avoid overflowing the holding crate. Upon detecting that the holding crate fill level exceeds a threshold level, the environmental control system 1004 may be configured to send a signal to the planting probe 1010 to suspend harvesting operations until the holding crate is emptied.

[0123] While described above with reference to a track-based planting and harvesting probe configuration, various embodiments may be configured to plant seeds and / or harvest produce via an unmanned aerial vehicle (UAV) equipped with a planting probe and / or a harvesting probe having a configuration similar to that described above. The UAV may be autonomous and configured to navigate within a growing habitat according to a defined planting plan. In certain embodiments, the planting plan may define a map of intended seed planting locations, whereby the autonomous UAV may be configured to autonomously navigate between multiple intended seed planting locations and deposit seeds within a growing medium.

[0124] The autonomous UAV may additionally comprise a harvest probe configuration similar to that described herein. A UAV with a harvest probe configuration may be configured to autonomously navigate within a growing habitat to harvest agricultural products grown therein.

[0125] The grow habitat irrigation system may be embodied as one or more tubes that may be connected to a water distribution mechanism, such as a spray sprinkler, drip irrigation tubing, and / or the like. The tubes may include flexible tubing made of plastic and may be embodied as a self-healing material configured to self-seal cracks, cuts, and / or punctures through the tube wall. These tubes may be connected to the condensation system drain of the AWG system, the water holding tank of the AWG system, and / or the like.

[0126] Additionally, the irrigation system may include a fertilizer supply mechanism configured to automatically mix a measured amount of fertilizer (e.g., liquid fertilizer) into the water supplied to the irrigation system. The fertilizer supply mechanism may be in electronic communication with the environmental control system 1004, which may be configured to provide signals to the fertilizer supply mechanism to modify the amount of liquid fertilizer introduced into the water stream.

[0127] Exemplary Operation of an Atmosphere Generating System The following description provides an exemplary operation of an embodiment utilizing a membrane-based water extraction process as part of an atmospheric water generation system and method. The membrane-based water extraction process is one of several exemplary water extraction processes. Other exemplary water extraction processes may include, for example, a membraneless water extraction process (such as described in co-pending U.S. patent application Ser. No. 18 / 179,750, filed March 7, 2023, the contents of which are incorporated herein by reference in their entirety), an electrostatic nucleation water extraction process (such as described in co-pending U.S. patent application Ser. No. 18 / 199,427, filed May 19, 2023, the contents of which are incorporated herein by reference in their entirety), a photosensitive adsorbent material water extraction process (such as described in co-pending U.S. patent application Ser. No. 18 / 538,832, filed December 13, 2023, the contents of which are incorporated herein by reference in their entirety), and / or the like.

[0128] It should be noted that recitation of the term "about" with reference to a numerical value (e.g., a temperature value, a pressure value, and / or the like) encompasses both the value itself and deviations from that value that provide the same functional operation of the methodology for which the value is recited. For example, recitation of a temperature range of "about" 10°F to "about" 50°F is intended to encompass a temperature range between 10°F and 50°F, as well as slight deviations at the upper and lower ends of the temperature range that provide the same functionality as the recited temperature range. Moreover, the concepts discussed herein may be utilized with any of a variety of desiccant fluids, gels, aerogels, and / or the like, as the described processes are independent of the desiccant utilized. Any one of several desiccant fluids may be utilized, including, but not limited to, CaCl, NaCl, LiCl, KCOOH, MgCl, ionic liquids, deep eutectic solvents, organic liquids, and / or any combination thereof.

[0129] In an exemplary operation, the desiccant fluid exits the absorber at a concentration of about 10% by weight to about 50% by weight and a temperature of about 75°F to about 130°F. The desiccant fluid flows (optionally using one or more pumps) along a flow path through one or more heating mechanisms, such as a condenser (which, as discussed in more detail herein, utilizes the relatively cool temperature of the diluted desiccant on a first side of the condenser to condense water vapor flowing in the water flow path, which flows on the opposite side of the heat exchanger embodied as a condenser). The water vapor condenses in the condenser as it warms the diluted desiccant fluid as it traverses through the condenser. In another embodiment, the condenser is a dual-pipe heat exchanger in which the diluted desiccant fluid enters the dual-pipe heat exchanger through an inner pipe. Water vapor enters the heat exchanger through the outer pipe. The water vapor condenses on the outer surface of the inner pipe as it traverses through the condenser.

[0130] In certain embodiments, the flow pattern through the condenser is a counter-current orientation, with the dilute desiccant solution flowing in a first direction and the water flow path flowing in an opposite second direction through the heat exchanger (as noted, the dilute desiccant fluid is physically separated from the water flow path within the condenser.) In other embodiments, the flow pattern through the condenser is a co-current orientation, depending on the shape and orientation of the heat exchanger embodying the condenser.

[0131] Upon exiting the condenser, the diluted desiccant fluid temperature is between about 132°F and about 170°F. The diluted desiccant fluid then passes to a second heat exchanger 629, where sensible heat transfer from the concentrated desiccant fluid exiting the membrane-based water extraction device further heats the diluted desiccant fluid. In one particular embodiment, the second heat exchanger is a shell-and-tube heat exchanger, with the diluted desiccant fluid entering the tube side of the second heat exchanger. The concentrated desiccant fluid enters the shell side of the second heat exchanger. In another embodiment, the diluted desiccant fluid enters the shell side of the second heat exchanger, and the concentrated desiccant fluid enters the tube side of the second heat exchanger. In another embodiment, the second heat exchanger is a plate-and-frame heat exchanger, with the diluted desiccant fluid entering one set of plates and the concentrated desiccant fluid entering the other set of plates. In another embodiment, the second heat exchanger is a dual-pipe heat exchanger in which the dilute desiccant fluid enters the heat exchanger through an inner pipe and the concentrated desiccant fluid enters the heat exchanger through an outer pipe.

[0132] The diluted desiccant fluid exits the second heat exchanger at a temperature of about 140°F to about 210°F. The diluted desiccant fluid then traverses through a heater. The heater can be an in-line electric heater with a bundle of elements to heat the fluid. The heater may have a different orientation to reduce the possibility of fluid gushing over the elements. In certain embodiments, the heater can reside in the membrane-based water extraction device housing. In certain embodiments, the heater is a shell-and-tube or plate-and-frame type heat exchanger. In certain embodiments, the heater is a solar heater that utilizes photovoltaic panels to generate energy through electricity from sunlight capture. In certain embodiments, the heater is a Fresnel lens that utilizes solar energy to generate thermal energy in the form of heat. In certain embodiments, the heater is a geothermal heater as discussed herein. In certain embodiments, the heater is a thermal heater that utilizes a hydrocarbon fuel source combined with oxygen to create heat from combustion. In certain embodiments, the oxygen is used from ambient air. Upon exiting the heater, the diluted desiccant fluid is at a temperature of about 150°F to about 270°F.

[0133] The diluted desiccant fluid then flows into the membrane-based water extraction device. In certain embodiments, a reduction orifice is provided at the outlet of the membrane-based water extraction device to increase the pressure on the desiccant side of the membrane within the housing of the membrane-based water extraction device.

[0134] The diluted desiccant stream enters the housing of the membrane-based water extraction device on a first side of the membrane. The diluted desiccant stream is at high temperature (e.g., about 150°F to about 270°F) and high pressure. As the water stream flows along the water flow path, it passes through the opposite second side of the membrane. The water flow path can pass through a reduction orifice at the inlet to the membrane-based water extraction device to reduce the pressure on the water flow side of the membrane and encourage water to pass through the membrane to the water flow side of the membrane. Additionally, the water stream passes through a cooling unit before entering the housing of the membrane-based water extraction device.

[0135] As the diluted desiccant and water flow through the housing of the membrane-based water extraction device on opposite sides of the membrane, the water absorbed within the fluid desiccant migrates through the membrane (water is considered to enter the membrane on the desiccant side of the membrane as a liquid and exit the membrane on the water side of the membrane as a vapor). This re-concentrates the desiccant fluid while recovering the water vapor in the permeate flow path for later condensation and storage for use as liquid water. The re-concentrated desiccant fluid then flows back to the absorber by passing through one or more chiller devices to reduce the temperature of the desiccant fluid before entering the absorber (e.g., passing through one or more heat exchangers for sensible heat transfer to the diluted desiccant solution, as discussed above). The overall desiccant flow path is a closed desiccant circulation loop configured for negligible mass transport of desiccant salts through the membrane of the membrane-based water extraction device and / or into the atmosphere that contacts the desiccant stream in the absorber. As the desiccant flows along the flow path in the closed desiccant circulation loop, water is absorbed into and / or extracted from the desiccant as it flows through the AWG system.

[0136] The water flow path exiting the membrane-based water extraction device passes out of the membrane-based water extraction device and through one or more condensers. In certain embodiments, a sweep gas (e.g., a gas with properties that reduce the likelihood of water evaporation into the gas and / or high humidity air flowing along a closed air circulation loop coextensive with the water flow path) flows with the liquid water so that water vapor exits the membrane-based water extraction device and flows to the condenser. In certain embodiments, the water flow passes through a compressor to increase the pressure of the water flow (thereby increasing the vapor pressure of the water vapor and promoting condensation) before passing through one or more condensers. After the water flow path passes through the condenser, the water flow path enters a storage tank. A certain amount of water from the storage tank is recirculated through the water flow path (to the membrane-based water extraction device) to promote the extraction of additional water from the diluted desiccant fluid flowing through the membrane-based water extraction device.

[0137] Example AWG Control System The following description provides exemplary optimized control operations of embodiments of an AWG system optimized for a particular environment. In various embodiments, the AWG system may be configurable based at least in part on the water usage requirements of a particular environment. For example, the AWG system may generate water based at least in part on input data such as demographic information (e.g., age, gender, weight, height, etc.) of the environment's occupants, water usage information (e.g., current and / or historical water usage at different times, etc.), water-dependent facilities (e.g., number of sinks, toilets, showers, sprinklers, etc.), water storage capacity (e.g., storage tank size, etc.), and / or any other data reflecting end-user water requirements. Due to the power requirements of the AWG system, the system may be optimized by selectively powering the system to meet the water requirements for a particular environment using the lowest possible input energy.

[0138] 8 shows a schematic diagram of an exemplary AWG computing ecosystem 800 according to one embodiment. As depicted, the AWG computing ecosystem 800 may include an AWG system 110, a centralized system 816, a user interface 802, an organizational interface 814, and / or one or more information sources 818. In some examples, each of the AWG system 110, the centralized system 816, the user interface 802, the organizational interface 814, and / or the one or more information sources 818 may be communicatively coupled via one or more networks 820.

[0139] In various embodiments, the AWG system 110 may include a membrane-based water extraction system, as described herein. Additionally or alternatively, the AWG system 110 may implement a membrane-less extraction process. In some examples, the AWG system 110 may include electrostatic nucleation, photosensitive adsorbent materials, and / or the like.

[0140] In some embodiments, the AWG system 110 may be integrated with a control system 804 and one or more controllers 812. The control system 804 may include one or more controllers 812 coupled (e.g., via one or more wired and / or wireless connections, etc.) to and / or in communication with (e.g., via one or more wired and / or wireless connections, etc.) one or more subsystems of the AWG system 110, such as, for example, a humidity augmentation system, an air preconditioning system, a power generation module, one or more valves (e.g., electromechanical mixing valves, etc.), motors, actuators, thermal condensers, and / or the like. In some examples, the control system 804 may be and / or include an environmental control system (see FIG. 6 ). For example, the control system 804 may be physically and / or wirelessly integrated with a controlled environment (and / or one or more components thereof), such as a greenhouse (see FIGS. 4-7 ), and / or with the artificial intelligence AWG system 110 within the controlled environment. In some examples, the control system 804 may be physically and / or wirelessly integrated with the AWG system 110 in an at least partially uncontrolled environment (e.g., an apartment complex, a recreational facility, etc.).

[0141] In various embodiments, the control system 804 may be configured to selectively power, run, initiate, schedule, and / or the like, one or more components and / or subsystems of the AWG system 110 through communication with one or more controllers 812. For example, the control system 804 may be configured to selectively control operation of the AWG system 110 (e.g., via one or more control instructions to the one or more controllers 812, etc.) based at least in part on one or more optimization inputs. In some examples, the control system 804 may autonomously operate the AWG system 110 using an optimization module configured to meet the water requirements for a particular environment using the lowest possible input energy.

[0142] Control system 804 may include memory 806 and one or more processors 808 communicatively coupled to memory 806. In some examples, memory 806 may include one or more non-transitory computer-readable storage media containing instructions that, when executed by the one or more processors 808, cause the one or more processors 808 to perform one or more training and / or optimization operations of the present disclosure. For example, the one or more processors 808 may communicate with other elements in control system 804 via a bus.

[0143] In some embodiments, the one or more processors 808 may be embodied as one or more complex programmable logic devices (CPLDs), microprocessors, multi-core processors, co-processing entities, application specific instruction set processors (ASIPs), microcontrollers, and / or the like. Additionally, the one or more processors 808 may be embodied as one or more other processing devices or circuits. The term circuit may refer to an entirely hardware embodiment or a combination of hardware and a computer program product. Thus, the one or more processors 808 may be embodied as an integrated circuit, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a programmable logic array (PLA), a hardware accelerator, other circuitry, and / or the like.

[0144] Thus, it will be understood that one or more processors 808 may be configured for a particular use, or may be configured to execute instructions stored on volatile or non-volatile media, or may otherwise be accessible to one or more processors 808. Thus, whether configured by hardware or a computer program product, or a combination thereof, one or more processors 808, when configured accordingly, may be capable of performing steps or operations according to embodiments of the present disclosure.

[0145] In some embodiments, memory 806 may include non-volatile media (also referred to as non-volatile storage, memory, memory storage, memory circuitry, and / or similar terms used interchangeably herein). In some embodiments, non-volatile media may include one or more non-volatile memories, including but not limited to, hard disks, ROM, PROM, EPROM, EEPROM, flash memory, MMC, SD memory cards, memory sticks, CBRAM, PRAM, FeRAM, NVRAM, MRAM, RRAM, SONOS, FJG RAM, Millipede memory, racetrack memory, and / or the like. As will be appreciated, non-volatile media may store databases, database instances, database management systems, data, applications, programs, program modules, scripts, code (e.g., source code, object code, byte code, compiled code, interpreted code, machine code, etc.) that embody one or more machine learning models or other computer functions, executable instructions, and / or the like, as described herein. The terms database, database instance, database management system, and / or similar terms used interchangeably herein may refer to a collection of records or data stored in a computer-readable storage medium using one or more database models such as a hierarchical database model, a network model, a relational model, an entity-relationship model, an object model, a document model, a semantic model, a graph model, and / or the like.

[0146] In some embodiments, memory 806 may include volatile media (also referred to as volatile storage, memory, memory storage, memory circuitry, and / or similar terms used interchangeably herein). In some embodiments, volatile media may also include one or more volatile memories, including but not limited to RAM, DRAM, SRAM, FPM DRAM, EDO DRAM, SDRAM, DDR SDRAM, DDR2 SDRAM, DDR3 SDRAM, RDRAM, TTRAM, T-RAM, Z-RAM, RIMM, DIMM, SIMM, VRAM, cache memory, register memory, and / or the like. As will be appreciated, volatile storage or memory media may be used to store at least a portion of databases, database instances, database management systems, data, applications, programs, program modules, code (source code, object code, byte code, compiled code, interpreted code, machine code) embodying one or more machine learning models described herein or other computer functions, executable instructions, and / or the like, for execution by one or more processors 808, for example. Thus, code (source code, object code, byte code, compiled code, interpreted code, machine code) embodying a database, a database instance, a database management system, data, an application, a program, a program module, one or more machine learning models, or other computer functions, executable instructions, and / or the like described herein may be used to control certain aspects of the operation of the AWG system 110 with the aid of one or more processors 808, an operating system, and / or one or more controllers 812.

[0147] The one or more controllers 812 may include one or more mechanical, electromechanical, and / or electronic devices that may perform a control action in response to a control signal. By way of example, the controllers 812 may include any combination of actuators, flow controllers, level controllers, programmable logic controllers, switches, motor controllers and drivers, and / or the like. In some examples, each of the controllers 812 may correspond to (e.g., be attached to, etc.) a particular component and / or subsystem of the AWG system 110. Each of the components may be controlled by supplying and / or removing power (e.g., current, etc.) to the respective component, modifying one or more configuration settings, repositioning one or more valves, and / or the like.

[0148] For example, the controller may be equipped and / or configured to control the air preconditioning system (e.g., one or more compressors, centrifugal fans, their electric motors, etc.) by providing a level of power to the air preconditioning system's compressors, centrifugal fans, electric motors, and / or the like. As another example, the same or another controller may be equipped and / or configured to control the water extraction module of the AWG system 110 by providing a level of power, opening and / or closing one or more inlet / outlet valves, and / or the like. The same and / or another controller may be equipped and / or configured to control various other components of the AWG system 110, including the absorption module, blower, carbon dioxide capture system, and / or the like. In those embodiments utilizing a photosensitive material for water vapor absorption / desorption, the same and / or another controller may be configured to control the movement of the photosensitive material and / or to control an illumination device that illuminates the photosensitive material (e.g., by changing the wavelength of light) to change the photosensitive material into an absorption or desorption phase.

[0149] In some embodiments, control system 804 includes one or more network interfaces 810 for communicating with various computing entities (e.g., centralized system 816, user interface 802, organizational interface 814, information source 818, etc.), such as by transmitting, receiving, manipulating, processing, displaying, storing, and / or communicating data, code, content, information, and / or like terms that may be similar and may be used interchangeably herein. For example, control system 804 may communicate with another computing entity to upload or download data or code (e.g., data or code embodying or otherwise associated with one or more machine learning models). Such communications may be performed using one or more networks 820.

[0150] The one or more networks 820 may include a wired data transmission protocol such as Fiber Distributed Data Interface (FDDI), Digital Subscriber Line (DSL), Ethernet, Asynchronous Transfer Mode (ATM), Frame Relay, Data Service Interface Standard over Cable (DOCSIS), or any other wired transmission protocol. Additionally or alternatively, the one or more networks 820 may include a general packet radio service (GPRS), universal mobile telecommunications system (UMTS), code division multiple access 2000 (CDMA2000), CDMA2000, or any other wired transmission protocol. It may include one or more wireless external communications networks using any of a variety of protocols, such as 1X (1xRTT), Wideband Code Division Multiple Access (WCDMA), Global System for Mobile Communications (GSM), Enhanced Data Rates for GSM Evolution (EDGE), Time Division-Synchronous Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), Evolved Universal Terrestrial Radio Access Network (E-UTRAN), Evolution Data Optimized (EVDO), High Speed Packet Access (HSPA), High Speed Downlink Packet Access (HSDPA), IEEE 802.11 (Wi-Fi), Wi-Fi Direct, 802.16 (WiMAX), Ultra Wideband (UWB), Infrared (IR) protocol, Near Field Communication (NFC) protocol, Wibree, Bluetooth protocol, Wireless Universal Serial Bus (USB) protocol, and / or any other wireless protocol.

[0151] In some embodiments, the control system 804 is an edge device of the AWG system 110 configured to communicatively connect the AWG system 110 to the centralized system 816 and / or one or more other components of the AWG computing ecosystem 800. In some examples, the edge device may be physically disposed on the atmospheric water generating system.

[0152] The centralized system 816 may include one or more external computing devices including, for example, memory (including any of the example memories described herein with reference to memory 806), one or more processors (including any of the example processors described herein with reference to processor 808), and / or one or more network interfaces (including any of the example interfaces described herein with reference to network interface 810) external to the AWG system 110. For example, the centralized system 816 may include a connected cloud platform, including a distributed serverless cloud environment, a dedicated cloud server environment, and / or the like. In some examples, the memory of the centralized system 816 may include one or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors of the centralized system 816, cause the one or more processors to perform one or more training and / or optimization operations of the present disclosure.

[0153] In some embodiments, the centralized system 816 is an intermediary computing system between the control system 804 and / or one or more remote devices, such as one or more information sources 818, organizational interface 814, user interface 802, and / or the like. In some examples, the centralized system 816 may host one or more computing services for use by the control system 804. As an example, the centralized system 816 may host one or more portions of an optimization module that may be utilized by the control system 804 to perform one or more optimized control actions. In some examples, the optimization module may include one or more machine learning models that are at least partially trained by the centralized system 816. In some examples, the trained models may be provided to the control system 804 to enable the control system 804 to locally implement the machine learning models and / or locally generate one or more energy-based outputs based at least in part on multiple optimization inputs. In some examples, at least a portion of the optimization inputs may be provided by the centralized system 816.

[0154] Additionally or alternatively, centralized system 816 may be configured to implement the trained machine learning model to remotely generate one or more energy-based outputs on behalf of control system 804. In such cases, control system 804 may provide one or more optimization inputs to centralized system 816 and receive one or more optimized energy outputs from centralized system 816 based at least in part on the optimization inputs.

[0155] In some embodiments, centralized system 816 and / or control system 804 are configured to receive one or more optimization inputs from one or more external information sources 818. Information sources 818 may include, for example, one or more remote data stores configured to receive, manage, store, and / or publish data related to a particular domain. Exemplary information sources 818 may include, for example, one or more weather sources, one or more water usage sources, and / or the like.

[0156] In some embodiments, centralized system 816 and / or control system 804 are configured to receive one or more optimization inputs from organizational interface 814. Organization interface 814 may include a computing interface that may enable access to centralized system 816 through one or more elastic computing interfaces. For example, organizational interface 814 may include a virtual instantiation of centralized system 816, a web portal to centralized system 816, and / or the like.

[0157] In some embodiments, the centralized system 816 and / or the control system 804 are configured to receive one or more optimization inputs from a user interface 802. The user interface 802 may include a computing interface accessible from a user device. The user interface 802 may include a number of interactive widgets that may initiate one or more control commands and / or provide one or more optimization inputs to the centralized system 816 and / or the control system 804.

[0158] In some embodiments, control system 804 may include or communicate with one or more input elements, such as keyboard input, mouse input, touchscreen / display input, motion input, movement input, audio input, pointing device input, joystick input, keypad input, and / or the like. Control system 804 may also include or communicate with one or more output elements (not shown), such as audio output, video output, screen / display output, motion output, movement output, and / or the like.

[0159] Input / output elements may be physically integrated with control system 804. For example, input / output elements may include one or more tactile (e.g., physical switches, touchpads, buttons, etc.), audio (e.g., microphones, etc.), and / or visual (e.g., cameras, etc.) devices electrically connected (e.g., via one or more wired connections) to one or more components of AWG system 110. As another example, input / output elements may include one or more tactile (e.g., percussion, etc.), audio (e.g., speakers, etc.), and / or visual (e.g., display screens, etc.) output devices electrically connected (e.g., via one or more wired connections) to one or more components of AWG system 110.

[0160] Additionally or alternatively, the input / output elements may include one or more input / output devices of a remote device communicatively connected to the control system 804. By way of example, the user interface 802 and / or the tissue interface 814 may include one or more tactile (e.g., physical switches, touchpads, buttons, etc.), audio (e.g., microphones, etc.), and / or visual (e.g., cameras, etc.) input devices electrically connected to the respective interfaces (e.g., via one or more wired connections). In some examples, input to the one or more input devices may cause the respective interfaces to provide control signals to the control system 804 to enable remote control of the AWG system 110 through the respective interfaces. As another example, the user interface 802 and / or the tissue interface 814 may include one or more tactile (e.g., percussion, etc.), audio (e.g., speakers, etc.), and / or visual (e.g., display screens, etc.) output devices electrically connected to the respective interfaces (e.g., via one or more wired connections). In some examples, each interface may receive an output signal from the control system 804, which may cause each interface to present AWG data to a user, allowing remote monitoring of the AWG system 110 through each interface.

[0161] In some embodiments, the computing ecosystem 800 provides for the use of artificial intelligence to control and enhance production of the AWG system 110. Current AWG systems 110 are generally controlled through the use of programmable logic controllers, microcontrollers, or microprocessors. Generally, such systems are limited to input from local sensors and buttons on the system that are not augmented with artificial intelligence. The AWG system 110 of the computing ecosystem 800 can utilize artificial intelligence to process input from multiple sources to include local sensors, user demographics, historical water usage, location, and ambient weather, and / or the like, to control when and how the machine should operate. The embodiments described herein use artificial intelligence to control the AWG system 110 for optimal use.

[0162] 9-12 illustrate schematic diagrams of exemplary AWG computing ecosystems, according to various embodiments.

[0163] 9 shows a schematic diagram of an exemplary AWG computing ecosystem 900 including a centralized system 816, a control system 804, and an organizational interface 814. As shown, in some examples, the control system 804 may be integrated with the centralized system 816 (e.g., a cloud system), in which case artificial intelligence may be used to determine trends and optimize execution time based at least in part on data collected from the control system 804 and / or one or more information sources. Additionally or alternatively, the centralized system 816 may provide data collected from one or more information sources, the organizational interface 814, and / or the like to the control system 804, in which case artificial intelligence may be used to determine trends and optimize execution time based at least in part on the collected data.

[0164] In various embodiments, the AWG system is deployed at an end user's location associated with a unique local environment. The AWG system may include a control system 804 that can optimize the runtime of the AWG system based at least in part on data collected about the unique local environment. Additionally or alternatively, the control system 804 may include edge hardware and / or edge software deployed on the AWG system that enables communication with a centralized system 816 (e.g., a cloud server). The centralized system 816 and / or the control system 804 may process data (e.g., via one or more cloud services, etc.) to generate trends, optimized energy output, and / or the like that are (i) utilized by the control system 804 to control the operation of the AWG system, (ii) sent from the centralized system 816 to the control system 804 for optimization, and / or (iii) sent to the organization interface 814 for reporting and / or other uses, such as, but not limited to, data analysis, marketing, and local optimization for future design, and / or the like.

[0165] For example, the control system 804 and / or the centralized system 816 may include one or more portions of an optimization module 908. The optimization module 908 may include computer-readable instructions configured to generate one or more energy-based outputs (e.g., time-based energy predictions, optimized energy outputs, etc.) based at least in part on one or more optimization inputs. The computer-readable instructions may define, for example, machine learning, rule-based, and / or any other type of model configured to map multiple optimization inputs to an energy-based output. The energy-based output may include, for example, one or more time-based energy predictions and / or optimized energy outputs, such as real-time and / or scheduled operation instructions (e.g., execution times, etc.) configured to initiate operation of one or more components of the AWG system based at least in part on the one or more time-based energy predictions.

[0166] In various embodiments, the optimization module 908 may include an optimization machine learning model, such as one or more supervised, unsupervised, semi-supervised, reinforcement, and / or similar learning models, configured to generate one or more energy-based outputs based at least in part on one or more optimization inputs. The optimization machine learning model may be pre-trained, for example, using labeled and / or unlabeled training data reflecting multiple training optimization inputs. In some examples, the training data may include performance data (e.g., historical water output, etc.) from the AWG system 110 (e.g., recorded and / or transmitted by the control system 804 and / or the centralized system 816) paired with multiple corresponding historical optimization inputs (e.g., recorded and / or transmitted by the control system 804 and / or the centralized system 816). For example, the performance data may include ground truth labels of the multiple corresponding historical optimization inputs. In some examples, the optimization machine learning model may be trained continuously by recording performance data (e.g., real-time water output, etc.), generating training pairs by matching the performance data with corresponding optimization inputs, and retraining the optimization machine learning model with the new training pairs through one or more training techniques (e.g., backpropagation, gradient descent, etc.). This may advantageously allow for dynamic consideration of new and / or modified optimization inputs over time.

[0167] In some embodiments, control system 804 includes a memory configured to store at least a portion of optimization module 908. Additionally or alternatively, control system 804 may include one or more edge devices communicatively connected via one or more wireless networks to a centralized system 816 configured to store, train, and / or implement optimization module 908 on behalf of control system 804. By way of example, centralized system 816 may include a connected cloud platform, including a distributed serverless cloud environment, a dedicated cloud server environment, and / or the like. In some examples, centralized system 816 may train an optimization machine learning model and provide the trained model to control system 804. In such cases, control system 804 may locally implement the optimization machine learning model to locally generate one or more energy-based outputs. Additionally or alternatively, control system 804 may provide one or more optimization inputs to centralized system 816 and receive one or more energy-based outputs from centralized system 816 based at least in part on the optimization inputs.

[0168] In some embodiments, the centralized system 816 includes multiple interconnected modules configured to perform one or more operations of the present disclosure. Specific communications within the centralized system 816 may take several forms and go through multiple steps to define a desired output for the AWG system and / or tissue interface 814.

[0169] In some demonstrative embodiments, the centralized system 816 may include a gateway module 902. The gateway module 902 may include an Internet of Things (IoT) gateway that receives the optimized data and enables directing the optimized data to one or more other modules of the centralized system 816. In some embodiments, the centralized system 816 may include a pre-processing module 904 configured to perform one or more data cleaning and / or formatting operations to reformat the optimized data into an interpretable data format. The pre-processing module 904 may include serverless functionality. Additionally or alternatively, the pre-processing module 904 may include a dedicated server tailored to the one or more data cleaning and / or formatting operations.

[0170] The pre-processed optimization data may be stored in knowledge base 906. Knowledge base 906 may include, for example, one or more memory devices configured to store any type of data construct, including, by way of example, one or more graph data structures, relational databases, lookup tables, and / or the like. In some examples, knowledge base 906 may include a queryable database that is accessible to one or more other components of centralized system 816 and / or control system 804, such as optimization module 908. In some examples, the pre-processed optimization data may be temporarily stored in knowledge base 906 and provided to another functional block for further use.

[0171] In some embodiments, the preprocessed data is used by the optimization module 908 to generate one or more optimized energy outputs for the AWG system 110. Additionally or alternatively, the preprocessed data may be utilized to train one or more portions of the optimization module 908. For example, the optimization module 908 may include one or more machine learning algorithms that enable optimization of the performance of the AWG system. In some examples, the preprocessed data may include training data for training the machine learning algorithms, as described herein.

[0172] During inference, the optimization module 908 may generate one or more optimized energy outputs based at least in part on at least a portion of the data from the optimization module 908. The one or more optimized energy outputs may be stored in an optimization database for later querying. In some examples, the one or more optimized energy outputs are returned to the knowledge base 906 (e.g., for later training and / or evaluation operations, etc.). In some examples, the one or more optimized energy outputs may be provided to the control system 804 via the gateway module 902 to control one or more components of the AWG system 110.

[0173] In some examples, optimization database 910 may include one or more data constructs, such as a relational database, a graph database, and / or the like, which may be accessible to one or more interfaces, such as organization interface 814, via elastic compute interface 912. In this manner, optimization data collected by centralized system 816 may be utilized for other applications, such as increased weather monitoring, studying water use trends by location, demographics, use cases, etc. For example, the optimization data may enable researchers and regulators to optimize proposals or regulations around water to ensure best use, freedom of use, and / or regulation of water. In some examples, data may also be input directly into optimization database 910 via elastic compute interface 912 to centralized system 816.

[0174] 9, the optimization data may be collected from a variety of systems, including one or more connected control systems, one or more information source systems, one or more user interfaces, and / or the like. In some examples, the optimization data may include sensor data from the AWG system 110. For example, the control system 804 may use one or more sensors 914 of the AWG system 110 to record one or more real-time environmental attributes of its unique environment. The one or more sensors 914 may include one or more internal and / or external temperature sensors, gas sensors, pressure sensors, humidity sensors, wind sensors, weight sensors, and / or the like.

[0175] In various embodiments, the sensor data may reflect the effectiveness of one or more water-producing operations at a particular time and / or location (e.g., optimization data, etc.). For example, the sensor data may indicate one or more ambient environmental attributes such as temperature, ambient pressure, CO2 concentration, relative humidity, cloud cover, wind speed and directional irradiation, interior and exterior water tank levels, water conductivity, and / or the like. In some examples, the sensor data may be time-stamped to reflect the time and date corresponding to the recorded ambient environmental attribute. In some examples, the optimization data (and / or one or more derivatives thereof reflecting the real-time environmental attribute) may be provided to the centralized system 816 and / or a locally stored optimization module to generate one or more time-based energy forecasts.

[0176] In various embodiments, the sensor data may reflect one or more water use requirements (e.g., water use data, etc.) at a particular time and / or location. For example, sensors 914 include one or more environmental sensors in an at least partially controlled environment, such as a greenhouse (e.g., as shown herein with reference to FIGS. 4-7). By way of example, sensors 914 may include one or more sensors in a growing medium of a growing habitat and / or any other sensors in communication with an environmental control system. In some examples, the water use requirements may be based at least in part on humidity levels (e.g., as measured by one or more environmental humidity sensors, etc.), soil moisture levels (e.g., as measured by one or more growing medium sensors, etc.), plant growth rate and / or plant condition (e.g., as measured by one or more video and / or image sensors, etc.), sun canopy opacity (e.g., an opacity sensor in the sun canopy), and / or the like. In some examples, the water use data (and / or one or more derivatives thereof) may be provided to centralized system 816 and / or a locally stored optimization module to generate one or more optimized energy outputs.

[0177] In various embodiments, the sensor data may reflect the energy capacity of an environment at a particular time and / or location. By way of example, sensors 914 (e.g., one or more environmental sensors, etc.) may be configured to record energy usage data regarding one or more electronic devices in the environment. By way of example, the energy usage data may include energy consumption of one or more robotic devices (e.g., a planter robot in a greenhouse, etc.), one or more lighting devices (e.g., LEDs, etc.), one or more climate control devices (e.g., heating, ventilation, and air conditioning units, etc.).

[0178] In some embodiments, the sensor data may reflect the performance of the AWG system 110 at a particular time and / or location (e.g., performance data, etc.). For example, the performance data may be based on the water output of the AWG system 110. The one or more sensors 914 may include one or more water output measuring devices (e.g., flow measurement sensors, weight scales, etc.) that may measure and record the flow, weight, and / or volume of water collected by the AWG system 110. In some examples, the performance data may include time-stamped performance data, allowing for a direct comparison of the performance data with time-stamped optimization data. In this manner, performance data from local system sensors 914 on the AWG system 110 may be leveraged in a recursive manner, allowing the optimization module 908 to continuously improve.

[0179] In various embodiments, optimization data, water usage data, energy usage data, and / or performance data recorded locally by one or more sensors 914 may be combined with data from external sources to generate an overall energy-based output for optimizing the AWG system 110. For example, as described with reference to Figures 10-12, local data received from the control system 804 about the local environment may be augmented with external data sourced from multiple external systems to improve optimized control of the AWG system 110.

[0180] 10 shows a schematic diagram of an exemplary AWG computing ecosystem 900 including a centralized system 816, a control system 804, and multiple connected control systems. By way of example, the multiple connected control systems 902A-D may correspond to multiple AWG systems that may be linked together to form an intelligent cluster of systems. In some examples, the intelligent cluster of systems may be based at least in part on location proximity to increase the data collected to enable greater optimization.

[0181] In some embodiments, multiple connected control systems 902A-D with edge capabilities may communicate data back and forth. The communication may occur via one or more network interfaces (e.g., Bluetooth, Wireless Fidelity (WiFi) networks, etc.). In some examples, the control system 804 may comprise one of multiple connected control systems 902A-D in an intelligent cluster.

[0182] In some examples, the intelligent clusters may form one or more hierarchical relationships. For example, a first connected control system 902B may act as a leader hub and collect data from control system 804 and / or one or more other connected control systems 902A. In some examples, the one or more hierarchical relationships may include location-based relationships. For example, a first connected control system 902B may act as a leader hub for multiple connected control systems 902A within a threshold distance from the first connected control system 902B (e.g., within a 100-foot radius forming an information network hub). Other connected control systems 902C and 902D outside the threshold distance may form one or more other hierarchical relationships.

[0183] Each hierarchical relationship may facilitate communication between the centralized system 816 and each of the subordinate control systems in the particular relationship. For example, the connected control system 902B may be configured to relay optimization data and / or optimized energy output between the control system 804 and / or the connected control system 902A and the centralized system 816.

[0184] In some examples, the first connected control system 902B may be determined as the control system with the greatest amount of other system connections. In certain embodiments, network expansion may be achieved through a chain network in which several hub units send data to a single hub unit, which then sends the data to the centralized system 816. In this way, data may be aggregated across multiple connected control systems 902B to reduce the data footprint reaching the centralized system 816 from control systems that may not be within a threshold distance of another connected control system 902B. Each of the control systems may also receive data from the centralized system 816 for optimization, which is sent as operating procedure updates to all other control systems.

[0185] FIG. 11 shows a schematic diagram of an exemplary AWG computing ecosystem 1100 including a centralized system 816, a control system 804, and one or more information sources 818.

[0186] In some embodiments, information sources 818 may include one or more weather sources, one or more water usage sources, and / or the like. Weather sources may include, for example, one or more real-time and / or forecast services (e.g., weather stations, etc.) accessible through one or more application programming interfaces (APIs). The one or more real-time and / or forecast services may provide recorded local ambient conditions, forecasted weather conditions, and / or the like. In some examples, historical weather data may be stored as training data for the optimization module 908.

[0187] Additionally or alternatively, information sources 818 may include one or more local water usage sources, such as one or more smart home devices configured to sense and record water usage from various water sources within the environment. The local water usage sources may be accessible through one or more APIs and may provide real-time, historical, and / or forecasted water requirements for the environment that reflect the average and / or expected water usage of one or more users within the environment. In some examples, water usage data may be time-stamped to reflect the corresponding time of day, day of the week, month, and / or the like.

[0188] FIG. 12 shows a schematic diagram of an exemplary AWG computing ecosystem 1200 including a centralized system 816, a control system 804, one or more information sources 818, and one or more user interfaces 802.

[0189] In some embodiments, user interface 802 includes one or more interfaces between (i) the centralized system 816 and (ii) user devices associated with end users of control system 804. In some examples, end users may provide one or more optimization inputs, such as habitual water usage at a particular time, electricity pricing, electricity source (e.g., solar, wind, grid, generator), and / or the like, to tailor the optimized energy output to one or more known criteria. In some examples, user inputs may be entered directly into centralized system 816. For example, end users may utilize user interface 802 to enter data such as demographic information, age, gender, weight, height, water usage at different times, number of occupants, number of sinks, number of toilets, number of showers, sprinkler system usage, size of the end user's storage tank, and / or other user data to tailor the optimized energy output to maintenance issues, current machine performance, historical usage, and / or the like.

[0190] Example Optimized Control Process 13 is a flowchart illustrating an example training process 1300 for generating an optimized machine learning model, according to some embodiments discussed herein. The flowchart depicts a training process 1300 for improving the performance of an AWG system. Process 1300 may be implemented by one or more computing devices, entities, and / or systems described herein. For example, through various steps / operations of process 1300, a computing system such as control system 804, centralized system 816, and / or a combination thereof may leverage improved machine learning techniques to generate predictive energy insights tailored to the AWG system. By doing so, process 1300 enables automated, adaptive, and real-time control of the AWG system tailored to its unique environment.

[0191] 13 illustrates an example process 1300 for explanatory purposes. Although the example process 1300 depicts a particular sequence of steps / actions, the sequence may be changed without departing from the scope of the present disclosure. For example, some of the depicted steps / actions may be performed in parallel or in a different sequence without significantly affecting the functionality of the process 1300. In other examples, different components of an example device or system implementing the process 1300 may perform functions substantially simultaneously or in a particular sequence.

[0192] In some embodiments, process 1300 includes receiving training data for a machine learning model at step / act 1302. For example, a computing system may receive training data for an optimized machine learning model.

[0193] The optimization machine learning model may include parameters, hyperparameters, and / or defined behaviors of a rule-based and / or machine learning model (e.g., a model including at least one of one or more rule-based layers, one or more layers depending on trained parameters, coefficients, and / or the like). The optimization machine learning model may include, for example, one or more machine learning models trained to generate time-based energy predictions and / or optimized energy outputs for the AWG system. The optimization machine learning model may include one or more of any type of machine learning model, including one or more supervised, unsupervised, semi-supervised, and / or reinforcement learning models. In some embodiments, the optimization machine learning model may include multiple sub-models configured to perform one or more different stages of the prediction process.

[0194] In some examples, the optimization machine learning model may include one or more classification and / or regression models, which may include one or more neural networks, decision trees, logistic regression, state vector machines, and / or the like. In some examples, the optimization machine learning model may be trained using one or more supervised training techniques, such as backpropagation, gradient descent, and / or the like. In some examples, the optimization machine learning model is trained to optimize a loss function configured to optimize performance of the AWG system by minimizing the amount of energy consumed while achieving one or more water requirements of a location associated with the AWG system. In some examples, the loss function may be used with labeled training data to improve the performance of the model.

[0195] In some embodiments, the training data may include a plurality of labeled optimization training entries. Each of the plurality of labeled optimization training entries may include a set of historical optimization inputs and ground truth labels corresponding to the set of historical optimization inputs. In some examples, the ground truth labels may include performance data. For example, the performance data may indicate ground truth water output from the AWG system based at least in part on one or more historical atmospheric water production operations. Additionally or alternatively, the ground truth labels may include energy usage data. For example, the energy usage data may indicate ground truth energy consumption from the AWG system based at least in part on one or more historical atmospheric water production operations.

[0196] In some embodiments, process 1300 includes, at step / operation 1304, using the machine learning model to generate one or more training time-based energy predictions based at least in part on the set of historical optimization inputs from the training data. For example, a computing system may input the set of historical optimization inputs to the optimization machine learning model and receive one or more training time-based energy predictions as output from the optimization machine learning model.

[0197] In some examples, one or more training time-based energy predictions are output by a first sub-model of the optimized machine learning model. For example, the training time-based energy predictions may include one or more intermediate outputs for a second sub-model of the optimized machine learning model. Additionally or alternatively, the training time-based energy predictions may include a final output of the optimized machine learning model.

[0198] In some embodiments, process 1300 includes generating one or more training-optimized energy outputs based at least in part on the time-based energy predictions at step / operation 1306. For example, the computing system may generate one or more training-optimized energy outputs based at least in part on the time-based energy predictions.

[0199] In some embodiments, the optimization machine learning model may generate a training-optimized energy output. For example, one or more training-optimized energy outputs may be output by a second sub-model of the optimization machine learning model. For example, the optimization machine learning model may include multiple machine learning models trained end-to-end to generate an optimized energy output based at least in part on one or more optimization inputs.

[0200] Additionally or alternatively, the training optimized energy output may be generated by a subsequent model based at least in part on the time-based energy forecast. For example, the subsequent model may include one or more rule-based models configured to apply one or more optimization rules to the one or more training optimized energy output and water usage data for the AWG system. The optimized rules may include, for example, one or more statistical models for deriving optimal run times and / or optimal run parameters for the AWG system based at least in part on the time-based energy forecast and one or more water requirements for a location associated with the AWG system.

[0201] In some embodiments, process 1300 includes, at step / operation 1308, using a loss function to generate a model loss based at least in part on historical performance data corresponding to the training time-based energy predictions, the training optimized energy output, and / or the historical optimization inputs. For example, the computing system may use the loss function to generate a model loss for the optimized machine learning model.

[0202] The loss function may include any type of supervised loss function, including, for example, Huber loss, cross-entropy loss, mean squared error, and / or the like. For example, the loss function may measure the error between (i) the training time-based energy predictions and (ii) one or more ground truth water output and / or energy costs. For example, the optimized machine learning model, and / or its first sub-model, may be trained to reduce the loss (e.g., prediction error) between the predictions and the ground truth corresponding to the predictions to improve the model's predictive accuracy.

[0203] In some examples, the loss function may include a reinforcement learning function such as a greedy method and / or the like. For example, the loss function may focus on measuring the overall reward of the optimized energy output based at least in part on the ground truth water output and / or energy consumption. For example, the reinforcement learning function may be configured to reward (i) higher water output (e.g., consumption above a threshold), (ii) lower energy consumption (e.g., consumption below a threshold), and / or (iii) achievement of one or more water requirements. Additionally or alternatively, the reinforcement learning function may be configured to punish (i) lower water output (e.g., consumption below a threshold), (ii) higher energy consumption (e.g., consumption above a threshold), and / or (iii) achievement of one or more water requirements.

[0204] In some embodiments, process 1300 includes, at step / action 1310, updating one or more parameters of the machine learning model using a model training technique based at least in part on the model loss. For example, the computing system may update one or more parameters of the optimized machine learning model using a model training technique based at least in part on the model loss. The model training technique may include any type of model training technique, including backpropagation, gradient descent, and / or the like.

[0205] In some examples, the optimized machine learning model may include one or more submodels. The one or more submodels may be trained end-to-end at least in part using a first model loss, such as a reinforcement learning loss output by a reinforcement learning function. Additionally or alternatively, (i) a first submodel, such as a prediction submodel configured to output one or more time-based energy predictions, may be trained at least in part individually using a first model loss, such as a supervised learning loss output by a supervised loss function, and (ii) a second submodel, such as an optimization submodel configured to output one or more optimized energy outputs, may be trained at least in part individually using a second model loss, such as a reinforcement learning loss output by a reinforcement loss function.

[0206] 14 is a flowchart illustrating an example control process 1400 for optimally controlling an AWG system using a trained, optimized machine learning model, according to some embodiments discussed herein. The flowchart depicts the control process 1400 for improving the performance of an AWG system. The process 1400 may be implemented by one or more computing devices, entities, and / or systems described herein. For example, through various steps / operations of the process 1400, a computing system such as the control system 804, the centralized system 816, and / or a combination thereof may leverage improved machine learning techniques to generate predictive energy insights tailored to the AWG system. By doing so, the process 1400 enables automated, adaptive, and real-time control of the AWG system tailored to its unique environment.

[0207] 14 illustrates an example process 1400 for explanatory purposes. Although the example process 1400 depicts a particular sequence of steps / actions, the sequence may be changed without departing from the scope of the present disclosure. For example, some of the depicted steps / actions may be performed in parallel or in a different sequence without significantly affecting the functionality of the process 1400. In other examples, different components of an example device or system implementing the process 1400 may perform functions substantially simultaneously or in a particular sequence.

[0208] Process 1400 may begin at step / operation 1310 of process 1300, which includes updating one or more parameters of a machine learning model to generate a trained machine learning model. In some embodiments, the trained machine learning model may be stored in memory and retrieved, invoked, and / or otherwise accessed to perform one or more steps / operations of process 1400. Additionally or alternatively, the trained machine learning model may be provided to another computing system (e.g., a control system, etc.) for use in performing one or more steps / operations of process 1400.

[0209] For example, the steps / operations of process 1400 may be performed by a control system of the AWG system including one or more controllers electrically connected to one or more subsystems of the AWG system (e.g., a humidity augmentation system, an air preconditioning system, a power generation module, one or more valves (e.g., electromechanical mixing valves, etc.), motors, actuators, a thermal condenser, etc.) and configured to interact with the one or more controllers to control operation of the AWG system. Additionally or alternatively, the steps / operations of process 1400 may be performed by a centralized system, and the control system of the AWG system may act as an edge device connecting the AWG system to the centralized system.

[0210] In some embodiments, process 1400 includes receiving one or more optimization inputs at step / operation 1402. For example, a computing system may receive the one or more optimization inputs. The one or more optimization inputs may correspond, for example, to a time and / or location associated with operation of the AWG system. In certain embodiments, the optimization inputs may include data generated by the AWG system and / or data generated by a third-party system (e.g., a third-party weather forecasting system). In some embodiments, the optimization inputs may include data indicative of the likely effectiveness of the AWG system in producing water (e.g., weather and / or location data indicative of the likely efficiency of producing water from air). In some embodiments, the optimization inputs may alternatively or additionally include data indicative of water need (e.g., data indicative of soil moisture levels in a greenhouse attached to the AWG system). In some embodiments, the optimization inputs indicative of water need may be weighted more or less heavily than data indicative of the likely efficiency of water production.

[0211] The optimization inputs may include one or more real-time, historical, forecast, and / or static inputs. For example, the real-time inputs may indicate current performance, current water usage, and / or current environmental conditions associated with the AWG system. The one or more historical inputs may include one or more historical performance, water usage, and / or environmental condition trends associated with the AWG system. The one or more forecast inputs may include one or more forecast performance, water usage, and / or environmental condition trends associated with the AWG system. The one or more static inputs may include one or more user-defined metrics that may reflect performance, water usage, and / or environmental conditions associated with the AWG system. By way of example, some optimization inputs may include local current machine sensor data, historical machine sensor data, historical water usage, user demographics, user applications, AWG system location, historical weather patterns, predicted weather forecasts, precipitation, fog conditions, and / or the like.

[0212] In some examples, optimization inputs may be aggregated and streamed to the computing system from multiple different sources. For example, the optimization inputs may include one or more AWG system inputs, connected AWG system inputs, source inputs, organizational inputs, and / or user inputs.

[0213] For example, one or more optimization inputs may include local sensor data from the AWG system. The AWG system inputs may include, for example, local sensor data recorded by one or more sensors of the AWG system. The local sensor data may reflect, for example, the performance of the AWG system (e.g., water output, energy usage, etc.), one or more local environmental conditions (e.g., temperature, humidity, etc.), and / or the like.

[0214] By way of example, the local sensor data may include local optimization data, water usage data, energy usage data, performance data, and / or the like that reflect one or more attributes of the AWG system and / or associated environment (e.g., a controlled environment such as a greenhouse). The local optimization data may reflect the effectiveness of one or more water-producing operations at a particular time and / or location, such as, for example, one or more ambient environmental attributes, and / or the like. The water usage data may reflect one or more water usage requirements, such as humidity levels in the environment, soil moisture levels, plant growth rates and / or plant conditions, solar canopy opacity, and / or the like. The energy usage data may reflect the energy capacity of the environment, such as the energy consumption of one or more robotic devices (e.g., a planter robot in a greenhouse), one or more lighting devices (e.g., LEDs), one or more climate control devices (e.g., heating, ventilation, and air conditioning units), and / or the like. The performance data may reflect the water output of the AWG system, such as the amount of water collected by the AWG system.

[0215] As another example, an AWG system may be associated with a cluster of multiple connected AWG systems. Each of the multiple connected AWG systems may be associated with a different location. In some examples, one or more optimization inputs may include remote sensor data from each of the multiple connected AWG systems. For example, the connected AWG system inputs may include remote sensor data recorded by one or more sensors of one or more connected AWG systems in the cluster of connected AWG systems. The remote sensor data may reflect, for example, performance of each AWG system (e.g., water output, energy usage, etc.), one or more remote environmental conditions (e.g., temperature, humidity, etc.) at the location of each AWG system, and / or the like.

[0216] As yet another example, one or more optimization inputs may include current and / or future weather data from one or more external sources. The source inputs may include, for example, real-time, historical, and / or forecasted information recorded, generated, and / or maintained by one or more sources. The source may include, for example, a weather forecasting service, and the source inputs may reflect one or more real-time, historical, and / or forecasted weather conditions (e.g., humidity, temperature, precipitation, etc.) at a location corresponding to the AWG system.

[0217] Organizational and / or user inputs may include pre-set and / or dynamically modified user-defined parameters for the AWG system. The inputs may reflect required water usage, available energy output, one or more facility characteristics (e.g., number of toilets, sinks, people, etc.), and / or any other information that can tailor the performance of the AWG system to a particular environment.

[0218] In some embodiments, process 1400 includes, at step / operation 1404, using the trained machine learning model to generate one or more time-based energy predictions based at least in part on the one or more optimization inputs. For example, a computing system may use the optimization machine learning model to generate one or more time-based energy predictions for an AWG system based at least in part on the one or more optimization inputs. In some examples, the optimization machine learning model may be pre-trained using one or more supervised training techniques (e.g., backpropagation, etc.) based at least in part on a training dataset including a plurality of labeled optimization training entries.

[0219] The time-based energy forecast may reflect the expected energy consumption of the AWG system to produce water at and / or during a particular time period. In some examples, the time-based energy forecast may reflect the ratio of energy to water output at and / or during a particular time period. For example, the time-based energy forecast may indicate a threshold amount of energy to achieve a particular water output at and / or during a particular time period. In some examples, the time-based energy forecast may reflect the expected maximum achievable water output at and / or during a particular time period.

[0220] In some examples, one or more time-based energy forecasts may be generated for one or more different time periods. For example, the one or more time-based energy forecasts may include a current time-based energy forecast for a current time period and one or more subsequent time-based energy forecasts for one or more future time periods following in time the current time period. In some examples, the one or more time-based energy forecasts may enable the computing system to generate a forecast map of water production to energy consumption for a location associated with the AWG system. The forecast map of water production to energy consumption may reflect the ratio of energy to water production over one or more time periods and / or locations associated with the AWG system and / or one or more connected systems. In some examples, the forecast map of water production to energy consumption may be rendered within a user interface (see the user interfaces and organization interfaces of FIGS. 8-9 and 11-12). For example, data indicating the ratio of energy to water production over one or more time periods and / or locations may be provided (and / or accessed) by one or more user computing entities. In some examples, the ratio of energy to water production may be plotted across an interactive timeline that may be rendered via one or more user interfaces of the user computing entity.

[0221] In some embodiments, process 1400 includes, at step / operation 1406, generating one or more optimized energy outputs based at least in part on the time-based energy forecasts. For example, the computing system may generate the optimized energy outputs based at least in part on the time-based energy forecasts. In some examples, the computing system may generate the optimized energy outputs for the AWG system based at least in part on one or more time-based energy forecasts and water usage data for a location corresponding to the AWG system.

[0222] The water usage data may include and / or be determined using one or more of the optimization inputs, for example. For example, the water usage data may be based at least in part on user input or sensor data from the AWG system, one or more connected AWG systems, one or more interfaces (e.g., a user, an organization, etc.), and / or one or more sources.

[0223] In some examples, one or more optimization inputs may be utilized to generate water usage data reflecting water usage trends for a location associated with the AWG system. In this manner, a water usage forecast for the location may be constructed that reflects the time-based water requirements of the location associated with the AWG system. In some examples, the optimized energy output may be based at least in part on a comparison between the time-based water requirements and / or the time-based energy forecast. For example, the optimized energy output may include one or more recommended control instructions for optimizing operation of the AWG system to operate at the best time of day, week, year, etc. to achieve the maximum amount of water with the lowest possible input energy while achieving the water requirements of the particular location.

[0224] The optimized energy output may, for example, identify one or more optimal run times for collecting water using the AWG system. As an example, one or more optimization inputs may reflect current high humidity for the current week and forecasted low humidity for the subsequent week that temporally follows the current week. In such a case, the optimized energy output may identify the current week as the optimal run time for water collection to achieve the water requirements for both the current and subsequent weeks. In some examples, the output may be generated by balancing humidity data with other optimization inputs, such as water usage data, water tank levels, electricity prices, and / or the like. In this way, the computing system may maximize available water during low humidity conditions when the system's production efficiency is low (e.g., as reflected by time-based energy forecasts, etc.).

[0225] In some embodiments, the optimized energy output may modify one or more performance parameters for one or more run times. For example, the optimized energy output may include optimal performance parameters for current and / or scheduled run times. In some examples, the optimal performance parameters may be associated with one or more operating states (e.g., a standby state, a low energy state, a high energy state, and / or the like) to limit one or more energy consumptions of the AWG system during a particular time period. As another example, the optimal performance parameters may include one or more run time operating thresholds and / or conditions. For example, the one or more run time operating thresholds and / or conditions may reflect one or more temperature ranges of components of the AWG system, such as an in-line heater (e.g., for heating a desiccant), one or more valve control time constraints (e.g., particularly in the case of a batch-style AWG system), one or more valve positioning constraints (e.g., where the desiccant can be directed through one or more optional fluid paths), and / or the like.

[0226] In some embodiments, the optimized energy output is generated by applying one or more rule-based models to time-based energy forecasts and / or water usage data. Additionally or alternatively, the optimized energy output may be generated using one or more machine learning techniques. For example, in some embodiments, the optimized energy output may be generated by sub-models of an optimization machine learning model. By way of example, one or more time-based energy forecasts may be generated by a first sub-model of the optimization machine learning model as intermediate outputs. The intermediate outputs may then be processed by a second sub-model to generate the optimized energy output. In this manner, the optimization machine learning model may include an improved machine learning architecture that breaks down traditional complex optimization methods into a sequence of connected machine tilt models.

[0227] In some embodiments, process 1400 includes, at step / operation 1408, initiating control of the AWG system based at least in part on the optimized energy output. For example, the computing system may initiate control of the AWG system based at least in part on the optimized energy output. For example, the computing system may generate one or more control instructions based at least in part on the optimized energy output and / or a time-based energy forecast. In some examples, the computing system may communicate one or more control instructions to one or more controllers of the AWG system to initiate one or more atmospheric water production operations based at least in part on one or more time-based energy forecasts, optimized energy output, and / or the like. In this manner, data aggregated from multiple sources (e.g., optimization inputs, etc.) may be used to optimize the runtime of the AWG system.

[0228] In some embodiments, process 1400 includes, at step / operation 1410, receiving performance data associated with the optimized energy output. For example, the computing system may receive performance data associated with the optimized energy output and / or a time-based energy forecast. For example, the computing system may receive energy usage data and / or performance data corresponding to one or more atmospheric water generation operations performed based at least in part on the optimized energy output and / or the time-based energy forecast. The performance data may indicate, for example, an amount of water output by the AWG system in response to the atmospheric water generation operations. The energy usage data may indicate an amount of energy consumed to perform the atmospheric water generation operations.

[0229] In some embodiments, process 1400 includes, at step / operation 1412, modifying the training data based at least in part on the performance data. For example, the computing system may modify the training data based at least in part on the performance data. For example, the computing system may store one or more optimization inputs, energy usage data, and performance data as labeled optimization training entries in a training data set. In some examples, the optimization training entries may also include one or more time-based energy predictions and / or optimized energy output.

[0230] In some embodiments, process 1400 may then return to step / operation 1302 to retrain the optimized machine learning model. For example, a computing system may iteratively perform one or more steps / operations of process 1300 to recursively update the optimized machine learning model based at least in part on the performance of the AWG system over time. As a particular example, the computing system may retrain the optimized machine learning model based at least in part on the labeled optimization training entries generated using the steps / operations of process 1400. In this manner, optimization of the AWG system may improve with machine operation. In some cases, to improve machine learning, an end user may input downtime periods during which data is not used by the machine learning model, such as times when the end user does not use water consistently as usual. Over time, the machine learning model may learn and adapt to these downtime periods based at least in part on user input and recorded water usage data during the downtime periods.

[0231] conclusion Numerous modifications and other embodiments will come to mind to one skilled in the art to which this disclosure pertains having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. It is to be understood, therefore, that the disclosure is not limited to the specific embodiments disclosed, and that modifications and other embodiments are intended to be included within the scope of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.

Claims

1. 1. A computer-implemented method comprising: receiving, by one or more processors, one or more optimization inputs for a time and a location associated with operation of the atmospheric water generating system; generating, by the one or more processors and using an optimization machine learning model, one or more time-based energy forecasts for the atmospheric water generating system based at least in part on the one or more optimization inputs; and communicating, by the one or more processors, one or more control instructions to one or more controllers of the atmospheric water generating system to initiate one or more atmospheric water generating operations based at least in part on the one or more time-based energy forecasts.

2. 2. The computer-implemented method of claim 1, wherein the optimized machine learning model is pre-trained using one or more supervised training techniques based at least in part on a training dataset including a plurality of labeled optimized training entries, each of the plurality of labeled optimized training entries including a set of historical optimization inputs and historical performance data corresponding to the set of historical optimization inputs.

3. The computer-implemented method of claim 2 , wherein the historical performance data indicates ground truth water output from the atmospheric water generating system based at least in part on one or more historical atmospheric water generating operations.

4. receiving, by the one or more processors, energy usage data and performance data corresponding to the one or more atmospheric water generating operations; 3. The computer-implemented method of claim 2, further comprising storing, by the one or more processors, the one or more optimization inputs, the energy usage data, and the performance data as labeled optimization training entries in the training dataset.

5. 5. The computer-implemented method of claim 4, further comprising: retraining, by the one or more processors, the optimized machine learning model based at least in part on the labeled optimized training entries.

6. generating, by the one or more processors, an optimized energy output for the atmospheric water generating system based at least in part on the one or more time-based energy forecasts and water usage data for the location corresponding to the atmospheric water generating system; The computer-implemented method of claim 1 , further comprising: generating, by the one or more processors, the one or more control instructions based at least in part on the optimized energy output.

7. The computer-implemented method of claim 6 , wherein the water usage data is based at least in part on user input or sensor data from the atmospheric water generating system.

8. The computer-implemented method of claim 1 , wherein the one or more optimization inputs include sensor data from the atmospheric water generating system.

9. The computer-implemented method of claim 8 , wherein the one or more optimization inputs include current or future weather data from one or more external sources.

10. 2. The computer-implemented method of claim 1, wherein the atmospheric water generating system is associated with a cluster of a plurality of connected atmospheric water generating systems, each of the plurality of connected atmospheric water generating systems being associated with a different location, and the one or more optimization inputs include remote sensor data from each of the plurality of connected atmospheric water generating systems.

11. communicating the one or more control instructions 2. The computer-implemented method of claim 1, comprising providing, by the one or more processors, the one or more control instructions to an edge device that (i) is physically disposed on the atmospheric water generating system and (ii) is electrically connected to at least one of the one or more controllers of the atmospheric water generating system.

12. 1. A computing system comprising: a memory; and one or more processors communicatively coupled to the memory, the one or more processors: receiving one or more optimization inputs for time and location associated with operation of the atmospheric water generating system; using an optimization machine learning model to generate one or more time-based energy forecasts for the atmospheric water generating system based at least in part on the one or more optimization inputs; and a computing system configured to communicate one or more control instructions to one or more controllers of the atmospheric water generating system to initiate one or more atmospheric water generating operations based at least in part on the one or more time-based energy forecasts.

13. 13. The computing system of claim 12, wherein the optimized machine learning model is pre-trained using one or more supervised training techniques based at least in part on a training dataset including a plurality of labeled optimized training entries, each of the plurality of labeled optimized training entries including a set of historical optimization inputs and performance data corresponding to the set of historical optimization inputs.

14. The computing system of claim 13 , wherein the performance data indicates ground truth water output from the atmospheric water generating system based at least in part on one or more historical atmospheric water generating operations.

15. the one or more processors: generating an optimized energy output for the atmospheric water generating system based at least in part on the one or more time-based energy forecasts and water usage data for the location corresponding to the atmospheric water generating system; The computing system of claim 12 , further configured to generate the one or more control instructions based at least in part on the optimized energy output.

16. The computing system of claim 15 , wherein the water usage data is based at least in part on user input or sensor data from the atmospheric water generating system.

17. The computing system of claim 12 , wherein the one or more optimization inputs include sensor data from the atmospheric water generating system.

18. 20. The computing system of claim 17, wherein the one or more optimization inputs include current or future weather data from one or more external sources.

19. 1. An atmospheric water generating system comprising: one or more controllers electrically connected to one or more subsystems of the atmospheric water generating system; 1. A control system comprising: a memory; and one or more processors communicatively coupled to the memory, the one or more processors: receiving one or more optimization inputs for a time and a location associated with operation of the atmospheric water generating system; using an optimization machine learning model to generate one or more time-based energy forecasts for the atmospheric water generating system based at least in part on the one or more optimization inputs; a control system configured to communicate one or more control commands to the one or more controllers to initiate one or more atmospheric water generation operations based at least in part on the one or more time-based energy forecasts.

20. 20. The atmospheric water generating system of claim 19, wherein the atmospheric water generating system is associated with a cluster of a plurality of connected atmospheric water generating systems, each of the plurality of connected atmospheric water generating systems being associated with a different location, and the one or more optimization inputs include remote sensor data from each of the plurality of connected atmospheric water generating systems.

Images