System and method for generative design of energy storage devices

Abstract

To provide a system and method for generative design of an energy storage device.SOLUTION: A method automatically builds at least one model of an energy storage device. The building is based on a design parameter space and employs a machine learning process. The method automatically performs a simulation of the energy storage device using the design parameter space, a design evaluation space, and the at least one model built. The performing produces at least one prediction. The method automatically evolves at least one of the design parameter space and the design evaluation space. In an event the at least one prediction indicates that a product design objective or model design objective has been achieved, the method automatically converges on the design parameter space evolved, thereby completing a generative design of the energy storage device and, otherwise, repeats the building, performing, and evolving.SELECTED DRAWING: Figure 1A

Description

Background Art

[0001] Generative design refers to a technology where models, such as three-dimensional (3D) computer-aided design (CAD) models or another type of computer-based model, can be created and optimized by computer software. Designers / engineers / scientists can typically interact with the technology to create excellent designs and drive product innovation more quickly. Generative design can be considered a design exploration process that expands the designer's world of known effective solutions to their design challenges.

[0002] For example, as a non-limiting example, the designer may provide parameters such as materials, size, weight, strength, manufacturing methods, etc., and the generative design software may explore possible combinations of solutions and may quickly generate hundreds or thousands of design options. From there, the designer, or the software, can filter and select the results that best meet the goals of the design process. When using generative design, the designer is not limited by their imagination or past experience.

[0003] The input to generative design tools may be similar to the input to many optimization tools, but generative design generates multiple valid (e.g., high-performance but cost-effective) designs or solutions instead of one version of an optimized solution. In addition to creating completely new solutions, another area where generative design stands out is its ability to take manufacturability into account. When using generative design, simulations may be incorporated into the design process, and generative design software can, as a non-limiting example, generate only designs that can be fabricated using the specified manufacturing method(s) or according to the parameters and / or requirements of the specified manufacturing method(s).

Summary of the Invention

[0004] The various embodiments of this disclosure generally relate to generative designs for energy storage devices. Exemplary embodiments, in non-limiting examples, enable the design / specification of compounds, components, additives, formulations, and recipes through generative design for use in energy storage devices. Exemplary embodiments enable the design / specification of such compounds, components, additives, formulations, and recipes, individually or in combination, to influence, in non-limiting examples, the performance of energy storage devices such as batteries. Performance may, in non-limiting examples, be based on attributes such as capacity, temperature tolerance, maximum charge and discharge rates or profiles, cycle life, ease of manufacture, quality, safety, reprocessing or repurposing at the end of life, and the shape, configuration, or form of the battery.

[0005] In relation to batteries, there are several challenges that may be addressed through generative design applied to those batteries. Battery manufacturing is a complex, iterative, multi-factor process. For example, battery manufacturing may include the production of electrodes. Electrode production may include, in non-limiting examples, the production of anodes, cathodes, active materials, electrolytes, binders, and separators. Battery manufacturing may include the production of cells, where a single unit that performs the function of a "battery" is produced. Battery manufacturing may include module assembly, where multiple cells enclosed in a metal case are connected to an electronic management system. Battery manufacturing may further include performing pack assembly, where connections of multiple modules, sensors, and controllers are installed within a case.

[0006] As a non-limiting example, generative design for batteries may take into account charge state, battery chemistry and technology, efficiency, safety, aging, temperature diffusion, durability, and sustainability and recyclability. In this way, generative design may be optimized to enable battery users to stop worrying about insufficient charging time, and generative design may find the most high-performing and safest chemistry for a battery. Generative design may make a battery more efficient by generating enough power to increase functionality, and may assess battery safety to prevent accidents. Generative design may assess battery durability to prevent capacity loss, and may assess the effect of temperature diffusion on battery performance. Generative design may assess durability and enable the design of "green" batteries.

[0007] As a non-limiting example, and in relation to batteries, generative design may perform the identification of chemical components such as additives for the electrolyte, coatings for the electrodes, by-product scavengers, and degradation inhibitors. Generative design may perform the identification of formulations such as mixture design (quantity) and specification design (grade, purity, and impurities). Generative design may determine the geometric shape of the battery, such as size and form factor, electrode configuration and structure, and tab position and size. Generative design may determine the process design for the battery, such as process sequence, process operating parameters, selection and identifiers of manufacturing equipment, and coating stages and operations. Generative design may further provide test and quality results for the battery at the end of the manufacturing line, on the line, at installation, during use / aging, and during recycling.

[0008] Generative design can advantageously combine advanced data science, machine learning, cheminformatics-materialinformatics, and structure-based modeling to explore and optimize chemical, constituent, and process spaces. Exemplary embodiments may automate virtual generation, testing, and selection of novel components, their quantities during formulation, and may incorporate process-level parameters. The goals of generative design may not only reduce the volume and cost of physical testing, but, in non-limiting examples, may also be to perform trade studies between system variables.

[0009] According to an exemplary embodiment, a computer implementation method for generative design of an energy storage device includes automatically constructing at least one model of the energy storage device. The construction may be based on a design parameter space and may employ a machine learning process. The computer implementation method further includes automatically performing a simulation of the energy storage device. The simulation may employ a design parameter space, a design evaluation space, and the constructed at least one model. The execution may include generating at least one prediction of the energy storage device that achieves at least one product design objective or at least one model design objective achieved by the constructed at least one model.

[0010] The computer implementation method further includes automatically evolving at least one of (i) the design parameter space and (ii) the design evaluation space. The evolution may be based on at least one generated prediction, and may employ a machine learning process. If at least one prediction indicates that at least one product design objective has been achieved, or at least one model design objective has been achieved, the computer implementation method further includes automatically converging on the evolved design parameter space to complete the generative design of the energy storage device, and otherwise repeating the build, run, and evolve process.

[0011] The design parameter space and the design evaluation space may be associated with at least one of the energy storage device and at least one constructed model. The design parameter space may include variables. The variables may include material variables, system variables, or combinations thereof. The variables may be associated with the chemical space of the energy storage device, the formulation space of the energy storage device, the material space of the energy storage device, the configuration space of the energy storage device, the process space of the energy storage device, or combinations thereof.

[0012] The design evaluation space may include at least one test of the energy storage device, e.g., a real-world test, and performing a simulation may include simulating at least one test using at least one constructed model. A non-limiting embodiment of such a real-world test may identify the onset of cell failure in the energy storage device by repeatedly measuring the charge state over numerous charge cycles at different charge rates. For each design space, such measurements can be used to validate the model and predict pre-production results. This also allows for the determination of better process and material specifications, as well as modifications to the design, physical configuration, or recipe of the energy device's cell components.

[0013] At least one product design objective may include at least one user-specified criterion associated with the energy storage device, at least one machine-generated criterion associated with the energy storage device, or a combination thereof. At least one product design objective may include a target product profile (TPP). At least one model design objective may include at least one error threshold associated with the difference between simulated measurements and real-world measurements, a target size relative to the design parameter space, or a combination thereof.

[0014] Evolution may include (a) cutting out the design parameter space, (b) cutting out the design evaluation space, (c) expanding the design parameter space, (d) expanding the design evaluation space, or (e) a combination of (a) to (d).

[0015] Evolution may include maintaining diversity within the design parameter space. Maintaining diversity may include employing clustering methods, Pareto methods, or a combination thereof.

[0016] The computer implementation method may further include adopting at least one monitored parameter in building, evolving, or a combination thereof. The at least one monitored parameter may be adopted in at least one iteration of building, evolving, or a combination thereof. The at least one monitored parameter may represent at least one real-world outcome generated through at least one real-world experiment employing the energy storage device. The real-world outcome may be carried out in at least one real-world experiment based on the adoption of the evolved design parameter space.

[0017] The computer implementation method may further include employing generative adversarial networks (GANs), deep neural networks (DNNs), Bayesian optimization (BAOs), genetic function approximation methods, or a combination thereof in the machine learning process.

[0018] The design parameter space may include semantically structured real-world evidence (RWE) data associated with experiments on energy storage devices.

[0019] The computer implementation method may further include automatically storing in a database the constructed model, associated with at least one generated prediction, at least one input to the constructed model, and at least one output from the constructed model. Running the simulation may include inputting at least one input to the constructed model and generating at least one output from the constructed model in response to the at least one input.

[0020] The energy storage device may be a battery. Converging on an evolved design parameter space may involve identifying at least one of the compounds, components, additives, formulations, recipes, or combinations thereof that enable the achievement of at least one product design objective of the energy storage device.

[0021] According to another exemplary embodiment, a computer-based system for generative design of an energy storage device may comprise at least one memory and at least one processor coupled to at least one memory. The at least one processor may be configured to automatically construct at least one model of the energy storage device based on a design parameter space. Constructing may employ a machine learning process. The at least one processor may be further configured to automatically perform a simulation of the energy storage device. The simulation may employ a design parameter space, a design evaluation space, and at least one constructed model. Performing the simulation may include generating at least one prediction of the energy storage device that achieves at least one product design objective or at least one model design objective achieved by the constructed model.

[0022] At least one processor may be further configured to automatically evolve at least one of (i) the design parameter space and (ii) the design evaluation space. The evolution may be based on at least one prediction generated and may employ a machine learning process. If at least one prediction indicates that at least one product design objective has been achieved or at least one model design objective has been achieved, at least one processor may be further configured to automatically converge on the evolved design parameter space, thereby completing the generative design of the energy storage device, and otherwise iterate through building, running, and evolving.

[0023] Alternative computer-based system embodiments are similar to those described above in relation to embodiments of the exemplary method.

[0024] In yet another exemplary embodiment, a non-temporal computer-readable medium for the generative design of an energy storage device may be loaded by and executed by at least one processor, and may encode a sequence of instructions thereon that causes at least one processor to automatically construct at least one model of the energy storage device based on the design parameter space. The construction may employ a machine learning process.

[0025] The instruction sequence may further cause at least one processor to automatically perform a simulation of the energy storage device. The simulation may employ a design parameter space, a design evaluation space, and at least one constructed model. Performing the simulation may include generating at least one prediction of the energy storage device that achieves at least one product design objective or at least one model constructed that achieves at least one model design objective. The instruction sequence may further cause at least one processor to automatically evolve at least one of (i) the design parameter space and (ii) the design evaluation space.

[0026] Evolution may be based on at least one generated prediction or may employ a machine learning process. If at least one prediction indicates that at least one product design objective has been achieved or at least one model design objective has been achieved, the sequence of instructions may further cause at least one processor to automatically converge on an evolved design parameter space, thereby completing the generative design of the energy storage device, and, if not, may cause at least one processor to repeat constructing, executing, and evolving.

[0027] Alternative embodiments of the non-transitory computer-readable medium are similar to those described above in connection with the exemplary method embodiments.

[0028] Of course, the exemplary embodiments disclosed herein can be implemented in any combination and in any form of a method, an apparatus, a system, or a non-transitory computer-readable medium having program code embodied thereon. [[ENDEND]]END]]

[0029] The foregoing will be apparent from the following more particular description of exemplary embodiments, as illustrated in the accompanying drawings, wherein like reference numerals refer to the same parts throughout the different figures. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] [Figure 1A] FIG. 1A is a block diagram of an exemplary embodiment of a computing environment in which a materials scientist is using a computer-based system for the generative design of an energy storage device. [Figure 1B] FIG. 1B is a block diagram of an exemplary embodiment of a generative process. [Figure 1C]Figure 1C is a block diagram of an exemplary embodiment of an automated molecular design environment. [Figure 1D] Figure 1D is a flowchart illustrating an exemplary embodiment of a computer-implemented generative method. [Figure 2] Figure 2 is a block diagram of an exemplary embodiment of a computer-based system for generative design of energy storage devices. [Figure 3] Figure 3 is a block diagram of an exemplary embodiment of the design parameter space. [Figure 4] Figure 4 is a flowchart illustrating an exemplary embodiment of a computer implementation method for generative design of an energy storage device. [Figure 5] Figure 5 is a block diagram of an embodiment of the internal structure of a computer, in which various embodiments of the present disclosure may be implemented. [Modes for carrying out the invention]

[0031] An example of an embodiment is described below.

[0032] The exemplary embodiments disclosed herein may, in non-limiting examples, relate to generative systems and methods for designing compounds, components, additives, formulations, and recipes for use in energy storage devices, and, in non-limiting examples, for designing models of such energy storage devices, such as batteries. The exemplary embodiments may, in general, in non-limiting examples, relate to material modeling and simulation of complex material-based systems, such as batteries. The exemplary embodiments disclosed herein may also be applicable to other areas where chemistry, materials, nano, and domain microstructures may be used to predict and understand the performance and behavior of products. The exemplary embodiments disclosed herein may also be applicable, in non-limiting examples, to systems of general chemical reactivity, such as fuel cells and chemical process effects.

[0033] Generative materials design (GMD) is essentially the automation of exploring a model-based variable space. Variables may be considered as chemistry (e.g., structure, isomers, toxicity, substitution, and sequence), material (e.g., phase, polymorphism, insertion, spacing), formulation (e.g., quantity, purity, grade), and process (e.g., steps, action, sequence, conditions, parameters). Variables may be specified during the design phase, and the system may automate the construction of samples. These samples may then be evaluated using the next steps identified through model-based and evolutionary approaches.

[0034] Exemplary embodiments disclosed herein, in non-limiting examples, provide an efficient and automated method for determining optimal conditions in a characteristic space from material and / or system variables. Exemplary embodiments of computer-based systems disclosed herein, in non-limiting examples, may use virtual cycles, in which the system can explore a characteristic space for an energy storage device such as a battery by “learning” from the results of real-world experiments.

[0035] Exemplary embodiments of computer-based systems disclosed herein may, in non-limiting examples, use a combination of machine learning models and structure-based modeling and simulation methods to virtually screen and optimize candidate compounds. Multi-objective optimization methods may be employed to balance competing objectives and enable product designers to generate compounds that are improved toward a design or target product profile (TPP).

[0036] In the “actual” cycle phase, such product designers may synthesize and test the most promising hypothetical compounds in a laboratory setting. New data generated through such testing may be input into a computer-based system to improve predictive models and, in non-limiting examples, to refine exploration of the chemical space. Such execution of real-world experiments may be useful for generating new data used to improve and optimize the models. According to the exemplary embodiment, such virtual (V) and real (R), i.e., V+R, active learning cycles may continue until the product designer identifies a compound that satisfies the TPP, in non-limiting examples.

[0037] The exemplary embodiments disclosed herein may, in non-limiting examples, include computer methods that employ an adversarial deep neural computing approach using an iterative and probabilistic process to obtain a realistic computer model of a battery. In non-limiting examples, exemplary embodiments of computer-based systems or computer implementations may link the following four areas of science and modeling: 1. Quantitative Structure-Activity Relationship Modeling (QSAR) - QSAR modeling of empirical data concerning quality, chemistry, and material systems can be performed using established and validated machine learning methods. These yield highly predictive and scalable estimates of performance / activity for proposed chemistrys and mixtures. 2. Early attention to balancing all the attributes required for a TPP-chemical system to progress from the innovation stage to the pilot-scale and manufacturing stages of a new product may involve employing multi-parameter optimization at all stages (e.g., safety, manufacturability, novelty, etc.). 3. Active Learning (AL) – This special case of machine learning closely combines the real world with virtual activities, allowing the learning method to query the user (or another source of information) to label new data points with desired outputs. In GMD, this means intentionally synthesizing mixtures and recipes, expanding the domain of applicability of machine learning models. 4. Generative Chemistry – As a non-limiting example, while it is impossible with reasonable resources to enumerate all chemicals that could have a significant impact on the development of battery electrode coatings or electrolytes, chemists can explore large areas of the chemical space through iterative modifications of starting chemistry or formulation sequences induced by property-based multidimensional fitness functions.

[0038] The GMD disclosed herein may then, using evolutionary methods, drive material-based systems, such as batteries, to maximum synthesizability / processability, as well as maximum desirability, in non-limiting examples. Such processes may, in non-limiting examples, generate millions of options over multiple cycles, which can be automatically trimmed against user-defined criteria to ultimately obtain a small optimal set of options for a set of solutions, according to the exemplary embodiments. According to the exemplary embodiments, clustering or Pareto methods may be employed to maintain diversity and ensure that the set of solutions is not a substantially identical group of materials with only slight variations from one another. After the optimization, i.e., “virtual experiment” or simulation, is complete, the materials scientist may, in non-limiting examples, observe the convergence in a chart(s) to decide whether to synthesize and test some of the resulting structures. Alternatively, the materials scientist may adjust the optimization parameters and run another run, i.e., another virtual experiment / simulation. Exemplary embodiments of computer-based systems that may be used by such materials scientists are further disclosed below with respect to Figures 1A and 2.

[0039] Exemplary embodiments of computer-based systems, in non-limiting examples, may allow users such as materials scientists to advantageously perform the following: • Use the desirability profile to define the objectives for each optimization. • Graphically adjust each desirability profile based on probabilities determined by applying a machine learning model to known compounds. • Increase the "weight" of the most important objective (e.g., activity) to help ensure that these objectives are achieved during optimization. As a non-limiting example, specify the overall goals of optimization, such as whether immediate improvement of the compound or exploration of the chemical space should be used to improve the model, whether compound improvement functions should be applied to improve product production performance and quality indicators, and whether a series of low-scale models should be employed to enable the development of a better, more accurate, or more robust model.

[0040] Exemplary embodiments of computer-based systems or computer implementation methods disclosed herein may, as non-limiting examples, be used at two levels: firstly, in a closed-loop approach, to enrich the forms used in feature generation; and secondly, in a higher level of subsequent modeling, to use an abstraction of an energy storage device in the form of a model that has been identified. Exemplary embodiments disclosed herein may be used automatically with no or minimal user input, which enables rapid exploration of the design parameter space and autonomous model development as new data is accumulated from test or production systems, resulting in an improved user experience due to the availability of multiple models. An advantage of this approach is that it incorporates the synthesizability of compounds. Enumeration may also be based on Markush structures having a core molecule with a reaction-based (R) group position and a set of binding fragments, as non-limiting examples. Such approaches may be employed in computer-based systems such as computer-based system 104 in Figure 1A disclosed below.

[0041] Figure 1A is a block diagram of an exemplary embodiment of a computing environment 100 in which a materials scientist 102 is using a computer-based system 104 for the generative design of an energy storage device 106. A materials scientist is a person who studies the structural and chemical properties of various materials in order to develop new products or enhance existing ones. Naturally, the computer-based system 104 is not limited to being used by a materials scientist. Furthermore, in the exemplary embodiment of Figure 1A, as a non-limiting example, the energy storage device 106 is a battery. The exemplary embodiment of Figure 1A may describe the energy storage device 106 as a battery, more specifically as a lithium-ion battery, but naturally, the exemplary embodiments disclosed herein are not limited to a battery or a lithium-ion battery.

[0042] In the exemplary embodiment shown in Figure 1A, a materials scientist 102 interacts with two user interfaces provided by a computer-based system 104, namely a first user interface (UI) 103 and a second UI 105. In one computer operating system, user interfaces 103 and 105 are implemented as so-called windows. In a global network (Internet) server-based computer system, user interfaces 103 and 105 may be implemented as different tabs, screen views, or similar. The first UI 103 and the second UI 105 are presented on the display screen 107 of the computer-based system 104 for user interaction. Naturally, the computer-based system 104 is not limited to providing two user interfaces (UIs).

[0043] In the exemplary embodiment, the energy storage device 106 is a lithium-ion battery. While lithium-ion batteries can last for many years, they sometimes exhibit rapid, nonlinear degradation that significantly limits their lifespan. In the exemplary embodiment of Figure 1A, a first UI 103 includes a first simulated representation 109 at the beginning of the lifespan of a slice of the lithium-ion battery's "jelly roll," and a second UI 105 includes a second simulated representation 111 after a cycle life simulation of the slice of the jelly roll. In the second simulated representation 111 of the slice, a deformation 113 is formed such that the jelly roll has deformed inward toward the core 115 (i.e., the center) of the jelly roll. Such deformation 113 may be attributed to a simulated increase in the internal pressure of the jelly roll during the cycle life simulation. A materials scientist 102 may employ a computer-based system 104 to perform GMD of such a lithium-ion battery to generate a design that avoids such deformation 113, as a non-limiting example.

[0044] According to an exemplary embodiment, the computer-based system 104 may be configured as follows, in a non-limiting example: Exploring chemical spaces • Includes reaction-based / R-group-based enumeration, • Filter out hypothetical compounds for undesirable substructures based on universal, company, and project-specific criteria. • Incorporate design preferences based on customized battery component profiles for material function domains, and Prioritize based on synthesizability.

[0045] Generative Modulation (GMD) offers a variety of methods for generating virtual compounds. Evolutionary methods may generate molecules for each input molecule using a variety of molecular transformations, including matched molecular pairs (MMPs), ring assembly substitution, molecular scaffolds, molecular variations (splitting, trimming, extension, rearrangement, crossover), and molecular mutations (changes in atomic and bond types, addition and deletion of atoms, ring opening and closing). There is an option to bias generative methods to generate compounds similar to a target structure or those containing a particular scaffold.

[0046] One of the challenges associated with lead compound searching is understanding the trade-offs to make when optimizing multiple design criteria. GMD enables materials scientists, chemists, etc., to design improved lead compounds by maximizing an overall desirability function based on individual desirability profiles in TPP. For each objective in TPP, the materials scientist 102 may use the profile to specify how "good" is considered "good enough" to achieve the project's objectives. GMD may be used to assist in this process, for example, via a display screen 107, to display the model predictive distribution and empirical probability values ​​(e.g., the probability that the compound will be active if the model score exceeds a certain value). Alternatively, GMD may be used to provide a reasonable default desirability profile based on TPP.

[0047] Instant access to both positive and negative data enables scientists, such as materials scientists102, to enhance predictive models and track compound lineages from conception to development. This “digital continuity” is useful for tracking intellectual property (IP) generated within discovery organizations. An exemplary embodiment of the generative process that enables this is described below with respect to Figure 1B.

[0048] Figure 1B is a block diagram of an exemplary embodiment of the generative process 150. The generative process 150 may include a data entry phase 152, an ideation phase 154, a prototype phase 156, a test phase 158, and an evaluation phase 160. The generative process 150 may also perform an iterative design 162 (returning to the ideation phase 154 based on the evaluation phase 160), which ultimately leads to a product design 164 based on the results of the evaluation phase 160. The generative process 150 may be employed as part of an automated molecular design environment, such as the environment 170 disclosed below with respect to Figure 1C.

[0049] Figure 1C is a block diagram of an exemplary embodiment of the automated molecular design environment 170. The automated molecular design environment 170 includes data 172, tools 174, and a client 176. In a non-limiting example, the data 172 may include assays 172-1, candidates 172-2, inputs 172-3, and / or history 172-4. The tools 174 may include data access 174-1, data preparation and exploration 174-2, machine learning methods 174-3, and / or delivery, integration, and deployment 174-5.

[0050] Machine learning methods 174-3 may include, as non-limiting examples, performance and scalability methods 175-1, automation methods 175-2, active learning methods 175-3, and / or multi-parameter optimization methods 175-4. Client 176 may include, as non-limiting examples, optimization workbenches 176-1, project management 176-2, model management 176-3, collaboration and analysis 176-4, and / or visual programming environments 176-5. Such automated molecular design environments 170 may be contained within the computing environment 100 of Figure 1A, which is disclosed above and further disclosed below. Chemical and materials design approaches may include the design of anti-aging additives in battery electrolytes. Given numerous competing design requirements, the ability to screen a large chemical space down to a human-understandable set and then prioritize that set for either performance or model development goals can be achieved by artificial intelligence. In addition, competing requirements can be evaluated using such models to achieve a balance between competing needs.

[0051] Referring to Figures 1A and 1C, the automated molecular design environment 170 represents the components of the computer-based system 104 according to an exemplary embodiment. As shown in Figure 1C, these components include data sources and real-world evidence, i.e., data 172, assays 172-1, candidates 172-2, inputs 172-3, and history 172-4. The components further include tools 174, which are used to access, transform, materialize, and analyze these various data sources 172. The components further include a consuming system, called a client technology or simply a client 176. The client 176 can interact with the tools 174, its derivatives, models about them, and native data 172. This framework, i.e., the automated molecular design environment 170, is far more than just the sum of its components. The environment 170 implicitly includes not only the lifecycle of entities but also a state and message framework within a multi-user / multi-role environment identified by the data 172, tools 174, and client 176.

[0052] Returning to Figures 1A and 1B, the generative process 150 may be performed by a computer-based system 104, which may perform a generative method disclosed below with respect to Figure 1D, for example.

[0053] Figure 1D is a flowchart of an exemplary embodiment of a computer-implemented generative method (180). The generative method (180) starts (182), (1) collects and semantically sorts data, (2) builds a model of a characteristic of interest, such as open-circuit cell voltage, flammability, capacity, maximum cycle life, and maximum charge current (C rate) during discharge and / or charge, for example, a product design 164 which is for a battery, (3) performs a design space experimental design across a set of variables using the model as a predictor, (4) stores in a database, (5) filters on user or defined machine criteria, (6) selects the best candidate and confirms such selection, and (7) if such selection (6) does not satisfy at least one criterion, the method may then be iterated from (2) or, as convergence occurs (e.g., the selection satisfies the at least one criterion), the method then terminates in an exemplary embodiment (184). At least one criterion may be associated with at least one product design objective or at least one model design objective, as is further disclosed below with respect to Figure 4, for example, in non-limiting examples. In one embodiment, at least one criterion may be set by the user or may be set automatically based on requirements.

[0054] Continuing to refer to Figure 1D, the alignment of data in (1), the model building based on real-world evidence in (2), and the optimization of the design space for either characteristic / performance targets or improvements in model building in (3) lead to a better virtual evaluation in evaluation stage 160. The storage action taken in (4) above provides storage that inspires or enables iteration into new ideas and limiting the design space to a manageable set of tests and evaluations in ideation stage 154. Prototyping stages 156 and testing stages 158 may be real-world actions that provide results that are input into evaluation stage 160, which is performed by the computer-based system 104, and by returning to ideation stage 154, the computer-based system 104 is made to perform iterative design 162. Further details regarding the actions (1) to (7) disclosed above are provided below. These actions may also be called generative actions and may be performed by the computer-based system 104 and the computer-based system 204 in Figure 2, which is further disclosed below.

[0055] Generative methods

[0056] (1) Collect data and sort it semantically.

[0057] Data for devices such as batteries may come from numerous different levels. As a non-limiting example, data may come from the subcell chemical level (e.g., atomic structure, quantum electronic properties) and the material level (e.g., not only purity, grade, structure, and / or phase, but also thermodynamic state and dynamic or reactive behavior, chemical decomposition pathways, and / or other information). Data may further come from mesoscopic information such as particle size, particle distribution, structure, or layered arrangement in the electrodes, and the distribution structure of the material, as well as porosity, flexibility, and permeability on the electrodes (i.e., separator, anode, and / or cathode). In addition, numerous material physical characteristics, such as pore size distribution on the electrodes and the effects of processing (e.g., calendering, winding, and compression or stacking), can directly affect the cell's performance.

[0058] At the cell level, information may come not only from the geometric shape of the cells, e.g., cylinder vs. pouch vs. prism vs. blade, but also, in non-limiting examples, from their size (e.g., in the case of cylinders, there are different dimensions such as the diameter and height of the cell (e.g., 18×65, 23×70, or 46×80)). The characteristics of these cells and subcells are manufactured, and therefore, in non-limiting examples, there are secondary attributes around the ambient conditions of manufacture, and the processes or equipment used for coating the electrodes, the geometric shape of the press for calendering, and the stacking configuration of the "jelly roll" or pouch. Once the battery is assembled, in its first "forming" step, it is tested to fabricate the solid electrolyte interface, and then tested for repeated cycle performance over its life at different charging speeds and different test temperatures.

[0059] The semantic alignment of this data is intended to identify not only the context (e.g., which tests relate to which cells), but also which processes passed through those cells and from which batches of material those cells were constructed. This alignment provides connections for material characteristics at the raw material, subcell, cell, and test levels.

[0060] The complexity of what constitutes a batch vs. sample vs. test vs. experiment vs. quality gate depends not only on simple linguistic semantics but also on the interconversion of time and length scales (from nanoseconds to hours, or cycles), the connection of the design to the model and source data (master data alignment), and the persistence of requirements and validation. Exemplary embodiments, as non-limiting examples, enable the alignment of chemical, material, and test indices for both virtual and real data within a unified data model that incorporates logical, functional, and physical requirements, and results.

[0061] In action (2) above, generative method 180 constructs, as a non-limiting example, a model of a characteristic of interest, such as open-circuit cell voltage, flammability, capacity, maximum cycle life, and maximum charge current (C rate) during discharge and / or charge. Generative method 180 may also construct, as a non-limiting example, a model that estimates from a hybrid of atomic / material and in-use data, such as telemetry, life, and aging behavior of the battery. Generative method 180 may also construct a model that represents the expected degree of variance in production quality metrics of the battery in volume and scale-up production.

[0062] Once data from testing is integrated into the system, models can be constructed using full three-dimensional fidelity, enabling high-fidelity monitoring of processes and changes occurring within storage devices (e.g., batteries) and across different parts of the storage device's lifecycle. Such models may be pyramids constructed from physical characteristic-based features, geometric shape-based features, process-based features, and test-based data. Furthermore, as additional data is generated, these models may be versioned and subsequently incremented against their preceding models, and evaluated for the accuracy and error prediction of their values.

[0063] In action (3) above, the computer-based system 104 may use a model as a predictor and perform design space experimental design across variables using the model as an imitation for physical testing, and may use Monte Carlo and evolutionary approaches to identify the optimal solution for the design space requirements.

[0064] In action (4) above, the computer-based system 104 may store the data in a database (not shown). As a non-limiting example, the computer-based system 104 may store models, their versions, and their inputs / outputs / parameters / components in the database for both tracking and tracing activities, as well as to build increased storage of knowledge and predictions. Since simulations may be rapid and may perform screening and validation of complex solutions, the ability to make approximate trade-offs at lower resolutions is useful for identifying domains of interest in the design space (e.g., chemistry, recipes, and processes).

[0065] In action (5) above, the computer-based system 104 may filter by user or defined machine criteria. By filtering and sorting the output of the models in the system, the user is provided with the ability to perform Pareto-based optimization within the loop, enabling the generation of additional values ​​and sub-scenario analysis of the generated data. This is useful in the post-base model building stage of a hybrid or adversarial training process.

[0066] In action (6), the computer-based system 104 may select the best candidate and verify the design. Real-world evidence of the prediction may be created and linked to experimental data. Deviations in the model prediction may be linked to the model version, and a retraining process may be automatically initiated to improve and optimize the model in a cyclical manner.

[0067] In action (7), the computer-based system 104 may be iterative. In this approach, useful elements are integrated data and the cyclicality of the process. The data is generated from both experiments (reality) and models (virtual). As the computer-based system 104 builds the models and data within its framework, the balance between reality and virtuality shifts. The value is that the specifications and limits of materials, processes, and innovations in chemistry can be virtually screened and optimized without requiring continuous testing, and can be verified at the end of the cycle. Figure 2 disclosed below is a block diagram of an exemplary embodiment of the computer-based system 204, which may be adopted as the computer-based system 104 disclosed above and which may be adopted to carry out actions (1) to (7).

[0068] Figure 2 is a block diagram of an exemplary embodiment of a computer-based system 204 for the generative design of an energy storage device 206. The computer-based system 204 may be adopted as the computer-based system 104 of Figure 1A disclosed above. In non-limiting examples, the computer-based system 204 comprises at least one memory 208 and at least one processor 218 coupled to the memory 208, such as a central processing unit 518 coupled to the memory 508, which is further disclosed below with respect to Figure 5.

[0069] Continuing to refer to Figure 2, at least one processor 218 of the computer-based system 204 is configured to automatically construct at least one model 212 of the energy storage device 206 based on the design parameter space 210. Such construction may employ a machine learning process 214. At least one processor 218 may be further configured to automatically perform a simulation 216 of the energy storage device 206. The simulation 216 may employ the design parameter space 210, the design evaluation space 220, and the constructed at least one model 212. The execution 216 may include generating at least one prediction 222 of the energy storage device 206 that achieves at least one product design objective (not shown) or at least one model design objective (not shown) and achieves the constructed at least one model 212.

[0070] As described above, building at least one model 212 may involve employing a machine learning process 214. Machine learning can be used in several stages of model building and evaluation. The model effectively maps from the parameter space to the characteristic space and includes the generation of features. Features may include real-world measurements, and in non-limiting examples, a collection, derivatives, or transformations of such real-world experiments. In non-limiting examples, signals measured per second may be averaged over several days or weeks for long-term trend models. The extension of real-world evidence with derived or generated features can be performed using the model. Furthermore, when a large descriptor space is thus generated, the selection of key variables that actually map to the characteristics of interest can be performed using the model. Naturally, the evaluation of the model, i.e., its construction, can be performed by machine methods. Some of these evaluations may include the model form, the functions used in the model, the model domain of applicability, the model error function, and other features or metrics. Furthermore, the evaluation of the model against optimal conditions or trade-off regions can be performed using a system of multiple models, consensus models, adversarial models, or even system models.

[0071] At least one processor 218 may be further configured to automatically evolve at least one of (i) the design parameter space 210 and (ii) the design evaluation space 220. The evolution may be based on at least one prediction 222 generated, and may employ a machine learning process 214. If at least one prediction 222 indicates that at least one product design objective (e.g., longer lifespan, lower weight, lower cost) has been achieved, or at least one model design objective (e.g., accuracy, range) has been achieved, at least one processor 218 may be further configured to automatically converge on the evolved design parameter space 210, thereby completing the generative design 224 of the energy storage device 206, and otherwise repeating the build, run, and evolve process.

[0072] In numerous real-world examples, the accuracy and applicability of a model are influenced by constraints on the design space on a manageable set of axes, or by the choice of forms and functions used within the model. As a non-restrictive example, to model the gradient of performance, a simple straight line linking two points can be used, which can then evolve to present more information in a polynomial fit, and then further evolve to present even more information in a discontinuous surface function.

[0073] Alternatively, the number of similar terms may increase; for example, a linear model y=f(x) may evolve into a quadratic model y²=f(x)+f(y)+f(x*y) using some level of cross-variable synergy. As the number of variables increases to four or more, the number of combinations of x, y, z, and w increases to 14, and an evolutionary process (i.e., evolution) may be used to identify which set or combination of these variables is best for the known data, i.e., best fits the known data. In the evolutionary process, the variables and their functions, polynomials, or nonlinearities may be described by coded patterns. The patterns may be ranked against the data, and then a random crossover of pairs may be performed to form a second generation. The second generation may be ranked and compared against the first generation, and the best set may advance for the next complete iteration of function evolution. The advantages of such evolution may be realized in terms of both the range and breadth of fitting functions that the system can handle, as well as the undiscovered discoveries it can reveal. This identification of key variables and their combinations reduces the dimensionality of the design space, as well as allows for prioritization of virtual and real-world testing.

[0074] At least one Model 212 may include, as a non-limiting example, a material model, a mixture and recipe model, a configuration and design model, a process and production model, a use and end-of-life model, or a combination thereof.

[0075] At least one processor 218 may implement a machine learning process 214, which may be further configured to employ a generative adversarial network (GAN), a deep neural network (DNN), Bayesian optimization (BAO), a genetic function approximation method, or a combination thereof.

[0076] The design evaluation space 220 may include at least one test (not shown) of the energy storage device 206. To run the simulation 216, at least one processor 218 may be further configured to simulate at least one test using at least one constructed model 212.

[0077] At least one product design objective may include at least one user-specified criterion (not shown) associated with the energy storage device 206, at least one machine-generated criterion (not shown) associated with the energy storage device 206, or a combination thereof. At least one product design objective may further include TPP.

[0078] At least one model design objective may include at least one error threshold associated with the difference between simulated measurements and real-world measurements, the target size relative to the design parameter space 210, or a combination thereof.

[0079] The design parameter space 210 may include semantically structured real-world evidence (RWE) data (not shown) associated with experiments of the energy storage device 206. The design parameter space 210 and the design evaluation space 220 may be associated with the energy storage device 206 and at least one of the constructed models 212. The design parameter space 210 may include variables, for example, disclosed with respect to Figure 3 below.

[0080] Figure 3 is a block diagram of an exemplary embodiment of the design parameter space 310, which may be adopted as the design parameter space 210 of Figure 2 disclosed above. In the exemplary embodiment of Figure 3, the design parameter space 310 includes variables 330. Referring to Figures 2 and 3, the variables 330 may include, as non-limiting examples, material variables (not shown), system variables (not shown), or a combination thereof. The variables 330 may be associated with the chemical space 332 of the energy storage device 206, the formulation space 334 of the energy storage device 206, the material space 336 of the energy storage device 206, the configuration space 338 of the energy storage device 206, the process space 340 of the energy storage device 206, or a combination thereof.

[0081] Referring back to Figure 2, in order for the computer-based system 204 to automatically evolve the design parameter space 210, the design evaluation space 220, or a combination thereof, at least one processor 218 may further be configured to (a) cut off the design parameter space 210, (b) cut off the design evaluation space 220, (c) expand the design parameter space 210, (d) expand the design evaluation space 220, or (e) a combination of (a) to (d). At least one processor 218 may further be configured, as a non-limiting example, to maintain diversity within the design parameter space 210 by employing a clustering method, a Pareto method, or a combination thereof.

[0082] At least one processor 218 may be further configured to employ at least one monitored parameter (not shown) in construction, evolution, or a combination thereof. The at least one monitored parameter may be employed in at least one iteration of construction, evolution, or a combination thereof. The at least one monitored parameter may represent at least one real-world result generated through at least one real-world experiment employing the energy storage device 206. The real-world result may be implemented in at least one real-world experiment through the use of the evolved design parameter space 210.

[0083] At least one processor 218 may be further configured to automatically store in at least one memory 208 at least one constructed model 212 associated with at least one generated prediction 222, at least one input (not shown) to the constructed model 212, and at least one output (not shown) from the constructed model 212. To run the simulation 216, at least one processor 218 may be further configured to input at least one input to the constructed model 212. The constructed model 212 may be configured to produce at least one output in response to at least one input.

[0084] In a non-limiting embodiment, the energy storage device 206 may be a battery, for example, as disclosed above with respect to Figure 1A. To converge on an evolved design parameter space 210, at least one processor 218 may be further configured to identify at least one of compounds, components, additives, formulations, recipes, or combinations thereof that enable the achievement of at least one product design objective of the energy storage device 206. Referring back to Figure 1A, in a non-limiting example, at least one product design objective may be that no deformation 113 is formed in the center 115 of the jelly roll of the energy storage device 106 over a specified life cycle. The generative design of the energy storage device 106 may be implemented via a computer implementation method, for example, as disclosed below with respect to Figure 4.

[0085] Figure 4 is a flowchart of an exemplary embodiment of a computer implementation method (400) for generative design of an energy storage device. The method starts (402) and automatically constructs (404) at least one model of the energy storage device. Construction (404) may be based on a design parameter space or may employ a machine learning process. The computer implementation method automatically runs a simulation of the energy storage device (406). The simulation may employ a design parameter space, a design evaluation space, and at least one constructed model. Execution (406) may include generating at least one prediction of the energy storage device that achieves at least one product design objective or at least one model constructed to achieve at least one model design objective. The computer implementation method automatically evolves (408) at least one of (i) the design parameter space and (ii) the design evaluation space. Evolution (408) may be based on at least one generated prediction or may employ a machine learning process. The computer implementation method checks whether at least one prediction indicates that at least one product design objective has been achieved or at least one model design objective has been achieved (410). If at least one prediction indicates that at least one product design objective has been achieved or at least one model design objective has been achieved, the computer implementation method automatically converges on the evolved design parameter space (412), thereby completing the generative design of the energy storage device, and in the exemplary embodiment, the computer implementation method then terminates (414). Otherwise, the computer implementation method repeats the build (404), run (406), and evolve (408) disclosed above.

[0086] Evolution (408) may include (a) cutting out the design parameter space, (b) cutting out the design evaluation space, (c) extending the design parameter space, (d) extending the design evaluation space, or (e) a combination of (a) to (d). Evolution (408) may also include maintaining diversity within the design parameter space. Maintaining diversity may include employing clustering methods, Pareto methods, or a combination thereof.

[0087] The computer implementation method (400) may further include adopting at least one monitored parameter in construction, evolution, or a combination thereof. The at least one monitored parameter may be adopted in at least one iteration of construction (404), evolution (408), or a combination thereof. The at least one monitored parameter may represent at least one real-world result generated through at least one real-world experiment employing an energy storage device. The real-world result may be carried out in at least one real-world experiment based on the adoption of the evolved design parameter space.

[0088] The computer implementation method (400) may further include employing GANs, DNNs, BAOs, genetic function approximation methods, or combinations thereof in the machine learning process.

[0089] The computer implementation method (400) may further include automatically storing in a database the constructed model, associated with at least one generated prediction, at least one input to the constructed model, and at least one output from the constructed model. Running the simulation may include inputting at least one input to the constructed model and generating at least one output from the constructed model in response to the at least one input.

[0090] The energy storage device may be a battery. Convergence on the evolved design parameter space (412) may include identifying at least one of compounds, components, additives, formulations, recipes, or combinations thereof that enable the achievement of at least one product design objective of the energy storage device. The computer implementation method (400) may, as a non-limiting example, be implemented via a computer having an internal structure, for example, disclosed below.

[0091] Figure 5 is a block diagram of an embodiment of the internal structure of a computer 500, in which various embodiments of the present disclosure may be implemented. The computer 500 includes a system bus 502, which is a set of hardware lines used for data transfer between components of a computer or digital processing system. The system bus 502 is essentially a shared conduit that connects different elements of a computer system (e.g., processor, disk storage, memory, input / output ports, network ports, etc.) enabling information transfer between elements. An I / O device interface 504 is coupled to the system bus 502 to connect various input and output devices (e.g., keyboard, mouse, display monitor, printer, speaker, microphone, etc.) to the computer 500. A network interface 506 enables the computer 500 to connect to various other devices attached to a network (e.g., a global computer network, wide area network, local area network, etc.). Memory 508 provides volatile or non-volatile storage for computer software instructions 510 and data 512, which may be used to implement embodiments of the present disclosure (e.g., methods 150, 400), where volatile and non-volatile memory are embodiments of non-temporary media. Disk storage 513 also provides non-volatile storage for computer software instructions 510 and data 512, which may be used to implement embodiments of the present disclosure (e.g., method 400). Central processing unit 518 is also connected to system bus 502 and provides execution of computer instructions.

[0092] Further exemplary embodiments disclosed herein may be configured using computer program products, for example, control may be programmed in software to carry out the exemplary embodiments. Further exemplary embodiments may include non-temporary computer-readable media containing instructions which, when loaded and executed, cause the processor to complete the methods and techniques described herein. Naturally, elements of block diagrams and flow diagrams may be implemented in software or hardware, or their equivalents, firmware, combinations thereof, or other similar implementations to be determined in the future, for example, through one or more arrangements of the circuit of Figure 5 disclosed above.

[0093] In addition, the elements of the block diagrams and flow diagrams described herein may be combined or separated in any manner in software, hardware, or firmware. When implemented in software, the software may be written in any language capable of supporting the exemplary embodiments disclosed herein. The software may be stored in any form of computer-readable media, such as random access memory (RAM), read-only memory (ROM), or compact disk read-only memory (CD-ROM). In operation, a general-purpose or application-specific processor or processing core loads and executes the software in a manner well understood in the art. Naturally, the block diagrams and flow diagrams may also contain more or fewer elements, may be arranged or oriented differently, or may be represented differently. Naturally, implementations may indicate the execution of the embodiments disclosed herein by referring to the block diagrams, flow diagrams, and / or network diagrams, as well as the numbers of the block diagrams and flow diagrams.

[0094] While illustrative embodiments have been specifically shown and described, it will be understood by those skilled in the art that various modifications of form and detail may be made without departing from the scope of embodiments included in the appended claims.

Claims

1. A computer-based method for generative design of energy storage devices, The method involves automatically constructing at least one model of an energy storage device, wherein the construction is based on a design parameter space and employs a machine learning process. The automatic execution of a simulation of the energy storage device, wherein the simulation employs the design parameter space, the design evaluation space, and the constructed at least one model, and the execution of the simulation generates at least one prediction of the energy storage device that achieves at least one product design objective or at least one model design objective achieved by the constructed at least one model, (i) Automatically evolving the design parameter space and (ii) the design evaluation space, wherein the evolution is based on the generated at least one prediction, employs the machine learning process, and includes (a) trimming the design parameter space, (b) trimming the design evaluation space, (c) expanding the design parameter space, (d) expanding the design evaluation space, or (e) a combination of (a) to (d), In constructing, evolving, or a combination thereof, employing at least one monitored parameter, wherein the at least one monitored parameter represents at least one real-world result generated through at least one real-world experiment employing the energy storage device, and the real-world result is carried out based on employing the evolved design parameter space in the at least one real-world experiment. If the at least one prediction indicates that the at least one product design objective has been achieved or the at least one model design objective has been achieved, the generative design of the energy storage device is completed by automatically converging in the evolved design parameter space; and in response to the at least one prediction indicating that the at least one product design objective has not been achieved and the at least one model design objective has not been achieved, the build, execute, and evolve process is repeated. A method that includes this.

2. The method according to claim 1, wherein the design parameter space and the design evaluation space are associated with at least one of the energy storage device and the constructed at least one model, the design parameter space includes a plurality of variables, the plurality of variables include material variables, system variables, or a combination thereof, and the plurality of variables are associated with the chemical space of the energy storage device, the formulation space of the energy storage device, the material space of the energy storage device, the configuration space of the energy storage device, the process space of the energy storage device, or a combination thereof.

3. The method according to claim 1, wherein the design evaluation space includes testing at least one of the energy storage devices, and performing the simulation includes simulating the at least one test using the constructed at least one model.

4. The method according to claim 1, wherein the at least one product design objective includes at least one user-specified criterion associated with the energy storage device, at least one machine-generated criterion associated with the energy storage device, or a combination thereof, the at least one product design objective includes a target product profile (TPP), and the at least one model design objective includes at least one error threshold associated with the difference between simulated measurements and real-world measurements, a target size for the design parameter space, or a combination thereof.

5. The method according to claim 1, wherein the evolution includes maintaining diversity in the design parameter space, and the maintenance includes employing a clustering method, a Pareto method, or a combination thereof.

6. The method according to claim 1, further comprising employing a generative adversarial network (GAN), a deep neural network (DNN), Bayesian optimization (BAO), a genetic function approximation method, or a combination thereof in the machine learning process.

7. The method according to claim 1, wherein the design parameter space includes semantically structured real-world evidence (RWE) data associated with experiments on the energy storage device.

8. The method according to claim 1, further comprising automatically storing in a database the constructed at least one model, associated with the generated at least one prediction, at least one input to the constructed at least one model, and at least one output from the constructed at least one model, the method comprising running the simulation, further comprising inputting the at least one input to the constructed at least one model, and generating the at least one output from the constructed at least one model in response to the at least one input.

9. The method according to claim 1, comprising identifying at least one of compounds, components, additives, formulations, recipes, or combinations thereof that converge in the evolved design parameter space to achieve the at least one product design objective of the energy storage device, wherein the energy storage device is a battery.

10. A computer-based system for generative design of energy storage devices, At least one memory, At least one processor coupled to the at least one memory, the at least one processor is Based on the design parameter space, at least one model of an energy storage device is automatically constructed, and this construction employs a machine learning process. The process includes automatically performing a simulation of the energy storage device, wherein the simulation employs the design parameter space, the design evaluation space, and the constructed at least one model, and the execution generates at least one prediction of the energy storage device that achieves at least one product design objective or at least one model design objective achieved by the constructed at least one model, (i) Automatically evolve at least one of the design parameter space and (ii) the design evaluation space, wherein the evolution is based on the generated at least one prediction, employing the machine learning process, and including (a) cutting off the design parameter space, (b) cutting off the design evaluation space, (c) expanding the design parameter space, (d) expanding the design evaluation space, or (e) a combination of (a) to (d), In constructing, evolving, or a combination thereof, at least one monitored parameter is employed, the at least one monitored parameter represents at least one real-world result generated through at least one real-world experiment employing the energy storage device, and the real-world result is carried out based on employing the evolved design parameter space in the at least one real-world experiment. If the at least one prediction indicates that the at least one product design objective has been achieved or the at least one model design objective has been achieved, the generative design of the energy storage device is completed by automatically converging in the evolved design parameter space; and if the at least one prediction indicates that the at least one product design objective has not been achieved and the at least one model design objective has not been achieved, the build, execute, and evolve process is repeated. A system comprising at least one processor, A computer-based system equipped with the following features.

11. The design parameter space and the design evaluation space are associated with the energy storage device and at least one of the constructed models. The aforementioned design parameter space includes multiple variables, The aforementioned plurality of variables include material variables, system variables, or combinations thereof. The aforementioned multiple variables are associated with the chemical space of the energy storage device, the compounding space of the energy storage device, the material space of the energy storage device, the component space of the energy storage device, the process space of the energy storage device, or a combination thereof. The design evaluation space includes at least one test of the energy storage device, and the at least one processor is further configured to simulate the at least one test using the constructed at least one model in order to perform the simulation. The at least one product design objective includes at least one user-specified criterion associated with the energy storage device, at least one machine-generated criterion associated with the energy storage device, or a combination thereof. The aforementioned at least one product design objective includes a target product profile (TPP), and The at least one model design objective includes at least one error threshold associated with the difference between simulated measurements and real-world measurements, the target size relative to the design parameter space, or a combination thereof. The computer-based system according to claim 10.

12. In order to automatically evolve at least one of the design parameter space and the design evaluation space, the at least one processor By employing clustering methods, Pareto methods, or combinations thereof, diversity within the design parameter space is maintained. The computer-based system according to claim 10, further configured as follows.

13. The computer-based system according to claim 10, wherein the at least one processor is further configured to implement the machine learning process and to employ a generative adversarial network (GAN), a deep neural network (DNN), Bayesian optimization (BAO), a genetic function approximation method, or a combination thereof in the machine learning process.

14. The computer-based system according to claim 10, wherein the design parameter space includes semantically structured real-world evidence (RWE) data associated with experiments on the energy storage device.

15. The computer-based system according to claim 10, wherein the at least one processor is further configured to automatically store in the at least one memory the constructed at least one model, associated with the at least one prediction generated, the at least one input to the constructed at least one model, and the at least one output from the constructed at least one model, and the at least one processor is further configured to input the at least one input to the constructed at least one model in order to perform the simulation, and the constructed at least one model is configured to produce the at least one output in response to the at least one input.

16. The computer-based system according to claim 10, wherein the energy storage device is a battery, and the at least one processor is further configured to identify at least one of compounds, components, additives, formulations, recipes, or combinations thereof that enable the energy storage device to achieve the at least one product design objective.

17. A non-temporary computer-readable medium for the generative design of an energy storage device, which, when loaded and executed by at least one processor, the at least one processor, Based on the design parameter space, at least one model of an energy storage device is automatically constructed, and this construction employs a machine learning process. The process includes automatically performing a simulation of the energy storage device, wherein the simulation employs the design parameter space, the design evaluation space, and the constructed at least one model, and the execution generates at least one prediction of the energy storage device that achieves at least one product design objective or at least one model design objective achieved by the constructed at least one model, (i) Automatically evolve the design parameter space and (ii) the design evaluation space, wherein the evolution is based on the generated at least one prediction, employing the machine learning process, and including (a) trimming the design parameter space, (b) trimming the design evaluation space, (c) expanding the design parameter space, (d) expanding the design evaluation space, or (e) a combination of (a) to (d), In constructing, evolving, or a combination thereof, at least one monitored parameter is employed, the at least one monitored parameter represents at least one real-world result generated through at least one real-world experiment employing the energy storage device, and the real-world result is carried out based on employing the evolved design parameter space in the at least one real-world experiment. If the at least one prediction indicates that the at least one product design objective has been achieved or the at least one model design objective has been achieved, the generative design of the energy storage device is completed by automatically converging in the evolved design parameter space; and if the at least one prediction indicates that the at least one product design objective has not been achieved and the at least one model design objective has not been achieved, the build, execute, and evolve process is repeated. A non-temporary, computer-readable medium that encodes a sequence of instructions that cause something to happen.

Images