End-to-end platform for managing a circular economy of waste materials

Abstract

Systems and methods of managing a chemical recycling process include accessing characterization data for a feedstock, the characterization data including one or more spectra collected according to one or more spectroscopic methods. The method includes predicting a set of constituent materials included in the feedstock using the characterization data. The method includes predicting a material composition of the feedstock using the predicted set of constituent materials. The method includes identifying one or more target products using, at least in part, the predicted material composition of the feedstock. The method includes generating a set of chemical reaction schemes capable of converting at least a portion of the feedstock into the one or more target products. The method further includes storing the material composition of the feedstock, the one or more target products, and the identification of the set of chemical reaction schemes in a data store.

Description

[0001] Cross-references to related applications

[0002] This application claims the benefit and priority of U.S. Application No. 17 / 033,512, filed on September 25, 2020, which is incorporated herein by reference in its entirety for all purposes. Technical Field

[0003] This manual relates to methods and systems for the circular economy in managing waste materials. Background Technology

[0004] Plastic products are primarily for single use and are generally not recycled. Global annual plastic production is approximately 350 million tons, of which roughly 10% is eventually recycled, 12% is incinerated, and the remainder (78%) ends up in landfills or the natural environment, where it takes nearly 500 to 1000 years to decompose. Plastic production is projected to double by 2030 and triple by 2050.

[0005] Mechanical recycling is the primary strategy for recycling plastics and involves grinding, melting, and re-extruding plastic waste. High contamination rates and mixed material streams are major causes of low productivity and low value in recycling processes because recycling facilities are often designed to handle streams of sorted material with high purity to maintain high levels of material properties in the recycled products. Feed impurities reduce recycling effectiveness even after only a few mechanical recycling cycles due to complex formulations with additives and physical degradation of materials. For example, polylactic acid (PLA) is a common waste plastic that is often undetectable in polyethylene terephthalate (PET) sorting and mechanical recycling operations. As another example, chlorinated compounds such as polyvinyl chloride (PVC) are unacceptable in both mechanical and chemical recycling operations because corrosive compounds are generated during the recycling process, limiting the value of hydrocarbon outputs. Summary of the Invention

[0006] Provide technologies (e.g., methods, systems, storage of non-transitory computer-readable media containing code or instructions executable by one or more processors) for the reuse of molecular components bound in waste materials.

[0007] Specifically, the technology can target the chemical or material properties of the constituent materials entering the waste stream. Constituent materials can be identified using chemical fingerprints derived from a comprehensive library including spectral information, physical properties, computational algorithms, and machine learning. Material characterization data can be used to develop chemical processes to transform the materials into a target product. The identification of the target product and process can be communicated through logistics information, market data, and real-time transaction data. For example, a target product can be identified as one associated with relatively high market demand and that can be produced relatively efficiently using at least one constituent material.

[0008] In some embodiments, the method may include accessing characterization data of a feed. The characterization data may include one or more spectra collected according to one or more spectroscopic methods. The method may include using the characterization data to predict a set of constituent materials included in the feed. The method may include using the predicted set of constituent materials to predict the material composition of the feed. The method may include using the predicted material composition of the feed to identify one or more target products. The method may include generating a set of chemical reaction schemes capable of converting at least a portion of the feed into one or more target products. The method may also include storing the material composition of the feed, one or more target products, and the identification of the set of chemical reaction schemes in a data storage device.

[0009] In some embodiments, the method may further include identifying one or more inputs to a fitness function, the one or more inputs describing chemical reaction schemes in a set of chemical reaction schemes. The method may further include generating an output of the fitness function using the one or more inputs. The method may further include selecting an implementation scheme from the set of chemical reaction schemes based on the fitness function, one or more inputs, and one or more target products. Identifying one or more target products may include: accessing inventory information describing a set of products; and using the inventory information to identify an incomplete subset of the set of products as one or more target products. The inventory information may include one or more of the quantity of feed available for conversion or the quantity of target products among one or more target products available in a geographic area. The method may further include directing a portion of the feed to a material recycling facility configured to convert that portion of the feed into at least one of one or more target products. Generating a set of chemical reaction schemes may include: accessing a chemical reaction inventory, the chemical reaction inventory including representations of chemical reactions describing the conversion of the feed into a target product among one or more target products; and populating the set of chemical reaction schemes based on the chemical reaction inventory. Generating a set of chemical reaction schemes may include: simulating a first constituent reaction of a chemical reaction scheme in the set of chemical reaction schemes using a machine learning model. Generating a set of chemical reaction schemes may include: estimating the output of a reward function, wherein the output of a machine learning model is used as input to the reward function. Generating a set of chemical reaction schemes may also include: estimating the maximum value of the reward function by modifying the input to the machine learning model, wherein the input is the output of a second constituent reaction preceding the first constituent reaction in the chemical reaction scheme.

[0010] In some embodiments, a computer system includes one or more processors and a memory in communication with the one or more processors, the memory being configured to store computer-executable instructions, wherein executing the computer-executable instructions causes the one or more processors to perform one or more aspects of the methods described above.

[0011] In some embodiments, a computer-readable storage medium stores computer-executable instructions that, when executed, cause one or more processors of a computer system to perform one or more aspects of the methods described above. Attached Figure Description

[0012] Figure 1 Example techniques for managing the reuse of molecular components in a feed, according to some embodiments of the present disclosure, are shown.

[0013] Figure 2 An example workflow for predicting the material composition of a feed is shown according to some embodiments of this disclosure.

[0014] Figure 3An example workflow for generating a set of chemical reaction schemes according to some embodiments of this disclosure is shown.

[0015] Figure 4 An example workflow for tuning a chemical reaction process using chemical and logistics data is shown according to some embodiments of this disclosure.

[0016] Figure 5 An example flow diagram of a method for managing the reuse of molecular components of a feed, according to some embodiments of the present disclosure, is shown. Detailed Implementation

[0017] Mechanical recycling is limited in its applicability to mixed, complex, and contaminated waste streams, partly because it employs mechanical separation and reforming processes that are insensitive to chemical contaminants and may not be able to modify the chemical structure of the waste materials. Chemical recycling overcomes the limitations of mechanical recycling by breaking down the chemical bonds of waste materials into smaller molecules. For example, in the case of polymeric materials, chemical recycling can provide a pathway to recover oligomers, monomers, or even basic molecules from plastic waste feed. In the case of polymers, chemical recycling processes can include operations to depolymerize and dissociate the chemical composition of complex plastic products, allowing their byproducts to be recycled upwards as feed for new materials.

[0018] Chemical recycling elements may allow materials to be repeatedly dissociated into primary feedstocks. In this way, chemical recycling can be integrated into an 'end-to-end' platform to facilitate the reuse of molecular components of recyclable materials, rather than being limited to a finite number of physical processes due to chemical structure and material integrity, as is the case in mechanical recycling. For example, the products of chemical recycling may include basic monomers (ethylene, acrylic acid, lactic acid, polyethylene, etc.), feed gases (carbon monoxide, methane, ethane, etc.), or elemental substances (sulfur, carbon, etc.). Based on the molecular structure of the input waste material, products that can be synthesized from intermediate chemicals can be identified, rather than being limited to a single set of recycled products, where intermediate chemicals can be generated from the waste through chemical reactions. Thus, an end-to-end platform can manage the waste stream by generating chemical reaction schemes to transform waste materials into one or more target products. For example, an end-to-end platform can direct waste feedstock to a chemical recycling facility for the chemical conversion of waste materials into target products. In this way, end-to-end platforms can improve the implementation of reuse and recycling strategies and increase the transfer of waste materials from disposal to recycling systems.

[0019] End-to-end platforms can collect data describing the quantities of waste materials, molecular components, and finished products, and use this information to proactively manage recycling processes to produce target products. Chemical reaction schemes can be modified or updated using data to alter the quantity, endpoint, or chemical structure of the target product. For example, processes for converting waste into feed monomers and then back into feed polymers can be tracked and integrated into local, regional, and / or global waste recycling or uplink systems. Systems may include recycling operators such as chemical processors, material recycling facilities, waste sources, and endpoints for refining polymer feeds. In turn, waste sources can include, but are not limited to, industrial, institutional, or household waste sources. Downstream processors may incorporate chemically recovered pure material products without accepting reused waste materials.

[0020] The potential advantages of the chemical recycling schemes described herein include the generation of byproducts from waste materials that cannot be obtained through mechanical recycling and can increase the proportion of recyclable waste materials. For example, plastic feedstocks can be completely converted into non-plastic materials, such as hydrocarbon gases, which can then be synthesized into new and different polymer materials. Furthermore, the development of assisted chemical processes using machine learning features can potentially reduce process development timelines and improve the efficiency of chemical recycling processes, making them feasible at scale. For example, the embodiments described herein can include accelerating the development timeline of new chemical reaction schemes, from laboratory scale to pilot scale, and finally to industrial scale (which can typically take up to 17 years), and then to real-time processes using multi-scale simulations of chemical recycling processes already active in logistics networks.

[0021] Another advantage is the potential to address the limitations of traditional recycling methods, which are typically designed to process relatively pure streams of waste with minimal contaminants. The techniques described herein can improve the recycling process by characterizing waste materials and managing recycling programs to produce desired products with enhanced efficiency and performance. Advantages may include, but are not limited to, product yield per unit weight of waste material, energy consumption, environmental impact of the recycling process, or the proportion of recyclable waste transferred to landfills or disposed of in water bodies.

[0022] Figure 1 An example workflow 100 for managing the reuse of molecular components of a feedstock according to some embodiments of the present disclosure is shown. Generally, workflow 100 may include one or more systems for characterizing waste materials, predicting the composition of waste materials, and developing chemical recycling protocols for waste materials, wherein the waste materials can be used as feedstock to be converted into one or more target products through one or more chemical recycling processes.

[0023] In some embodiments, workflow 100 may include a material characterization system 110, which may be implemented as a point-of-use device, such as a tablet, smartphone, laptop, or a dedicated sensor device that may include one or more sensor tools to facilitate the spectroscopic, imaging, or chemical characterization of waste material 111. Waste material 111 may be or includes materials that can be used as feed for a recycling process. For example, waste material 111 may be or include materials that are typically recycled, such as polyethylene terephthalate (PET), such that waste material 111 may be tagged prior to being characterized as feed for a recycling process. In some cases, waste material 111 may also include contaminants or additives that can be identified by analysis by characterization system 110 and can inform the use of waste material 111 as feed, as described in more detail below. In some cases, waste material 111 may be tagged with additional metadata to inform subsequent analysis of the material as part of workflow 100. For example, the tag may be or include a CAS number, which may allow standard characterization data to be retrieved or accessed from a database of standard data.

[0024] The material characterization system 110 can provide one or more types of characterization data 113 describing the waste material 111. Characterization data 113 may include spectral data generated by measuring the interaction of light of one or more wavelengths with the waste material 111. For example, characterization data 113 may include, but is not limited to, spectroscopic methods such as surface light reflection / absorption data 115, transmission absorption data 117, or hyperspectral image data measured by irradiating the waste material 111 with light within one or more spectral ranges. In some cases, characterization data 113 may include infrared absorption data, infrared reflection data, visible light absorption or reflection data, near-infrared data, ultraviolet absorption data, or microwave or X-ray interaction data (e.g., X-ray fluorescence). In some embodiments, characterization data 113 may include physical and chemical characterizations, including but not limited to surface resistivity data, physical characterization data such as hardness or tensile properties, or other physical or chemical properties that may include information distinguishing the waste material 111 from other types of waste materials.

[0025] In some embodiments, waste material 111 may include, but is not limited to, polymers, plastics, plastic-containing composites, non-plastics, lignocellulosic materials, metals, glass, and / or rare earth materials. Polymer and plastic materials may include materials formed through one or more polymerization processes and may include highly crosslinked and linear polymers. In some cases, waste material 111 may include additives or contaminants. For example, plastic materials may include plasticizers, flame retardants, impact modifiers, rheology modifiers, or other additives included in waste material 111, for example, to impart desired properties or promote formation properties. In some cases, waste material 111 may contain constituent chemicals or elements that may be incompatible with a wide range of chemical recycling processes, and in such cases, characterization data 113 may include information specific to these chemicals. For example, the decomposition of halogen- or sulfur-containing polymers may produce corrosive byproducts that may inhibit or impair the chemical recycling of waste material 111 containing these elements. An example of waste material 111 containing halogen components is polyvinyl chloride (PVC). For example, the decomposition of PVC generates chlorine-containing compounds that can act as corrosive byproducts.

[0026] Once collected, the characterization data 113 can be accessed by a computer system 120 that implements one or more elements of workflow 100. In some embodiments, the computer system 120 may include servers, one or more servers, virtual machines, or multiple virtual machines that may be implemented in a physical computer system or a distributed computer system (e.g., a cloud computing system). In some cases, the computer system 120 may communicate with one or more external systems, such as the materials characterization system 110, via a network 130. The network may be a public network, such as the Internet, or a private network, such as a client network, a restricted network, or a local area network.

[0027] In some embodiments, computer system 120 may perform a process for predicting the constituent materials included in waste material 111. (See below for reference.) Figure 2 In more detail, computer system 120 can access spectral libraries corresponding to one or more spectroscopic methods. For example, database 131 may contain spectral data of various standard materials, combinations of materials, and empirical characterization data of real-world materials. In some embodiments, database 131 may communicate with computer system 120 via network 130. Furthermore, computer system 120 may store at least a portion of the spectral library in its memory.

[0028] In some embodiments, predicting the set of components to be included in waste material 111 may include performing a material identification application 140. The material identification application 140 may include a spectrometer 143 that receives characterization data 113. In some embodiments, the spectral data 141 may also be used as input to the spectrometer 143, which may be provided by accessing a spectral library in the memory of computer system 120 or from database 131. In some embodiments, the spectral data 141 may be simulated or measured empirically. As described below, the material identification application 140 may identify one or more bands of interest in the characterization data 113, and one or more bands of interest may be used as part of a chemical fingerprint 145 for generating waste material 111. In summary, the chemical fingerprint 145 of waste material 111 describes a set of characteristic information derived from the characterization data 113, which may identify, for example, the material components of waste material 111 that may be introduced into a chemical recycling process as feedstock.

[0029] In the context of material identification application 140, chemical fingerprint 145 can describe a prediction of the constituent materials and material composition of waste material 111 based at least in part on characterization data 113 and spectral data 141. For example, chemical fingerprint 145 can describe the major component compounds and additives or contaminants indicated by characterization data 113. For example, when characterization system 110 implements a calibrated spectral method that facilitates an absolute composition approach, chemical fingerprint 145 can also describe the relative composition of each constituent material of waste material 111. In some embodiments, material composition can be predicted based on standard data or can be predicted as part of a machine learning model trained using a dataset that includes information from mixed materials, as referenced below. Figure 2 To describe in more detail.

[0030] For reference Figure 3 In more detail, chemical fingerprinting 145 can allow the identification of one or more targeted or desired chemically recycled products. For example, in some embodiments, as part of performing material identification application 140, computer system 120 can access inventory information of chemical reactions describing one or more products (such as product sets), which may be generated by a chemical recycling process that uses waste material 111 as a feedstock. For example, computer system 120 can identify bands of interest, and computer system 120 can also provide one or more targeted or desired products that can be generated from waste material 111 from the bands of interest. (See reference...) Figure 3In more detail, the identification of a target product can be facilitated by accessing chemical recycling process data (such as feed-product pairing) in a searchable table (e.g., a lookup table). As an illustrative example, a chemical fingerprint 145 can be used to predict that waste material 111 may be or include PET, having one or more additives or impurities that can eliminate one or more potential chemical recycling processes, or can lead to a balance or adjustment of the feed ratio to allow recycling and / or reduce wear on the recycling process system. Based on this information, a computer system 120 can access chemical reaction data to provide information describing one or more target products. In this example, the computer system 120 can cross-reference chemical reaction data for the feed with chemical reaction data for impurities, additives, and contaminants to reduce the likelihood of the target product being incorrectly identified.

[0031] Additionally or alternatively, computer system 120 may receive one or more desired product identifiers from a user of computer system 120 as manual input based on chemical fingerprint 145. For example, the computer system may include a user interface or console application through which one or more users interact with one or more applications of computer system 120. In some embodiments, the user interface may allow a user to view data constituting chemical fingerprint 145, search for potential chemically recoverable products, and indicate one or more desired products.

[0032] In some embodiments, the computer system 120 may execute a chemical reaction modeling application 150, which allows the computer system 120 to simulate one or more chemical recycling processes for which waste material 111 acts as a feed to generate one or more target or desired products. (See below for reference.) Figure 3 In more detail, the chemical reaction modeling application 150 can access one or more representations of chemical reactions describing the transformation of feed materials into target products, which can be stored in a database 151 of chemical reaction data. The chemical reaction data can be or include a machine-searchable catalog of basic chemical reactions used for depolymerizing polymers, for dissociating covalent bonds in reactants, or for physically or chemically converting waste material 111 into the target product.

[0033] In some embodiments, the database 151 of chemical reaction data may be or include a list of chemical reactions, which can be used as an initial set of chemical reactions input into a chemical process simulation, as referenced. Figure 3A more detailed description follows. Like database 131, database 151 of chemical reaction data can be a network data storage device or a storage device located in the same physical location as computer system 120. In some cases, chemical fingerprint 145 can be used as additional input to chemical reaction modeling application 150. For example, chemical fingerprint 145 may include information describing the phase, structure, and quantity of one or more feeds and products, as previously described. In this way, the input to chemical reaction modeling application 150 can be or include input molecules, output molecules, catalysts, reagents, solvents, and chemical processing parameters, including but not limited to residence time, reaction temperature, reaction pressure, or mixing rate and mode.

[0034] In some embodiments, the chemical reaction modeling application 150 may be or include one or more unit operation models, which can be implemented to simulate the constituent reactions of a chemical reaction scheme 153. The chemical reaction modeling application 150 can generate multiple chemical reaction schemes 153, which may include different constituent reaction processes or describe different reaction products. In some embodiments, the chemical reaction modeling application 150 may use machine learning models to simulate one or more of the unit operation models, such as artificial neural networks implementing deep learning features, "black box" optimization techniques, supervised learning, reinforcement learning, or other standard machine learning methods. In this way, when the chemical reaction scheme 153 includes multiple constituent reactions, as represented in a series of unit operation models, the chemical reaction modeling application 150 may implement one or more machine learning models, wherein the output of a first model serves as the input of a second model for the machine learning model. (See reference...) Figure 3 In more detail, the chemical reaction modeling application 150 can implement a model tuning protocol via a reward function, which allows iterative modification of one or more parameters of the unit operation model to optimize or improve the model. In some embodiments, tuning the model may include estimating the output of a reward function as a function of one or more values ​​computed by the chemical reaction modeling application 150 and modifying one or more model parameters to maximize the output of the reward function. In addition to the reward function, the training of a machine learning model implemented as part of the chemical reaction modeling application 150 will also be described in more detail below.

[0035] In some embodiments, one or more unit operation models may be based on first principles rather than machine learning methods. As an illustrative example, chemical recycling processes, such as polymer catalytic degradation unit operations, can be simulated using chemical rate equations, for example, by providing input variables to the chemical rate equations using the aforementioned unit operation models or by using a heuristic approach based on table lookups. In this way, a series of unit operation models simulated by the chemical reaction modeling application 150 may include both machine learning models and first-principles models. In some embodiments, such as when the material characterization system 110 includes an online sensor system as part of a material sorting process, the chemical reaction modeling application 150 may progressively access or receive chemical fingerprint 145 data and may update the reaction scheme 153 in response to receiving updated information. Real-time updates to the chemical reaction simulation can improve the performance of the chemical recycling process managed by the computer system 120. For example, waste material 111 may be redirected from its initial receiving destination to another receiving destination after the chemical reaction scheme 153 is updated, which can improve one or more performance factors, as described below.

[0036] In some embodiments, the chemical reaction scheme 153 or constituting a chemical unit operation can be filtered by one or more selection operations performed by the computer system 120. For example, a fitness function can be defined, and the implementation scheme can be selected by that fitness function, as referenced below. Figure 3 and Figure 4 A more detailed description follows. The fitness function can be an object model with multiple inputs, which may include, but are not limited to, predicted input quantities, output quantities, energy input values, cooling water requirements, material costs, or fuel consumption of the logistics operations involved in transporting waste material 111. In some embodiments, the fitness function may receive derived values ​​as inputs, including but not limited to reaction yield, conversion efficiency, chemical reaction selectivity, heat balance value, energy consumption, or environmental impact. Environmental impact may describe the generation of regulated byproducts, including but not limited to greenhouse gases, chemical wastewater, or vitrified slag. For example, a “greenness” approach may be used to establish a comprehensive index that provides a comprehensive quantitative measurement of environmental impact and the sustainability of the proposed reaction conditions. Similarly, “green chemistry and life cycle assessment principles” may be used to promote safe processes that minimize the generation of hazardous substances. In some embodiments, each parameter provided to the fitness function may be assigned a weight that may affect the favorability of a given chemical reaction scheme 153 or constituting a chemical unit operation.

[0037] Chemical reaction modeling application 150 can provide optimization engine 160 with an output including chemical reaction scheme 153. Optimization engine 160 may be or include a machine learning model and can facilitate real-time modification or selection of chemical reaction scheme 153 based on inputs, including but not limited to inputs generated by chemical reaction modeling application 150, chemical fingerprint 145, or inventory information 163. In some embodiments, inventory information 163 may be accessible from a networked system of recycling information 161. Recycling information 161 may be stored in a progressively (e.g., real-time) updated database that can detail the material supply chain and track waste feeds by breaking them down and subsequently resynthesizing them into new materials. For example, inventory information 163 may include the quantity or quality of feeds available in a logistics network that may correspond to a geographic region. Similarly, inventory information 163 may include inventory information of target products available in a geographic region.

[0038] In some embodiments, optimization engine 160 may use inventory information 163 to modify the target or desired product used as input to chemical reaction modeling application 150. For example, computer system 120 may access inventory information 163. Using inventory information 163, computer system 120 may identify a subset of a large number of target products to limit the number of chemical reaction schemes 153 generated. As an illustrative example, waste material 111 may be identified as a potential feed for various chemical recycling methods that provide a variety of possible reaction products. By accessing inventory information 163 corresponding to possible reaction products, one or more possible reaction products can be selected as target products with limited supply to avoid an already prevalent oversupply of the product or to generate a product whose consumption reflects the likelihood of high demand. In some embodiments, reference is made as follows. Figure 4 In more detail, real-time exchange can connect recyclers, chemical companies, and other consumers or producers of recyclable materials. Real-time exchange can implement inventory planning, supply and demand management, and market and logistics management of recycled products. For example, computer system 120 can direct waste material 111, or a portion thereof, to a material recycling facility or other processing operation, where waste material 111 can be converted into one or more target products. Examples of directing waste material 111 may include identifying receiving and sending facilities and generating logistics information that can be provided to either the receiving or sending facility.

[0039] As part of the chemical recycling process for waste material 111, computer system 120 may provide external computer system 170 with one or more of the following: chemical reaction scheme 153, model output, chemical fingerprint 145, characterization data 113, or other information generated, processed, or accessed by computer system 120. External computer system 170 may be or include a control server at the material recycling facility. For example, computer system 120 may receive characterization data 113 from a field characterization system 110 including multiple sensors and probes, generate chemical reaction scheme 153 as described above, and provide external computer system 170 with chemical reaction scheme 153 and / or implementation schemes for execution using the chemical processing units of the material recycling facility. In this way, external computer system 170 may receive information from computer system 120 via network 130.

[0040] In some embodiments, computer system 120 may store the same or similar information in a data storage device, such as database 131 or chemical reaction inventory database 151. For example, computer system 120 may maintain a chemical reaction inventory by storing representations of all chemical inputs and outputs of a reaction, as well as the catalysts and reaction conditions involved. Information stored in the chemical reaction inventory can then be used to optimize known and widely used reactions and to help explore and discover novel catalysts and reaction conditions that may be suitable for decomposing plastic waste. Similarly, bands of interest developed by material identification application 140 may be stored for training and improvement of material fingerprinting methods, as referenced below. Figure 2 To describe in more detail.

[0041] Figure 2 An example workflow 200 for predicting the material composition of a feed is shown according to some embodiments of this disclosure. This includes the development and management of waste materials (e.g., waste materials) used as feed. Figure 1 As part of the chemical recycling process of waste material 111), workflow 200 can combine different datasets, data processing technologies, and analytical operations. The system implementing workflow 200 can be or includes references... Figure 1 The computer system described (e.g., Figure 1 The computer system 120 may be a communication system, such as a hosted spectral analysis application (e.g., Figure 1 The material identification application (140) is used on the client device. The output data of workflow 200 can, for example, facilitate the prediction of chemical reaction schemes by generating chemical fingerprint data, as described below, wherein this chemical fingerprint data can be used to identify target products, select candidate chemical reaction unit operations, or collect inventory information from a logistics network, as well as the above references. Figure 1 Other uses described.

[0042] The workflow 200 may include one or more data acquisition and processing operations. In some embodiments, it may be... Figure 1 The spectral database 210 of the database 131, as an example, can store and process spectral data 220 and related data 230 as part of identifying experimental materials and compositions via the spectral analysis system 240. Spectral data 220 may be or include calibrated or uncalibrated spectral data prepared to facilitate the generation of chemical fingerprint data 250 by the spectral analysis system 240. Spectral data 220 may include spectral characterization data 221 of pure controls. Pure controls may include a base polymer membrane without additives or contaminants, also known as a spectral standard, for multiple individual membranes. The base polymer membrane may be or include polypropylene, polyethylene, polystyrene, high-density polyurethane, low-density polyurethane, polyethylene terephthalate, acrylonitrile butadiene styrene, polycarbonate, or polyamide. Furthermore, the spectral characterization data 221 of pure controls may include, but is not limited to, spectral data of control polymer membranes with known amounts of additives or even individual additives. Different spectral datasets may also be generated for control samples using various forms across the electromagnetic spectrum (X-ray fluorescence, radio frequency, near-infrared, short-wave infrared, mid-wave infrared, THz, or mm range), as described above. The molecular and elemental composition of the control material can also be characterized and included in the spectral characterization data 221 of the pure control.

[0043] In some embodiments, spectral data 220 may be or include spectral characterization 223 of a material sample, which may include data generated by one or more spectroscopic techniques applied to a sample of the waste material or its constituent components. For example, deformulation techniques may be applied to further derive ground-based information about the waste material using destructive methods such as gas chromatography-mass spectrometry (GCMS), laser-induced breakdown spectroscopy (LIBS), or non-destructive methods (ATR-FTIR). Spectral data 220 may include spectral characterization 221 of material samples collected from a recycling network and progressively provided to spectral database 210. For example, a network of material recycling facilities may collect material characterization data (e.g., Figure 1 Characterization data 113) can be used as part of the waste material intake or sorting process, and the data can be provided to the spectral database 210 as part of the implementation of cross-network chemical recycling management.

[0044] Spectral data 220 can be coordinated with related data 230, for example, by labeling it with name data 231, formula data 233, or other metadata 235. In some cases, related data 230 may correspond to spectral data 220 to facilitate spectral analysis techniques implemented by the spectral analysis system 240, including but not limited to model training techniques, as described below. Spectral data 220 and related data 230 can be stored in separate data storage devices connected to the spectral database 210 via a network. For example, spectral data 220 can be accessed by the spectral database 210 as part of distributed data system operations, such as extraction, transformation, and loading (ETL) processes. Similarly, related data 230 can be collected in one or more databases located in one or more physical locations and can be accessed or received by the spectral database 210.

[0045] In preparing for subsequent data processing, the data acquisition 211 operation can be applied to the spectral data 220 and the associated data 230. For example, data acquisition 211 may include one or more data transformations, such as ETL processes, that can modify the format or representation of the data. For instance, the spectral file 213 from the spectral data 220 may be combined with associated related data 215 as part of data acquisition 211 to generate spectral data entries in a normalized format 217. This may include converting the spectral file 213 from a standard data format, such as a comma-separated value format, to a key-value pair format. The key may be or include searchable database tags, such as unique identifiers. The normalized format 217 may include fields from the associated data 215, such as tags for the chemical composition of the sample. Examples of tags may include, but are not limited to, name data 231, formula data 233, molecular weight data, and associated metadata 235, such as SMILES string data, MOL file data, CAS numbers, or structural representations.

[0046] like Figure 2As described, the spectral analysis system 240 can access data in a standardized format 217 as part of generating chemical fingerprint data 250. The spectral analysis system 240 can generate chemical fingerprint data 250 for standard control materials and characteristic waste materials, and can combine both automated and manual analysis techniques. For example, the spectral analysis system 240 can implement a set of visualization tools 241 and can implement machine learning methods or other computational spectral analysis 245 techniques as part of developing the chemical fingerprint data 250. The visualization tools 241 can be used to query the relevant spectra of specific materials, additives, or contaminating chemicals. In some embodiments, the visualization tools 241 can display sample spectra and align sample spectra with one or more control spectra for comparison by human eye or by machine image analysis (e.g., by a convolutional neural network trained to classify spectra). The visualization tools 241 can allow for rapid analysis of anomalous spectra and manual synthesis of spectral datasets used as training sets for machine learning. The visualization tools 241 can allow for the labeling of synthesized sample spectra, as used in reinforcement learning to refine training sets.

[0047] In some embodiments, the spectral dataset can be normalized by data preprocessing 243 including a modular normalization method. For example, intensity normalization can be applied to the raw spectral data based on, for example, the identification of key features such as peaks or bands, as part of identifying the band of interest 251. As described below, the band of interest 251 can be used to guide the material characterization system by configuring spectral probes and can be additionally or alternatively used to identify unlabeled spectra detected by broadband characterization techniques. For example, the band of interest 251 of a transmission scanner may be or include 1620 nm to 1787 nm from 1350 nm to 2450 nm, such that normalization or other data processing can be preferentially applied thereto. Similarly, the band of interest 251 of a reflectance scanner may be or include 1117 nm to 1261 nm from 900 nm to 1700 nm. Normalization can refer to intensity normalization and can be applied, including but not limited to, situations where the spectral sensor device is not intensity calibrated.

[0048] Data processing 243 may include baseline and other compensation techniques. For example, a baseline in the spectral file may be detected, which may correspond to a background signal or general trend in the original spectrum not attributable to the measured sample. In some cases, normalization may include multiple operations, including but not limited to baseline subtraction and intensity normalization performed by dividing the resulting intensity data for each band by the sum of all differences. In this way, the processed spectral data can be normalized across different material thicknesses and transparency. The normalized spectrum may be smoothed to improve subsequent computational spectral analysis 245.

[0049] In some embodiments, the processed spectral data can be used as training data 247 for a machine learning model implemented as part of computational spectral analysis 245. For example, the machine learning model can be or includes a support vector machine (SVM) classifier. The machine learning model can be trained using at least a portion of normalized spectral data, which may be labeled or unlabeled, via model training 249, which includes, but is not limited to, supervised learning or reinforcement learning. In some embodiments, model training 249 can be implemented using subband data, which can provide improved classification accuracy compared to training using the full spectrum. Model training 249 can implement adversarial learning methods, such as discriminators, which can be used to train the machine learning model implemented as part of computational spectral analysis 245.

[0050] Normalized spectral data can be processed by a trained machine learning model or other computational methods (such as process- or rule-based models) to look for patterns in signals associated with material labels 253, additive or contaminant labels 255, or other information indicating chemical type, composition, form, structure, or purity. In materials that combine a variety of different additives, contaminants, or impurities with the main material, such as units comprising different forms of recycled PET objects including different plasticizers (e.g., materials received by a material recycling facility), multiple regions covering the peak signal of the material can be identified as bands of interest 251. In some embodiments, up to 30 to 40 bands of interest 251 can be selected, excluding less informative bands that may be common in all forms of recycled feed materials. In an illustrative example, a classifier implementing an SVM trained to classify materials can be provided with the bands of interest 251 of the waste material sample based on the labels of the spectra included during data acquisition 211.

[0051] In some embodiments, chemical fingerprint data 250 may be stored in fingerprint database 260. The fingerprint database may communicate with the spectral analysis system 240, for example, via a network or at the same physical location as the spectral analysis system 240. As part of implementing computational spectral analysis 245, the spectral analysis system 240 may access the chemical fingerprint data 250 stored in fingerprint database 260. For example, by accessing the bands of interest 251 and material labels 253 for broad classes of materials (such as polymers), an SVM trained by model training 249 can provide classifications with sufficient accuracy to distinguish different polymer structures, side chains, main chains, or other information that may influence the identification of potential target products, as well as formulations for chemical reactions that convert the classified spectral data into potential target products, as referenced below. Figure 3 As stated above.

[0052] Figure 3An example workflow 300 for generating a set of chemical reaction schemes according to some embodiments of this disclosure is shown. As part of managing the chemical recycling process, a computer system (e.g., Figure 1 The computer system 120 can simulate one or more chemical recycling unit operations as part of the unit operation simulation 310. The unit operation simulation 310 can receive data generated by material identification and characterization applications, as described above, as part of a platform used to guide the chemical recycling process. The workflow 300 can include implementations of machine learning and rule-based models as part of generating a series of reaction conditions describing the chemical process that transforms the feed into the target product. The feed can be waste material received by the material recycling facility. The target product can be identified by the computer system as part of the workflow 300 and can be additionally or alternatively specified by external input.

[0053] Unit operation simulation 310 can receive or access material identification data 320 of waste materials (e.g., Figure 2 The chemical fingerprint data 250), and the material identification data 320, can be used as input 321 for the identity and composition of the waste material. The material identification data 320 may include desired input 323, which may be provided by an external system such as via an exchange system, as described below. Figure 4 The unit operation simulation 310 can also receive input from a chemical reaction list 330. The chemical reaction list 330 can store representations of all chemical inputs 333 and outputs 335 of the reaction, the catalysts involved 337, and the reaction conditions 331, such as embeddings. A reaction model 339 can also be stored as part of the chemical reaction list 330, which allows the unit operation simulation 310 to include rule-based reaction models in addition to machine learning methods, as part of the guided chemical recovery 350 formulation. Inputs 333 and outputs 335 can be cross-referenced in the chemical reaction list 330, allowing material identification data 320 to be paired with potential target products, which can be used to define a set of initial chemical reaction schemes that can be optimized, as described below.

[0054] As part of generating optimized reaction conditions 340, unit operation simulation 310 can modify known and widely used reactions and facilitate the exploration and discovery of novel catalysts, reagents, or solvents 343 and reaction conditions 341 that may be suitable for decomposing waste materials. In some embodiments, unit operation simulation 310 can incorporate molecular modeling techniques (such as density function theory and molecular dynamics) into known sets of catalysts or reagents to generate new catalyst data 337 that is not previously available in the chemical reaction inventory 330. Various clustering methods, Gaussian mixture models, factor analysis, and unsupervised ML algorithms that learn reaction embeddings through deep neural networks (DNNs) can be applied to the data from the chemical reaction inventory 330. In some embodiments, supervised ML algorithms, such as regression models or DNNs, can be used to improve the chemical reaction model 339. For example, in spectral analysis (e.g., Figure 2 As described in the context of the spectral analysis system 240), the machine learning method implemented as part of the unit operation simulation 310 may be trained using a dataset from the chemical reaction list 330. The training may include one or more preprocessing steps, such as labeling, synthesis processing, or other methods for selecting training data and guiding the development of the ML model.

[0055] When retrieving recommended catalysts and chemical reactions, multiple methods can be combined into a guided chemical recovery simulation 350 to generate optimized reaction conditions 340. In some embodiments, one or more chemical processes can be simulated as a series of reaction models 353a-n, each of which receives inputs 351a-n and generates outputs 355a-n. Each reaction model 353 can represent a chemical unit operation that forms a stage in the chemical recovery process. In some cases, a terminal reaction model 335n can output a final output, which can represent the target product identified from the chemical reaction list 330.

[0056] In some embodiments, simulation results representing intermediate reaction conditions can be provided to an online learning algorithm to fine-tune the model and simulation techniques. The online learning algorithm can incorporate a reward function 360, which can indicate the success of a reaction or reaction scheme. In some embodiments, the reward function 360 can generate threshold criteria representing one or more chemical process parameters (such as input 351 or output 355), through which the optimization of the entire guided chemical recovery simulation 350 can be judged. For example, one or more inputs 351 or outputs 355 can be provided to the reward function 360 at each iteration, and the unit operation simulation 310 can increment those inputs 351 or outputs 355 until a desired result is reached, such as the output of the reward function 360 crossing a threshold, which can indicate that the reaction conditions from the guided chemical recovery 350 have been optimized. In some embodiments, the output of the reward function 360 crossing a threshold can indicate that the reaction conditions from the guided chemical recovery 350 are above a minimum acceptable level, rather than at a maximum level.

[0057] In some cases, the outcome may be or includes an optimized chemical reaction scheme, such as a pyrolysis process, to efficiently break down a plastic into a desired set of molecules. In another embodiment, given inputs of carbon monoxide and hydrogen in a Fischer-Tropsch reaction, reward function 360 may receive pressure, temperature, and catalyst levels to produce liquid hydrocarbons that may be the feedstock for the plastic. Reward function 360 may be or includes a general optimization algorithm, such as steepest descent, to guide the increment of input 351 and output 355. Algorithms specific to each chemical reaction problem may also be included, depending on the desired input / output data or the conditions requiring optimization. For example, simulations of thermochemical processes (such as pyrolysis) with sufficient process data available may use different optimization techniques, such as reinforcement learning. Reward function 360 may form part of the ML framework of workflow 300, such as reinforcement learning or black-box / grey-box optimization techniques, and may be used to guide the learning process and evaluate the learning outcomes.

[0058] In addition to those inputs included as inputs 351 to reaction model 353, reward function 360 can receive many other inputs. For example, derived values ​​such as yield, selectivity, feasibility, energy use, or environmental impact can be used as inputs to reward function 360. As an example, yield can be used to determine how much plastic polymer can be successfully converted into its constituent monomers and how much plastic polymer can be converted into unusable byproducts. Similarly, selectivity can describe the ratio of desired monomer output to undesirable reaction output. Feasibility can be used to capture the concept of whether the proposed reaction conditions are feasible / practical for establishment or implementation, suggesting that some inputs to reward function 360 can be qualitative assessments based on non-physical criteria. Weights can be assigned to inputs to the parameters of reward function 360 to bias the reward function (and the learning process) toward specific goals or objectives. For example, if there are limitations on selectivity for certain reaction types, selectivity can be assigned a higher weight in the calculation of threshold criteria. In other cases, yield may be more important and can be assigned a higher weight.

[0059] In addition to optimized reaction conditions, workflow 300 may also include multiple outputs that can improve the implementation, adoption, and performance of the chemical recovery process. For example, optimized reaction conditions 340 can be visualized as a Markov process simulation 370, through which the various stages of the chemical reaction scheme 345 are treated as steps in a Markov process.

[0060] Generally, a Markov process represents different stages in a material flow or process chain as nodes connected by directional arrows, with visual or quantitative indications of the weights of the connections between the corresponding nodes. In this way, a Markov process simulation 370 can generate and / or present a dynamic visualization of reaction scheme 345 to demonstrate the overall effect of fine-tuning the constituent reactions of reaction scheme 345 on the entire recycling pipeline. Thus, the Markov process simulation 370 can receive material flow data describing the material recycling supply chain as input, as referenced below. Figure 4 To describe in more detail.

[0061] Figure 4 An example workflow 400 for tuning a chemical reaction process using chemical and logistics data according to some embodiments of this disclosure is illustrated. Elements of workflow 400 may include data provided by the logistics network and database described in the foregoing figures, which together can be used as input to an optimization engine 410. The optimization engine 410 may implement one or more methods as described below to modify or tune the chemical recycling process simulated by the system in the foregoing figures to better align with network factors such as material inventory levels, logistics constraints, or consumption trends. For example, in Figure 3During the iteration of the unit operation simulation 310, the output of the workflow 400 can be returned as input to one or more of the aforementioned systems.

[0062] In some embodiments, the optimization engine 410 may receive data that can be broadly grouped into: chemical recycling process data, such as material identification data 420 (e.g., Figure 2 Chemical fingerprint data 250) or optimized reaction conditions 430 (e.g., Figure 3 The optimization engine 410 can apply one or more computational methods to modify aspects of the optimized reaction conditions 430 and output optimized data 470. In some embodiments, the optimization engine can receive additional input provided by exchange software 480, which can provide a platform for network interaction with entity 490, which generates feed materials, consumed products, or chemically recovered intermediates, including but not limited to catalysts, solvents, or other consumables.

[0063] Material inventory data 440 may include, but is not limited to, inventory data describing molecule 441, which may describe products or byproducts, feed materials 443, chemical substances 445 (such as consumables, catalysts, or other reactants), or general materials 447. General materials may include, but are not limited to, other materials that can be used to operate the chemical recovery process, such as electricity, cooling water, heating fuel, or compressed gas. In this way, material inventory 440 may represent one or more constraints on the operation of the chemical recovery process identified in optimized reaction conditions 430. Thus, the information reflected in material inventory 440 may potentially indicate the optimized reaction conditions, for example, when a rate-limiting catalyst cannot be provided.

[0064] Similarly, utilization data of 450 can reflect the effectiveness of the reaction scheme (e.g., Figure 3 The reaction scheme 345) is used to promote or degrade local or regional trends in the chemical recycling infrastructure. For example, utilization data 450 may include, but is not limited to, data reflecting downstream demand 451, upstream supply 453, market data 453, or logistics data 455. In addition to the physical and chemical factors reflected by the reaction model and thermochemical optimization, these supply, demand, and market factors can allow the optimization engine to tune one or more parameters of the chemical recycling process simulation to reflect economic factors. For example, the reaction scheme may produce products for which supply exceeds demand and there is a lack of storage capacity in the logistics network. In this case, the optimization engine 410 may degrade the reaction scheme or may identify a target product subset 471. In turn, the target product subset 471 may be returned to the chemical process simulation (e.g., Figure 3 The workflow 300) is tuned to optimize the reaction conditions 430.

[0065] For reference Figure 3 The optimized reaction conditions 430 described above can describe the specific conditions of a single chemical reaction scheme. That is, the optimization engine 410 can simultaneously receive or access multiple chemical reaction schemes, as mentioned above. Figure 1 The selection is a subset of one or more implementations. For example, identifying a subset of target products 471 may allow selection of implementations that generate the subset of target products 471.

[0066] Similarly, optimization engine 410 can output optimized logistics data that describes the source of feed 443 and the recipient of the products generated by optimized reaction conditions 430. For example, logistics data may include real-time data 460, which may include, but is not limited to, data describing the operation of the Material Recycling Facility (MRF) 461, the operation of distributed collection 463, supply chain conditions 465, or material characterization sensor data 467, which describes the materials arriving at the MRF for processing in real time. Conversely, distributed collection 463 data may describe different sources of waste material feed, such as industrial, commercial, institutional, and residential sources. Real-time data 460 coupled with input from exchange software 480 allows optimization engine 410 to specify the recipient of the products and the feed source for entities participating in the chemical recycling network (e.g., through the network of entities 490).

[0067] The optimization engine 410 can implement a fitness function, which includes one or more computational techniques, such as rule-based models or machine learning models, to take into account various types of available chemical recovery process data and stream data, and generate fitness values ​​for optimized reaction conditions 430. (Similar to the reference...) Figure 3 The described reward function ( Figure 3The optimization engine 410 can receive weighted inputs (reward function 360), the weights of which can be externally specified by the operator or autonomous system, or developed by training the optimization engine 410 if a machine learning approach is employed. For example, the optimization engine can include an artificial neural network trained on a training dataset, which can be developed from historical operational data collected for a given chemical recovery process. Training can allow the optimization engine to develop weights corresponding to, for example, the process sensitivity to various logistics data (such as material inventory 440 or market data 455). For example, inventory data for rate-limiting catalysts can have a significant impact on the feasibility of a chemical reaction scheme. In this case, the weights for inputs describing catalyst supply can be higher than the weights for inputs with less influence. In the context of a loss function, the optimization engine can operate by minimizing the value of a loss function defined as the output of a machine learning model that receives chemical recovery process data and logistics data.

[0068] Figure 5 An example flow diagram of a method 500 for managing the reuse of molecular components in a feed, according to some embodiments of the present disclosure, is shown. (See reference...) Figures 1 to 4 The aforementioned one or more operations of the constituting method 500 can be performed by a computer system communicating with an additional system (e.g., Figure 1 The method is executed by a computer system 120, and additional systems include, but are not limited to, characterization systems, network infrastructure, databases, and user interface devices. In some embodiments, method 500 includes operation 510, in which the computer system accesses characterization data of the feed. Characterization data (e.g., Figure 1 The characterization data (113) can be generated in one or more wavelengths using in-situ spectroscopic techniques (such as reflectance spectroscopy, transmission spectroscopy, or fluorescence spectroscopy), as described above. Furthermore, the characterization data can include physical or chemical information based on one or more different techniques; examples of physical or chemical information include hardness, tensile properties, or thermal phase properties. The characterization data can be transmitted via a network (e.g., Figure 1 The network 130) is provided to the computer system.

[0069] In some embodiments, method 500 includes operation 520, wherein the computer system predicts the constituent material set included in the feed. The computer system may implement spectroscopic analysis methods, such as reference... Figure 2 A more detailed description is needed to identify chemical fingerprint data (e.g., Figure 2 Chemical fingerprint data 250). This can include data from spectral databases (e.g., Figure 1 Database 131) receives standard and reference data. Furthermore, operation 520 may include data transformation operations (e.g., Figure 2Data acquisition 211), one or more implementations of a machine learning model, wherein the machine learning model may be implemented using preprocessed training data (e.g., Figure 2 The training data (247) was used for training, and the preprocessed training data was prepared using spectral analysis techniques including but not limited to normalization, baseline subtraction, or smoothing.

[0070] In some embodiments, method 500 includes operation 530, wherein the computer system predicts the material composition of the feed. The material composition of the feed may include information about the relative prevalence of the feed, compared to the constitutive materials. For example, refer to... Figure 2 The described spectral analysis can be based on spectral features (such as material markers, e.g., Figure 2 Material marking 253) or additive marking (e.g., Figure 2 The additive label 255) is used to identify multiple constituent materials. However, for example, when the sensor is not intensity calibrated, such spectral analysis may fail to distinguish between major components and impurities. Therefore, in some cases, cross-referencing the spectral fingerprint with control data, or training a machine learning model with the composition data, can provide a predicted composition of the material, such as the weight-based composition.

[0071] In some embodiments, method 500 includes operation 540, wherein a computer system identifies one or more target products. Identification of target products may be achieved through a chemical reaction list (e.g., Figure 3 The chemical reaction list (330) facilitates this process, allowing the computer system to identify a set of candidate products for the feed. For example, the material composition may include information about a major component, which could be a polymer material, for which the chemical reaction list can describe a number of outputs that can be generated through the chemical recovery of the feed (e.g., ...). Figure 3 The output is 335). Similarly, the catalyst (e.g., Figure 3 Information about catalyst 337 can describe contaminants that may poison the catalyst, and thus, the corresponding chemical reaction and its products can be eliminated from the candidate product set. As mentioned above, halogenated plastics (e.g., chlorinated and fluorinated plastics) may produce corrosive byproducts, which may exclude halogenated plastics from some types of chemical recycling. As mentioned above, the candidate product set can be obtained using logistics data (e.g., Figure 4 The material inventory (440) is refined in one or more ways, which may allow for the identification of a limited quantity of target product or an incomplete subset of target product.

[0072] In some embodiments, method 500 includes operation 550, wherein the computer system generates a set of chemical reaction schemes. Based on the material composition and target product, the computer system can use the above references. Figure 3The described technology (e.g.) Figure 3 The workflow 300) generates a tuned reaction scheme. For example, a chemical recycling process can be simulated as a series of reactions represented by reaction models (e.g., Figure 3 The reaction model (353a-n) is a unit operation. In some cases, the reaction model can receive the output of the previous reaction model in the series as input, such as when the unit operation forms a stage in the process flow. See reference... Figure 3 To describe in more detail, unit operation simulation (e.g., Figure 3 The unit operation simulation 310) can be achieved through a reward function (e.g., Figure 3 The reward function (360°) is used for tuning, allowing multiple factors to influence the operation of a given reaction scheme during tuning. For example, the reward function can receive chemical and physical information (such as cooling water capacity, fuel consumption information, and environmental impact parameters) as input, or other inputs that can directly affect the operation constituting the reaction model of the process unit. Furthermore, the reward function can allow unit operation simulations to be optimized for derived values, including but not limited to yield, selectivity, or efficiency.

[0073] In some embodiments, method 500 includes operation 560, wherein the computer system stores the material composition of the feed, one or more target products, and identifiers of a set of chemical reaction schemes. The output generated by the computer system may include, but is not limited to, reaction schemes, visualization information (e.g., Figure 3 The Markov process simulation 370) and the predicted and generated data on material composition, constituent materials, and other data. In some embodiments, the data thus generated may be stored in a data memory by a computer system and sent to an external computer system (e.g., Figure 1 The external computer system 170 can be accessed, or the data can be returned as feedback during simulation iterations. Furthermore, material identification data, reaction scheme data, target product data, or other generated information can be stored for subsequent use in model training at one or more stages of method 500.

[0074] Various embodiments have been described in the foregoing description. Specific configurations and details have been set forth for illustrative purposes in order to provide a full understanding of the embodiments. However, it will be apparent to those skilled in the art that the embodiments can be practiced without specific details. Furthermore, well-known features may be omitted or simplified so as not to obscure the described embodiments. While the exemplary embodiments described herein focus on polymer materials, these embodiments are non-limiting illustrative examples. The embodiments of this disclosure are not limited to such materials but are intended to address material handling operations for which a broad array of materials can be used as potential feedstocks for material recycling and / or upcycle processes. Such materials may include, but are not limited to, metals, biopolymers (such as lignocellulose materials), viscoelastic materials, minerals (such as rare earth-containing materials), and complex composite materials or devices.

[0075] Some embodiments of this disclosure include a system comprising one or more data processors. In some embodiments, the system includes a non-transitory computer-readable storage medium containing instructions that, when executed on the one or more data processors, cause the one or more data processors to perform some or all of the one or more methods disclosed herein and / or some or all of one or more processes and workflows. Some embodiments of this disclosure include a computer program product tangibly embodied in a non-transitory machine-readable storage medium, the computer program product including instructions configured to cause the one or more data processors to perform some or all of the one or more methods disclosed herein and / or some or all of one or more processes.

[0076] The terms and expressions used are used as descriptive rather than limiting terms, and their use is not intended to exclude any equivalents of the features shown and described or portions thereof. However, it should be recognized that various modifications are possible within the scope of the claimed invention. Therefore, it should be understood that although the claimed invention has been specifically disclosed through embodiments and optional features, modifications and variations of the concepts disclosed herein can be adopted by those skilled in the art, and such modifications and variations are considered to be within the scope of the invention as defined by the appended claims.

[0077] This specification provides only preferred exemplary embodiments and is not intended to limit the scope, applicability, or configuration of this disclosure. Rather, the following description of preferred exemplary embodiments will provide those skilled in the art with enabling descriptions for implementing various embodiments. It should be understood that various changes may be made to the function and arrangement of the elements without departing from the spirit and scope set forth in the appended claims.

[0078] Specific details are given in the description to provide a thorough understanding of the embodiments. However, it should be understood that the embodiments can be practiced without these specific details. For example, specific computational models, systems, networks, processes, and other components may be shown as components in the form of block diagrams to avoid obscuring the embodiments with unnecessary details. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary details to avoid obscuring the configuration.

Claims

1. A data processing method, comprising: Access the characterization data of the feed, the characterization data including one or more spectra collected according to one or more spectroscopic methods; The characterization data is used to predict the set of constituent materials included in the feed; The material composition of the feed is predicted using the predicted set of constituent materials; The predicted material composition of the feed is used at least in part to identify one or more target products; Generate a set of chemical reaction schemes capable of converting at least a portion of the feed into one or more target products; The material composition of the feed, the identification of the one or more target products, and the set of chemical reaction schemes are stored in a data storage device; as well as A portion of the feed is directed to a material recycling facility configured to convert the portion of the feed into at least one of the one or more target products.

2. The method according to claim 1, further comprising: Identify one or more inputs to the fitness function, the one or more inputs describing chemical reaction schemes in the set of chemical reaction schemes, wherein the fitness function is configured to generate fitness values ​​of optimized reaction conditions as the output of the fitness function based on the one or more inputs; The fitness function is generated using the one or more inputs; and Based on the fitness function, the one or more inputs, and the one or more target products, an implementation scheme is selected from the set of chemical reaction schemes.

3. The method according to claim 1 or 2, wherein, Identifying one or more target products includes: Access inventory information describing the product set; and The inventory information is used to identify incomplete subsets of the product set as one or more target products.

4. The method according to claim 1 or 2, wherein, The predicted set of constituent materials included in the feed includes: Access the spectral library corresponding to one or more of the spectral methods and the metadata associated with the spectral library; Identifying the spectral bands in the one or more spectra of the characterization data; and The spectral bands are matched with spectra in the spectral library to predict the constituent materials in the constituent material set.

5. The method according to claim 1 or 2, wherein, Generating the set of chemical reaction schemes includes: Access a chemical reaction list, which includes representations of chemical reactions describing the conversion of the feed into a target product among the one or more target products; and Fill the set of chemical reaction schemes according to the chemical reaction list.

6. The method according to claim 1 or 2, wherein, Generating the set of chemical reaction schemes includes: The first constituent reaction of the chemical reaction schemes in the set of chemical reaction schemes is simulated using a machine learning model; Estimate the output of the reward function, wherein the output of the machine learning model is used as the input to the reward function; and The maximum value of the reward function is estimated by modifying the input to the machine learning model, wherein the input is the output of the second constituent reaction prior to the first constituent reaction in the chemical reaction scheme.

7. A data processing system, comprising: Memory, configured to store computer-executable instructions; and One or more processors, communicating with the memory and configured to execute the computer-executable instructions to: Access the characterization data of the feed, the characterization data including one or more spectra collected according to one or more spectroscopic methods; The characterization data is used to predict the set of constituent materials included in the feed; The material composition of the feed is predicted using the predicted set of constituent materials; The predicted material composition of the feed is used at least in part to identify one or more target products; Generate a set of chemical reaction schemes capable of converting at least a portion of the feed into one or more target products; The material composition of the feed, the identification of the one or more target products, and the set of chemical reaction schemes are stored in a data storage device; as well as A portion of the feed is directed to a material recycling facility configured to convert the portion of the feed into at least one of the one or more target products.

8. The system according to claim 7, wherein, Executing the computer-executable instructions also causes the one or more processors to: Identify one or more inputs to the fitness function, the one or more inputs describing chemical reaction schemes in the set of chemical reaction schemes, wherein the fitness function is configured to generate fitness values ​​of optimized reaction conditions as the output of the fitness function based on the one or more inputs; The fitness function is generated using the one or more inputs; and Based on the fitness function, the one or more inputs, and the one or more target products, an implementation scheme is selected from the set of chemical reaction schemes.

9. The system according to claim 7 or 8, wherein, Identifying one or more target products includes: Access inventory information describing the product set; and The inventory information is used to identify incomplete subsets of the product set as one or more target products.

10. The system according to claim 7 or 8, wherein, The predicted set of constituent materials included in the feed includes: Access the spectral library corresponding to one or more of the spectral methods and the metadata associated with the spectral library; Identifying the spectral bands in the one or more spectra of the characterization data; and The spectral bands are matched with spectra in the spectral library to predict the constituent materials in the constituent material set.

11. The system according to claim 7 or 8, wherein, Generating the set of chemical reaction schemes includes: Access a chemical reaction list, which includes representations of chemical reactions describing the conversion of the feed into a target product among the one or more target products; and Fill the set of chemical reaction schemes according to the chemical reaction list.

12. The system according to claim 7 or 8, wherein, Generating the set of chemical reaction schemes includes: The first constituent reaction of the chemical reaction schemes in the set of chemical reaction schemes is simulated using a machine learning model; Estimate the output of the reward function, wherein the output of the machine learning model is used as the input to the reward function; and The maximum value of the reward function is estimated by modifying the input to the machine learning model, wherein the input is the output of the second constituent reaction prior to the first constituent reaction in the chemical reaction scheme.

13. A computer-readable medium storing computer-executable instructions that, when executed by one or more processors of a computer system, cause the computer system to perform operations, the operations including: Access the characterization data of the feed, the characterization data including one or more spectra collected according to one or more spectroscopic methods; The characterization data is used to predict the set of constituent materials included in the feed; The material composition of the feed is predicted using the predicted set of constituent materials; The predicted material composition of the feed is used at least in part to identify one or more target products; Generate a set of chemical reaction schemes capable of converting at least a portion of the feed into one or more target products; The material composition of the feed, the identification of the one or more target products, and the set of chemical reaction schemes are stored in a data storage device; as well as A portion of the feed is directed to a material recycling facility configured to convert the portion of the feed into at least one of the one or more target products.

14. The computer-readable medium according to claim 13, wherein, When executed by one or more processors of a computer system, the computer-executable instructions further cause the system to perform operations, the operations including: Identify one or more inputs to the fitness function, the one or more inputs describing chemical reaction schemes in the set of chemical reaction schemes, wherein the fitness function is configured to generate fitness values ​​of optimized reaction conditions as the output of the fitness function based on the one or more inputs; The fitness function is generated using the one or more inputs; and Based on the fitness function, the one or more inputs, and the one or more target products, an implementation scheme is selected from the set of chemical reaction schemes.

15. The computer-readable medium according to claim 13 or 14, wherein, Identifying one or more target products includes: Access inventory information describing the product set; and The inventory information is used to identify incomplete subsets of the product set as one or more target products.

16. The computer-readable medium of claim 15, wherein, The inventory information includes one or more of the following: The quantity of the feed available for conversion; The mass of the feed that can be used for conversion; Market data on the feed that can be used for conversion; The quantity of the target product among the one or more target products available in the geographical region; The quality of the target product among the one or more target products available in the geographical area; or Market data for the target product among the one or more target products available in the geographic region.

17. The computer-readable medium according to claim 13 or 14, wherein, The predicted set of constituent materials included in the feed includes: Access the spectral library corresponding to one or more of the spectral methods and the metadata associated with the spectral library; Identifying the spectral bands in the one or more spectra of the characterization data; and The spectral bands are matched with spectra in the spectral library to predict the constituent materials in the constituent material set.

18. The computer-readable medium according to claim 13 or 14, wherein, Generating the set of chemical reaction schemes includes: The first constituent reaction of the chemical reaction schemes in the set of chemical reaction schemes is simulated using a machine learning model; Estimate the output of the reward function, wherein the output of the machine learning model is used as the input to the reward function; and The maximum value of the reward function is estimated by modifying the input to the machine learning model, wherein the input is the output of the second constituent reaction prior to the first constituent reaction in the chemical reaction scheme.

Images