WO2024160331A1

WO2024160331A1 - System and methods for using machine learning to recommend commodities to recover from recyclable materials

Info

Publication number: WO2024160331A1
Application number: PCT/DK2024/050022
Authority: WO
Inventors: Jesper Rømer HANSEN
Original assignee: Vestas Wind Systems A/S
Priority date: 2023-01-31
Filing date: 2024-01-31
Publication date: 2024-08-08

Abstract

A method for identifying and recovering target commodities from a composite wind turbine blade, in which a computing device receives quality grade data indicating quality grades of respective materials included within a turbine blade, and commodity data for a plurality of candidate commodities that are capable of being recovered from the one or more materials of the turbine blade. The computing device determines expected profitability for each of said plurality of respective candidate commodities, wherein the quality grade data for the respective materials and the commodity data for the plurality of candidate commodities are provided as inputs to a data model to cause the data model to determine the expected profitability for each of said plurality of respective candidate commodities. The computing device determines, based on the expected profitability, a recommendation indicating, for respective materials of the turbine blade, one or more target commodities to recover using the respective materials. The device delivers the recommendation to another device or recipient, so that recovery of the target commodities can occur.

Description

SYSTEM AND METHODS FOR USING MACHINE LEARNING TO RECOMMEND COMMODITIES TO RECOVER FROM RECYCLABLE MATERIALS

Technical Field

The present invention relates generally to methods of using machine learning to recommend commodities to recover from materials of a composite member, such as materials of a rotor blade of a wind turbine.

Background

Recycling is the act or process of converting a recyclable material into a reusable material. Typically, a recycling process involves manipulating or destroying the recyclable material (e.g., through a melting process, etc.) such that a resulting reusable material may be used to create new products.

A wind turbine is a device that converts kinetic energy of the wind into electrical energy. For example, a wind turbine may convert kinetic energy of the wind into electricity by using the aerodynamic force created by the wind interacting with the rotor blades to turn an electrical generator.

Modern wind turbine blades are complex composite structures. These structures are made using a range of materials with varying compositions, physical properties, and chemical properties. For example, a wind turbine blade may include various sensors, lightning conductors and receivers for lightning protection, a heating system (e.g., electrical- or fluid-based) for anti- icing or de-icing, glass fiber reinforced composites and carbon fiber reinforced composites for load carrying parts, polymer foams or balsa wood for stiffness of sandwich structures, coatings for protection and aerodynamical performance, protective shells for erosion protection of areas such as the leading edge, roots for connecting the blade root to the hub, adhesive or fixation means for bonding various parts together, and/or the like.

Due to the complexity of the wind turbine blade, each of the constituents of a wind turbine blade may be recovered and recycled in isolation or in combinations as a variety of commodities. However, efforts to recycle components or materials of a wind turbine have been largely absent or ineffective. One reason is that many wind turbine components are made using composite materials, where multiple materials are bound together using an epoxy resin. A large part of the value in recycling composite materials is the actual removal or disposal of "waste" material that would otherwise be put in a landfill or would need to be recycled in another (even more expensive) way (e.g., in situations where the landfill was not allowed). Without an efficient and/or effective way to separate these materials, many components and/or materials of the wind turbine have been placed or buried in landfills without being recycled.

However, new methods of recycling components and/or materials of a wind turbine have been discovered that involve separating each material of a composite material by exposing the composite material to a swelling agent (e.g., formic acid, etc.) for a certain time period. This causes the epoxy resin to disintegrate, allowing each material that had been bound together using the epoxy resin to separate.

In wind turbine blades and in other structures, disassembling allows for recovery of a variety of commodities with a wide range of applications in diverse markets. However, the process of recovering a commodity from a material is different for respective commodities and may involve materials of a composite member that have been separated using the new recycling process. For example, different types of fiber commodities may be recovered from fiber material, such as by use of the fiber material directly as fiber mats, or cutting, chopping, or crushing the fiber material into small pieces for use in later processes like injection molding, cement, or for remelting the fiber material into new fibers or other products, and/or the like. Each of these commodities are recovered using different processes, involve different alterations to the material, etc. Consequently, the time, effort, and cost associated with recovering each commodity is different.

This may make it may be difficult for a seller to determine to which commodity (or commodities) to recover. For example, typical supply and demand for a commodity will fluctuate based on a number of different factors. The introduction of new methods of recycling composite materials may cause an increase in the supply of commodities that may be recovered from those materials. If there are multiple sellers in the market, the increase in supply may have the effect of causing prices for related commodities to drop. Other factors that can affect the market for a raw materials commodity include changes in laws or regulations of a geographic region, a reduction in carbon dioxide caused by replacing non-recycled commodities with recycled commodities, a change in transportation costs of these materials/commodities (e.g., which may occur based on a large increase in the supply of the commodities recovered from these materials), and/or the like. Each of these factors make it difficult for a seller to select an optimal commodity to recover. It is an object of the invention to mitigate or overcome some or all of the problems described above. The invention is particularly advantageous when the composite member is a wind turbine blade including amine cured epoxy resin.

In an aspect of the invention, a method of recommending one or more target commodities to recover using materials within a composite member of a wind turbine is disclosed.

To provide a specific example, the method may be applied to a wind turbine blade that uses epoxy resin and/or an epoxy resin system for bonding materials together in glass or carbon reinforced composites. The epoxy resin and/or epoxy resin system may include an amine cured epoxy resin, such as Olin Airstone 760, a Hexion RIMR 035C infusion epoxy, an Aditya Birla Recyclamine epoxy resin system, and/or the like. Such epoxy resins and/or epoxy resin systems allow for disassembling via swelling and/or dissolving in relatively mild conditions in acidic solutions comprising formic acid and/or acetic acid. For formic acid, a solution of at least 50 wt-% formic acid at ambient temperature and pressure was found to be effective. However, increased temperature, pressure and concentration increases reaction speed.

To provide another example, the method may be applied to a glass fiber reinforced composite recovered by separating a cured resin matrix from the fibers such that the fibers may be recovered as long individual fibers, fiber mats of woven or stitched fibers, chopped fibers allowing for injection molding, and/or glass fiber mass for re-melting. The cured resin matrix material may recovered as a particulate resin matrix or may be chemically depolymerized into monomers or oligomers. Alternatively - but less desirable from a recycling perspective as this constitutes down-cycling - the glass fiber reinforced composite may be crushed or ground particles for filler, e.g., in concrete, or as a combination of fuel and filler in the production of cement.

The method includes receiving, by a computing device, materials data that includes quality grade data indicating quality grades of respective materials included within a composite member. The method further includes receiving, by the computing device, commodity data for a plurality of candidate commodities that are capable of being recovered from the one or more materials of the composite member. The method further includes determining, by the computing device, expected profitability for each of said plurality of respective candidate commodities. The quality grade data for the respective materials and the commodity data for the plurality of candidate commodities may be provided as inputs to a data model to cause the data model to determine the expected profitability for each of said plurality of respective candidate commodities. The data model may be trained using machine learning based on historical quality grade data indicating quality grades of materials included in composite members, and historical commodity data for candidate commodities that were made from the materials in the composite members. The method further includes determining, by the computing device and based on the expected profitability, a recommendation indicating, for respective materials of the composite member, the one or more target commodities to recover using the respective materials. The method further includes delivering, from the computing device, the recommendation to another device or recipient, so that the target commodities can be recovered.

In an embodiment of the invention, the historical commodity data for the candidate commodities may include sales price and sales quantity data from an online marketplace in which one or more of the candidate commodities was sold. In this embodiment, the data model may be used to determine the expected profitability for a candidate commodity, of the candidate commodities, based at least in part on the sales price and sales quantity data from the online marketplace.

In another embodiment of the invention, the data model may be used to determine the expected profitability for a candidate commodity, of the plurality of candidate commodities. For example, the data model may be used to determine an expected cost associated with recovering the candidate commodity (e.g., using one of the materials of composite member). The data model may be further used to determine an expected sales price of the candidate commodity. The data model may be then used to determine the expected profitability for the candidate commodity based on the expected cost and the expected sales price.

In another embodiment of the invention, the expected sales price may be determined based on historical sales prices of the candidate commodity, and historical sales prices of non-recycled commodities that share one or more characteristics with the candidate commodity.

In another embodiment of the invention, determining the expected sales price may include processing the quality grade data to identify a quality grade of the material, and determining the expected sales price based at least in part on the quality grade of the material.

In another embodiment of the invention, determining the expected sales price may include determining a pricing adjustment based on the candidate commodity being a recycled product, and determining the expected sales price based at least in part on the pricing adjustment. In another embodiment of the invention, the data model may be trained using historical market data for a market corresponding to a candidate commodity, of the plurality of candidate commodities. In this embodiment, determining the expected profitability for the candidate commodity may include determining one or more indicators of change in the market of the candidate commodity based on the historical market data, and determining the expected profitability for the candidate commodity based at least in part on the one or more indicators of change.

In another embodiment of the invention, the one or more indicators of change may include an indicator associated with a change to a law or regulation for a geographic region corresponding to a location of the composite member, an indicator associated with a reduction in carbon dioxide emission caused by replacing non-recycled commodities with the candidate commodity, where the non-recycled commodities share one or more characteristics with the candidate commodity, an indicator associated with a change in a transportation cost of the candidate commodity, an indicator associated with a location of production of the candidate commodity, or an indicator associated with a location of sale of the candidate commodity.

In another embodiment of the invention, the data model may be trained using historical market data for one or more markets corresponding to the respective candidate commodities. In this embodiment, determining the profitability for a candidate commodity, of the candidate commodities may include: processing the historical market data to determine a maturity level of a market to which the candidate commodity belongs, selecting generalized machine learning features or specific machine learning features based on the maturity level of the market, and determining the profitability for the candidate commodity using selected features.

In another embodiment of the invention, determining the recommendation may include determining a set of instructions that specify a manner in which to alter a material, of the one or more materials of the composite member, such that a target commodity is recovered from the material. In another embodiment of the invention, determining the recommendation may include determining a set of chemical treatment instructions that specify a manner in which to alter a material, of the one or more materials of the composite, such that a target commodity is recovered from the material.

In another embodiment of the invention, the materials of the composite member may include cured epoxy resin. In this embodiment, the target commodities may include one or more of: a particulate resin matrix, one or more chemically depolymerized monomers, and one or more chemically depolymerized oligomers. When determining the recommendation, the method may include determining to recommend recovery of at least one of the particulate resin matrix, the one or more chemically depolymerized monomers, and the one or more chemically depolymerized oligomers.

In another aspect of the invention, a device performing any of the above methods is disclosed. The device, comprises one or more memories and one or more processors, communicatively coupled to the one or more memories, to receive materials data that includes quality grade data indicating quality grades of respective materials included within a composite member of a wind turbine; to receive commodity data for a plurality of candidate commodities that are capable of being recovered from the one or more materials of the composite member; to determine expected profitability for each of said plurality of respective candidate commodities, wherein the quality grade data for the respective materials and the commodity data for the plurality of candidate commodities are provided as inputs to a data model to cause the data model to determine the expected profitability for each of said plurality of respective candidate commodities. Wherein the data model has been trained using machine learning based on historical quality grade data indicating quality grades of materials included in composite members, and historical commodity data for candidate commodities that were made from the materials in the composite members. Furthermore, the one or more memories and one or more processors, communicatively coupled to the one or more memories to determine, based on the expected profitability, a recommendation indicating, for respective materials of the composite member, one or more target commodities to recover using the respective materials; and deliver the recommendation to another device or recipient.

In another aspect of the invention, a computer-readable medium performing any of the above methods is disclosed. The non-transitory computer-readable medium storing instructions, and the instructions comprising one or more instructions that, when executed by one or more processors, cause the one or more processors to receive materials data that includes quality grade data indicating quality grades of respective materials included within a composite member of a wind turbine; to receive commodity data for a plurality of candidate commodities that are capable of being recovered from the one or more materials of the composite member; to determine expected profitability for each of said plurality of respective candidate commodities, wherein the quality grade data for the respective materials and the commodity data for the plurality of candidate commodities are provided as inputs to a data model to cause the data model to determine the expected profitability for each of said plurality of respective candidate commodities. Wherein the data model has been trained using machine learning based on historical quality grade data indicating quality grades of materials included in composite members, and historical commodity data for candidate commodities that were made from the materials in the composite members. When executed by one or more processors, the instructions furthermore cause the one or more processors to determine, based on the expected profitability, a recommendation indicating, for respective materials of the composite member, one or more target commodities to recover using the respective materials; and deliver the recommendation to another device or recipient.

The above summary presents a simplified overview of some embodiments of the invention to provide a basic understanding of certain aspects of the invention discussed herein. The summary is not intended to provide an extensive overview of the invention, nor is it intended to identify any key or critical elements, or delineate the scope of the invention. The sole purpose of the summary is merely to present some concepts in a simplified form as an introduction to the detailed description presented below.

Brief Description of the Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate various embodiments of the invention and, together with the general description of the invention given above, and the detailed description of the embodiments given below, serve to explain the embodiments of the invention.

Fig. 1 shows a wind turbine in which the products of the invention could be advantageously used.

Fig. 2 shows a cross-sectional view of a rotor blade of a wind turbine.

Fig. 3A is a diagram of a recycling management platform receiving historical data that is to be used to train a data model.

Fig. 3B is a diagram of the recycling management platform determining features to use to train the data model.

Fig. 3C is a diagram of the recycling management platform training the data model.

Fig. 4A is a diagram of the recycling management platform receiving a request for a recommendation from a user device. Fig. 4B is a diagram of the recycling management platform using the data model to determine expected profitability of candidate commodities that can be recovered from materials of a composite member.

Fig. 4C is a diagram of the recycling management platform determining a recommendation based on the expected profitability of the candidate commodities and providing the recommendation to a user device.

Fig. 5 is a diagram of an example environment in which systems and/or methods described herein may be implemented.

Fig. 6 is a diagram of example components of one or more devices of Fig. 5.

It should be understood that the appended drawings are not necessarily to scale, and may present a somewhat simplified representation of various features illustrative of the basic principles of the invention. The specific design features of the sequence of operations as disclosed herein, including, for example, specific dimensions, orientations, locations, and shapes of various illustrated components, may be determined in part by the particular intended application and use environment. Certain features of the illustrated embodiments may have been enlarged or distorted relative to others to facilitate visualization and a clear understanding.

Detailed Description

The following detailed description of example embodiments refers to the accompanying drawings that show, by way of illustration, specific details and embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. Other embodiments may be utilized and structural, logical, and electrical changes may be made without departing from the scope of the invention as defined in the appended claims.

Fig. 1 shows a wind turbine 100 in which the products of the invention could be advantageously used. The wind turbine 100 includes a tower 102, a nacelle 104 on which a rotor 106 is mounted, and a yaw bearing 108. The yaw bearing 108 rotatably connects the tower 102 to the nacelle 104. The rotor 106 comprises a rotor hub to which, in this wind turbine 100, three rotor blades (shown as rotor blades 110, 112, and 114, respectively) are attached. The yaw bearing 108 is configured such that nacelle 104, together with the rotor 106 mounted thereon, is rotatable relative to the tower 102. As such, the nacelle 104 may be arranged such that the rotor 106 is oriented towards the wind. When oriented towards the wind, the wind turbine 100 is operable to generate more electricity.

Although one rotor 106 is illustrated in Fig. 1 , it will be appreciated that multiple rotors may be carried by the tower 102. For example, a two-rotor configuration can be seen in US Pub. No. 2022/0025866 and a configuration with two pairs of rotors can be seen in WO-A1-2018/157897. Thus, aspects of the invention may apply to multi-rotor wind turbines or other types of wind turbines and should not be limited to that shown in Fig. 1.

Fig. 2 is a cross-sectional view of a rotor blade 200 of a wind turbine. In some embodiments, the rotor blade 200 may be a composite member as described herein. In some embodiments, the rotor blade 200 may correspond to rotor blade 110, rotor blade 112, and/or rotor blade 114. The rotor blade 200 has a plurality of spar caps 212, 214, 216, and 218. The rotor blade 200 has an outer shell 208, which is fabricated from two half shells 204, 206. The shells 204, 206 are moulded from glass-fiber reinforced plastic (GRP). Parts of the outer shell 208 are of a sandwich panel construction and comprise a blade core of lightweight form (e.g., polyurethane), which is sandwiched between inner and outer GRP layers or ‘skins.’

The rotor blade 200 comprises first and second pairs of spar caps 212, 214, 216, and 218, arranged between sandwich panel regions of the outer shell 208. One spar cap of each pair is integrated with the windward shell and the other spar cap of each pair is integrated with the leeward shell. The spar caps 212, 214, 216, and 218 of the respective pairs are mutually opposed and extend longitudinally along the length of the rotor blade 200. A first longitudinally extending shear web 220 bridges the first pair of spar caps 212, 214 and a second longitudinally extending shear web 222 bridges the second pair of spar caps 216, 218. The shear webs 220, 222 in combination with the spar caps 212, 214, 216, 218 form a pair of I-beam structures, which transfer loads effectively from the rotor blade 200 to the hub of the wind turbine (not shown). The spar caps 212, 214, 216, 218 in particular transfer tensile and compressive bending loads while the shear webs 220, 222 transfer shear stresses in the rotor blade 200.

In some embodiments, a spar cap (e.g., spar cap 212, 214, 216, 218) may be constructed using a stack of pultruded strips that are bonded together. The pultruded strips are bonded together by an adhesive, such as a resin (e.g., an epoxy resin). The pultruded strips are arranged in the stack with its respective adherend surfaces in mutually opposed relation. The adhesive may be applied directly to the adherend surfaces of the pultruded strip or via another technique such as a resin infusion process. In an infusion process, liquid resin is supplied to the stack, and the resin infuses between the opposed adherend surfaces of the pultruded strips. The wind turbine blade 200 described above is exemplary, and wind turbine blades have a wide range of designs and constructions. Thus, it should be understood that aspects of the invention may apply to wind turbine arrangements other than that described above.

Figs. 3A-3C are diagrams of an example method 300 for using machine learning to train a data model to recommend one or more commodities to recover using materials of a composite member. As used herein, recovering a commodity from a material may refer to any structural or physical alteration that, once made to the material, allows the candidate commodity to be recovered from the material.

The composite member described herein may be described as a rotor blade of a wind turbine. It is to be understood that this is provided by way of example. In practice, example method 300 may be used to train a data model to recommend commodities that are part of another type of composite member or assembly. For example, a composite member may refer to a specific component of a rotor blade, such as a spar cap, a specific component of a spar cap, such as a pultruded strip, etc. To provide another example, example method 300 may be used to train a data model to recommend commodities in situations where the composite member is another device, assembly, component, and/or sub-component. In these examples, the composite member may refer to a printed circuit board (PCB), a component or sub-component of the PCB, a vehicle (e.g., a boat, a train, a car, etc.), vehicle assembly, a component or sub-component of the vehicle (e.g., the body of a car or hull of a boat), and/or the like.

As shown in Fig. 3A, the recycling management platform 302 may receive historical data that is to be used to train the data model. For example, and as shown by reference number 304, the recycling management platform 302 may receive, from a data storage device, historical materials data for materials used within rotor blades (i.e. , composite members) of wind turbines. The historical materials data may be received over a network (e.g., the Internet, etc.) via a communication interface, such as an application programming interface (API) or similar type of interface. The historical materials data for respective materials may include a material identifier, type data, measurement data, quality grade data, component data, and/or the like.

The material identifier may be used to identify a specific material, such as via a numeric code, an alphanumeric code, and/or the like. The type data may include data indicating a type of material, such as a core element (e.g., a foam, a wood, etc.), a type of fiber (e.g., a synthetic fiber, a semi-synthetic fiber, a regenerated fiber, a carbon fiber, a basalt fiber, a glass fiber, a metal fiber, etc.), a type of adhesive (e.g., an epoxy resin, etc.), a type of material used within a lightning protection grid, a type of paint, and/or the like.

The measurement data may include data that identifies (or can be used to identify) boundaries of respective materials within the rotor blade (and/or within a given component of the rotor blade). For example, the measurement data may include data identifying a length of a material, a height of the material, a width of the material, an area and/or surface area of the material, and/or the like. In some embodiments, the recycling management platform 302 may process the measurement data to add the material to the 3-D parts map (e.g., such as by having coordinates define boundaries of the material as that material is used within the rotor blade).

Additionally, or alternatively, the materials data may include manufacturing data that provides information on conditions under which respective materials were manufactured. For example, the manufacturing data for a material may include data relating to specific acts of manufacturing. To provide a specific example, when the material is cured, the manufacturing data collected may include temperature data indicating a temperature at which the material was cured, time data indicating a duration during which the material was cured, pressure data indicating a pressure that the material was exposed to during a curing process, data indicating a difference between a measured characteristic of the material during the curing process and a baseline characteristic (e.g., which may be the standard/norm for curing that type of material), and/or the like. Similar types of data may be collected for other acts or processes associated with manufacturing. To the extent these acts or processes associated with manufacturing apply to specific components of the rotor blade, similar types of data may be collected for respective components of the rotor blade.

The quality grade data may include data identifying a quality grade for a material. In some embodiments, the quality grade data may include data identifying a quality grade for a specific portion or region within the material. For example, throughout the term of life of a material, different parts of that material may experience varied amounts of wear and tear. In this case, the quality grade data for that material may identify a quality grade for each part of the material that has a unique quality grade.

In some embodiments, the materials data may further include information describing properties and/or characteristics of the material. Additionally, or alternatively, the materials data may include information describing one or more components in which a material is used. For example, a material may be used in multiple components and/or sub-components of a rotor blade. In this case, the materials data may include component data identifying one or more components and/or sub-components in which the material was used.

As shown by reference number 306, the recycling management platform 302 may receive historical commodity data for recycled commodities that were recovered from the materials used within the rotor blades. For example, online marketplaces may provide records of transactions between buyers and sellers of commodities. The historical commodity data may be obtained using a data mining technique, as described further herein. The historical commodity data may be stored using a data storage device, and may be provided to the recycling management platform 302 when it is time to train the data model. The historical commodity data may be provided over a network (e.g., the Internet) via a communication interface, such as an API or a similar type of interface. The historical commodity data for a recycled commodity may include a commodity identifier, type data, description data, image data, sales price data, sales quantity data, status data, location data, and/or the like.

The commodity identifier may be used to identify a specific commodity, such as via a numeric code, an alphanumeric code, and/or the like. The commodity type data may include data indicating a type of commodity. The commodity description may include data describing one or more characteristics and/or properties of the commodity, such as one or more materials used within the commodity, a manner in which the commodity is altered or prepared (e.g., such that the commodity is recovered from a material), and/or the like. The image data may include data representing an image of the commodity and/or image metadata relating to the image of the commodity.

The sales price data may include price offer data indicating a price or price range at which a commodity was offered for sale, price acceptance data indicating a price at which the commodity was sold, etc. The sales quantity data may include data indicating an amount or volume of the commodity that was sold. The status data may indicate a status relating to the availability of the commodity for purchase. For example, if an entire supply of a commodity has been sold, a status for that commodity, as listed on an online marketplace, may be updated to reflect that the commodity is out of stock. The location data may include data indicating one or more locations in which the commodity is being sold. For example, the location data may indicate a country in which the commodity is being sold, a state or providence in which in the commodity is being sold, a geographic region in which the commodity is being sold, and/or the like. In some embodiments, in addition to receiving historical commodity data for commodities recovered from materials of the rotor blades, the recycling management platform 302 may receive historical commodity data for non-recycled commodities that share one or more characteristics with the recycled commodities. For example, the market for a commodity may involve sales of a recycled version of that commodity but may also involve sales of a nonrecycled and competing version of that commodity. As such, the recycling management platform 302 may receive historical commodity data for those non-recycled commodities that are in competition with the recycled commodity, that share one or more characteristics or properties with the recycled commodity, etc. A recycled commodity, as used hereafter, may be referred to as just a commodity, or as a candidate commodity.

In some embodiments, the recycling management platform 302 may obtain the historical commodity data from one or more online marketplaces. For example, the recycling management platform 302 may obtain the historical commodity data from the one or more online marketplaces by executing a data mining technique (often referred to as crawling the web). After obtaining the historical commodity data, the recycling management platform 302 may store the data using a data storage device, such that the recycling management platform 302 is able to access the historical commodity data when training the data model.

In some embodiments, the historical commodity data obtained from an online marketplace may not have a format, file type, etc., that is conducive to training the data model. In this case, the recycling management platform 302 may use one or more natural language processing techniques to analyze the historical commodity data. This may allow the recycling management platform 302 to identify respective types of historical commodity data and/or to convert said historical commodity data into a uniform file type and/or file format that is conducive for training the data model.

The historical product data provided above is provided by way of example. In practice, other types of historical product data characterizing commodities involved in prior commercial activity may be provided to the recycling management platform 302. For example:

- The genus or category of the commodity

- The appearance / form of the commodity (e.g., powder, liquid, granule, white crystal powder, spray, etc.).

- The commodity usage (electronic chemicals, plastic auxiliary agents, coating auxiliary agents, rubber auxiliary agents, water treatment chemicals, etc.). - The business of the seller and purchaser (exporter, manufacturer, wholesales, importer, agent)

- Applicable commodity certificates or regulatory authorizations

- Applicable seller or purchaser certificates or authorizations

- Supplier location, categorized by country and/or city or region

- Supplier ranking or feedback score, and experience

- Supplier lead time

As shown by reference number 308, the recycling management platform 302 may receive historical market data for one or more markets in which the recycled commodities were exchanged. For example, each respective recycled commodity may be part of a market. A market may have fluctuations in supply and demand for any number of different reasons. For example, an increase in the supply of recyclable materials may lead to an increased supply in recycled commodities. This could have the effect of causing a decrease in the price of those recycled materials in the market. One of ordinary skill in the art can appreciate that changes in the market can occur for any number of different reasons. In order to be able to accurately predict whether it would be more advantageous (e.g., profitable) for a seller to recover a first commodity from a material, or to recover a second commodity from the same material, the recycling management platform 302 may receive historical market data (e.g., for markets of the first and second commodity, respectively) in order to be able to make recommendations based on recent changes in one or more of the market conditions and/or make recommendations based on predicted changes to one or more of the markets.

The historical market data for respective markets may include a market identifier, type data, description data, regulatory data, market maturity data, supply and demand data for customers and/or suppliers, distribution channels data, supply chain data, market segmentation data, transportation data, environment data, and/or the like.

The market ID may be used to identify a specific market, such as via a numeric code, an alphanumeric code, and/or the like. The market type data may include data indicating a type of market. For example, if recycled fibers are recovered in a certain way (e.g., cut, chopped, crushed, etc.), the recovered commodity, a fiber-based commodity, may be part of a market for a recovered fiber-based commodity. The market description data may include data describing the type of market, type of commodity bought and sold in the market, etc. The regulatory data may include data describing aspects of one or more laws or regulations that impact the market. Specifically, the historical market data may include data identifying a law or regulation that impacts the market, data indicating a geographic region in which the law or regulation applies, data describing the law or regulation, and/or the like. The data describing the law or regulation may, for example, include information describing a law or regulation about imports and/or exports for that geographic region, information describing a change to that law or regulation, and/or any other type of information capable of impacting the market.

The market maturity data may include a maturity identifier that identifies a maturity level of a market in which a commodity belongs, data describing the maturity level of the market, and/or the like. For example, if a commodity is part of a relatively new market, the data describing the maturity level of the market may specify that this is a new market, that the commodity has only recently been available for sale, etc.

The supply and demand data may include supply data indicative of a supply and demand of a commodity. For example, the supply and demand data may include sales data and/or commodities data for sales of a commodity in a given time period, purchasing data and/or commodities data for purchases of the commodity in the given time period, etc. The market segmentation data may include data used to divide the market into segments based on things like preferences, habits, trends, etc. For example, the market segmentation data may include data indicating a preference of a seller of a commodity, data indicating a preference of a buyer of a commodity, data indicating a habit or trend specific to a subset of the market (e.g., specific to the seller, the buyer, a geographic location or region, etc.), and/or the like.

The distribution channels data may include data indicating a number of distribution channels available within the market, data identifying each respective distribution channel, data describing each respective distribution channel, and/or the like. The transportation data may include data that identifies transportation costs for a geographic location or region. For example, the transportation data for a region may include data indicating one or more modes of transportation, data indicating a cost associated with respective modes of transportation, data indicating average costs associated with a mode of transportation over a given time period, and/or the like. The environment data may include data indicating a change in the environment which can be correlated with recycled commodities. For example, an increase in the supply of recycled commodities may lead to a reduction in carbon dioxide emission due to replacing non-recycled commodities with the recycled commodities. In some embodiments, the historical market data may be collected from auctions for batches of various recycled commodities in various markets. Additionally, or alternatively, the historical market data may be provided by a domain expert. Additionally, or alternatively, the historical market data may be obtained using a data mining technique. Additionally, or alternatively, the historical market data may be obtained using a third-party API that permits queries for market data.

In some embodiments (not shown), prior to using the historical data to train the data model, the recycling management platform 302 may perform one or more preprocessing operations to standardize the historical data into a uniform data type, data format, and/or the like. For example, the historical data may be received in different file types and/or formats and the recycling management platform 302 may apply appropriate standardization techniques to different data types or data formats, such that the historical data is converted into a uniform data type, data format, etc.

In some embodiments (not shown), the recycling management platform 302 may receive historical cost data. The historical cost data may information indicating one or more costs associated with recovering a commodity from a material of the composite member.

In this way, the recycling management platform 302 receives historical data for training the data model.

As shown in Fig. 3B, and by reference number 310, the recycling management platform 302 may select features to use to train a data model. For example, the recycling management platform 302 may identify a feature set that includes features capable of being used to train the data model and may select a subset of the identified features.

A feature may be a measurable property or characteristic that can be used to train the data model using machine learning. The feature set may include features for historical data values, features for aggregated historical data values, features for combinations of historical data values, features for benchmarks for one or more of these values, features for relationships between two or more these values, and/or the like. Specific example features are provided below. The features for materials data may, for example, include a feature identifying a type of material, a feature identifying a property or characteristic of the material, a feature identifying a measurement of a material, a feature identifying a quality grade of the material, and/or the like.

Similar features may be determined for each of the other types of historical data. For example, the feature set may include features relating to historical commodity data of commodities that were recovered from materials used within rotor blades. This may include features relating to commodity type data, description data, image data, sales price data, sales quantity data, status data, location data, and/or the like.

Additionally, the feature set may include features relating to historical market data, which may include features relating to markets in which the commodities were exchanged. This may include features relating to market type data, market description data, regulatory data, market maturity data, supply and demand data, distribution channels data, market segmentation data, transportation data, environment data, and/or the like.

As described above, the feature set may include features for aggregated historical data values, features for combinations of historical data values, features for benchmarks, and/or features for relationships between two or more of these values. For example, the features for the aggregated data values may include features relating to totals, averages, means, modes, and/or the like. The features for different combinations of historical data values may include features for all (or some) combinations of two or more different historical data values. The features for the benchmarks may include benchmarked values for different types of historical data. For example, a benchmarked value for sales price data may be a value indicating a preferred sales price.

The features for the relationships may include a value indicating a relationship or association between two or more other values (i.e., two or more historical data values, combinations of values, aggregations of values, etc.), a value indicating a trend found within the data, and/or the like. For example, these features may include a value indicating a relationship between sales price and quality grade of a material. Specifically, if a material is of a low quality grade, the price at which that material is sold is likely to be reduced relative to a material of average quality grade. Similarly, if a material is of a high quality grade, the price at which that material is sold is likely to be increased relative to a material of average quality grade. One skilled in the art can appreciate that a multitude of different relationships between historical data values can be identified. For example, features may be identified for relations between product sales/product volume and each respective type of market data.

In some embodiments, the recycling management platform 302 may identify the feature set by processing the historical data using one or more feature identification techniques. The one or more feature identification techniques may include a text mining and latent semantic analysis (LSA) technique, a trend variable analysis technique, an interest diversity analysis technique, a neural networking technique, a composite indicators analysis technique, a clustering analysis technique, and/or the like.

In some embodiments, the recycling management platform 302 may select a subset of the identified features by processing the feature set and/or the historical data using one or more feature selection techniques. The one or more feature selection techniques may include one or more filtering techniques, one or more wrapper techniques, one or more embedded techniques, and/or the like. The one or more filtering techniques may be used to remove duplicate, redundant features. These techniques may include a chi-square test, a correlation coefficient, a variance threshold, and/or the like. The one or more wrapper techniques may involve an iterative approach to feature selection (and/or reduction), and may include a forward selection technique, a backward elimination technique, a bi-directional elimination technique, an exhaustive selection technique, a recursive elimination technique, and/or the like. The one or more embedded techniques may include a regularization technique, a tree-based selection technique, and/or the like.

Additionally, or alternatively, the recycling management platform 302 may receive features to use to train the data model. For example, a feature set, or a subset of the feature set, may be determined by a domain expert and provided to (or made accessible to) the recycling management platform 302.

In this way, the recycling management platform 302 determines features to use to train the data model.

As shown in Fig. 3C, and by reference number 312, the recycling management platform 302 may train the data model to determine expected profitability of commodities that were recovered from a material of a composite member. For example, the recycling management platform 302 may train the data model by using one or more machine learning techniques to analyze the historical data and/or the determined features. The one or more machine learning techniques may include classification-driven training technique, a logistical regression-based training technique, a Naive Bayesian classifier technique, a support vector machine (SVM) technique, a neural network, and/or the like.

In some embodiments, the recycling management platform 302 may train the data model using an iterative approach. For example, the recycling management platform 302 may use the data model to process historical training data for a composite member such as a rotor blade to cause the data model to determine expected profitability scores for the commodities. The historical training data may exclude known outputs such that the data model is effectively making a prediction as to the expected profitability of a candidate commodity.

In some embodiments, the recycling management platform 302 may, using the data model, determine the expected profitability using the following equation: Expected Profitability = Expected Sales Price - Expected Cost +/- Adjustments. The expected sales price may be determined based on historical price data and any features found in the historical training data that are indicative of causing a change in sales price. The expected cost may be determined based on historical cost data and any features found in the historical training data that are indicative of causing a change in expected cost.

Next, the recycling management platform 302 may account for one or more adjustments that further warrant modifying the expected profitability value. For example, when a recycled commodity is sold on the market, it can receive a sales premium due to being a recycled (i.e. , green) product. The sales premium may vary depending on the geographic region in which the commodity was sold. The sales premium may also vary based on other factors, such as the laws or regulations in that region as they pertain to the exchange of raw materials. In one embodiment, the recycling management platform 302 may add the sales premium as a positive adjustment value that is used to update the current total profitability value. Continuing with the example, the recycling management platform 302 may further adjust the total profitability value based on a quality grade of the material. For example, recycled materials may have a lower quality grade than new materials and may be discounted when sold on the market. To account for this, the recycling management platform 302 may adjust the total profitability value based on the quality grade of the materials. In some embodiments, while the data model is being trained, the recycling management platform 302 may update weights of the data model based on the accuracy of the predictions. For example, if the recycling management platform 302 is unable to properly estimate the profitability of a composite material, the platform can adjust weights within the data model such that certain values are weighted less, or more, than during previous profitability predictions.

In this way, the recycling management platform 302 trains the data model to determine quality grades of materials based the expected reusability of the materials.

As indicated above, Figs. 3A-3C are provided merely as an example. Other examples may differ from what is described with reference to Figs. 3A-3C. For example, there may be additional devices and/or networks, fewer devices and/or networks, different devices and/or networks, or differently arranged devices and/or networks than those shown in Figs. 3A-3C. Furthermore, two or more devices shown in Figs. 3A-3C may be implemented within a single device, or a single device shown in Figs. 3A-3C may be implemented as multiple and/or distributed devices. Additionally, or alternatively, a set of devices (e.g., one or more devices) of example embodiment(s) 300 may perform one or more functions described as being performed by another set of devices of example method 300.

Figs. 4A-4C are diagrams of an example method 400 for using machine learning to determine expected profits of candidate commodities capable of being recovered from a composite member. As used herein, recovering a candidate commodity member may refer to any process implemented upon the composite member that allows the candidate commodity to be recovered from the member.

In one or more embodiments described herein, the composite member may be a rotor blade of a wind turbine. It is to be understood that this is provided by way of example. In practice, example method 400 may be used to determine expected profits of candidate commodities recoverable from of other components of a wind turbine. Additionally, or alternatively, example method 400 may be used to determine expected profits of candidate commodities recoverable from of other objects/devices/machines, such as a printed circuit board (PCB), a vehicle (e.g., a boat, a train, a car, etc.), other composite materials, and/or the like.

As shown in Fig. 4A, example method 400 involves data communications between a user device 402, a recycling management platform 404, and a data storage device 406. As shown by reference number 408, a user may input a request for a recommendation on which candidate commodities to recover from a rotor blade. For example, the user may interact with a user interface of the user device 402 to input the request for the recommendation. In some embodiments, the user interface may be part of an application and/or API that permits users to submit requests to the recycling management platform 404. When the user submits the request, request data may be provided to the recycling management platform 404. The request data may include a rotor blade identifier and/or identifiers for each respective material in the rotor blade.

As shown by reference number 410, the recycling management platform 404 may obtain data for respective materials of the rotor blade and commodities data for candidate commodities recoverable from the blade. For example, the recycling management platform 404 may, based on receiving the request data, provide a request for the materials data and/or the commodities data to the data storage device 406. Data storage device 406 may use a data structure to store the materials data and/or the commodities data in association with the rotor blade identifier and/or the material identifier for respective materials. As such, the data storage device 406 may use the received identifier(s) to identify and provide the recycling management platform 404 with the materials data and/or the commodities data.

In some embodiments, the recycling management platform 404 may obtain other types of historical data from the data storage device 406. For example, the recycling management platform 404 may obtain historical market data that provides market information relating to the markets of the candidate commodities. One or more types of data described herein may have predictive value only if the data collected is current. As such, different types of data may be recollected periodically (e.g., and used to re-train the data model), thereby allowing the recycling management platform 404 to make use of current data relating to commodities and corresponding markets.

As shown in Fig. 4B, and by reference number 412, the recycling management platform 404 may determine, using the data model, expected profitability of respective candidate commodities. For example, the recycling management platform 404 may provide the materials data, the commodities data, and/or any other collected data as input data to the data model. Training of the data model is described in connection with Figs. 3A-3C. The data model may process the input data using machine learning to determine, for each candidate commodity, an expected profitability for the seller were that candidate commodity to be recovered using one or more materials of the rotor blade.

In some embodiments, the recycling management platform 404 may use the data model to determine an expected profitability score for a candidate commodity. The expected profitability score may represent a profitability of a seller were the candidate commodity score to be recovered from a material. By processing the input data using the data model, the recycling management platform 404 may determine that the historical materials data includes quality grade data indicating that the material is of high or low quality, and commodities data that specifies recent prices at which the candidate commodity was sold in the market. Further, the recycling management platform 404 may determine that the market is new (e.g., based on receiving a low quantity of commodities data, based on receiving market maturity data indicating that the market is a new market, etc.). Each of these considerations may impact how the recycling management platform 404 computes the expected profitability score, how the recycling management platform 404 weights certain variables in an equation to determine the expected profitability score, etc.

In this way, the recycling management platform 404 determines the expected profitability of respective candidate commodities.

As shown in Fig. 4C, and by reference number 414, the recycling management platform 404 may determine, based on the output of the data model, a recommendation indicating target commodities to recover from the rotor blade. In addition to recommending target commodities to recover, the recommendation may include one or more sets of instructions indicating how to recover respective commodities.

In some embodiments, the recommendation may include a set of instructions that specify a manner in which to process materials from the rotor blade, such that a candidate commodity can be recovered from the rotor blade. The instructions may specify an amount of the material that is to be altered and/or a way in which the material is to be altered. For example, if the material is a fiber, then based on the use of the fiber in a given commodity, the fiber may need to be cut into larger portions for use in fiber mats, may need to be cut, chopped, or crushed into smaller-sized portions for uses in injection molding, cement, the fiber may need particularly good cleaning to remove metal elements to allow for better remelting for making of new fibers and/or the like. Additionally, or alternatively, the recommendation may include a set of chemical treatment instructions that specify a manner in which to alter a material of the rotor blade, such that a candidate commodity can be recovered from the material.

In some embodiments, the recycling management platform 404 may recommend one commodity to recover from a given material. In some embodiments, the recycling management platform 404 may recommend multiple commodities or a combination of commodities to recover from the given material. In some embodiments, the recycling management platform 404 may determine the recommendation by referencing a data structure. For example, the recycling management platform 404 may have access to a data structure that associates data identifying expected profitability scores with commodity identifiers for respective candidate commodities, and with instructions such as those described above. The recycling management platform 404 may reference the data structure to identify the target commodities to recommend and to determine which instructions to include in the recommendation.

As shown by reference number 416, the recycling management platform 404 may deliver the recommendation to the user device 402. In some embodiments, the recycling recommendation may be delivered to another recipient, such as an electronic mail (email) account or related account. As shown by reference number 418, the user device 402 may display the recommendation. This may allow the user to use the recommendation to implement the process of recovering recommended commodities, such that recommended commodities may be offered for sale on the market.

In some embodiments, the recommendation may be provided to a controller (e.g., a computing device with a processor) that is part of an automated or semi-automated recycling facility. The automated or semi-automated recycling facility may have one or more pieces of equipment configured to automatically perform certain recycling tasks. The controller may be configured to send instructions to respective pieces of equipment to permit the equipment to perform the recycling tasks autonomously. In this case, the controller may use the recommendation to generate instructions that permit or cause respective pieces of equipment to perform recycling tasks such as cutting a material in a certain way, placing a material into a certain recycling container, and/or the like.

In some embodiments, the recommendation may be provided to a user device of an operator of recycling equipment. For example, the recommendation may be provided to and displayed on a user interface of the user device. The user interface may display instructions that the operator can use to perform a recycling task, such as the cut instructions and/or recycling container instructions described herein.

As indicated above, Figs. 4A-4C are provided merely as an example. Other examples may differ from what is described with reference to Figs. 4A-4C. For example, there may be additional devices and/or networks, fewer devices and/or networks, different devices and/or networks, or differently arranged devices and/or networks than those shown in Figs. 4A-4C. Furthermore, two or more devices shown in Figs. 4A-4C may be implemented within a single device, or a single device shown in Figs. 4A-4C may be implemented as multiple and/or distributed devices. Additionally, or alternatively, a set of devices (e.g., one or more devices) of example method 400 may perform one or more functions described as being performed by another set of devices of example embodiment(s) 400.

Fig. 5 is a diagram of an example environment 500 in which systems and/or methods described herein may be implemented. As shown in Fig. 5, environment 500 may include a user device 502, a data storage device 504, a recycling management platform 506 supported within a cloud computing environment 508, and/or a network 512. Devices of environment 500 may interconnect via wired connections, wireless connections, or a combination of wired and wireless connections. User device 502 may correspond to user device 402. Data storage device 504 may correspond to data storage device 406. Recycling management platform 506 may correspond to recycling management platform 302 and/or recycling management platform 404.

User device 502 includes one or more devices capable of receiving, generating, storing, processing, and/or providing information associated with a recommendation. User device 502 may include a device, such as a tablet computer (e.g., an iPad, etc.), a mobile phone (e.g., a smart phone, a radiotelephone, etc.), a laptop computer, a handheld computer, a server computer, a gaming device, a wearable communication device (e.g., a smart wristwatch, a pair of smart eyeglasses, etc.), or a similar type of device. In some embodiments, the user device 502 may provide a request for a recommendation to the recycling management platform 506. In some embodiments, the user device 502 may receive a recommendation from the recycling management platform 506.

Data storage device 504 includes one or more devices capable of receiving, storing, processing, and/or providing historical data. Data storage device 504 may include a server device or a group of server devices. In some embodiments, data storage device 504 may support a data structure that associates different types of historical data. In some embodiments, data storage device 504 may receive a request (e.g., a query) for historical data from the recycling management platform 506. This may cause the data storage device 504 to provide the recycling management platform 506 with the historical data.

Recycling management platform 506 includes one or more devices capable of receiving, storing, processing, and/or providing information associated with a composite member. Recycling management platform 506 may include a server device (e.g., a host server, a web server, an application server, etc.), a data center device, or a similar device.

In some embodiments, the recycling management platform 506 may receive a request for a recommendation from the user device 502. In some embodiments, the recycling management platform 506 obtains, from data storage device 504, historical data associated with one or more rotor blades of a wind turbine. For example, the recycling management platform 506 may provide the data storage device 504 with a request for historical data. This may cause the data storage device 504 to provide the historical data to the recycling management platform 506.

In some embodiments, the recycling management platform 506 stores or has access to a data model that has been trained using machine learning. In some embodiments, the recycling management platform 506 may train the data model using the historical data of composite members such as a group of wind turbines. In some embodiments, the recycling management platform 506 may receive a trained data model from another device. In some embodiments, the recycling management platform 506 may provide a recommendation to the user device 502.

In some embodiments, as shown, the recycling management platform 506 may be hosted in the cloud computing environment 508. Notably, while embodiments described herein describe the recycling management platform 506 as being hosted in the cloud computing environment 508, in some embodiments, the recycling management platform 506 may not be cloud-based (i.e. , may be implemented outside of a cloud computing environment) or may be partially cloudbased.

Cloud computing environment 508 includes an environment that hosts recycling management platform 506. Cloud computing environment 508 may provide computation, software, data access, storage, etc. services that do not require end-user knowledge of a physical location and configuration of system(s) and/or device(s) that hosts the recycling management platform 506. As shown, the cloud computing environment 508 may include a group of computing resources 510 (referred to collectively as “computing resources 510” and individually as “computing resource 510”).

Computing resource 510 includes one or more personal computers, workstation computers, server devices, or another type of computation and/or communication device. In some embodiments, the computing resource 510 may host the recycling management platform 506. The cloud resources may include compute instances executing in the computing resource 510, storage devices provided in the computing resource 510, data transfer devices provided by the computing resource 510, and/or the like. In some embodiments, the computing resource 510 may communicate with other computing resources 510 via wired connections, wireless connections, or a combination of wired and wireless connections.

As further shown in Fig. 5, computing resource 510 may include a group of cloud resources, such as one or more applications (APPs) 510a, one or more virtual machines (VMs) 510b, virtualized storage (VSs) 510c, one or more hypervisors (HYPs) 510d, and/or the like.

Application 510a may include one or more software applications that may be provided to or accessed by the user device 502 and/or the recycling management platform 506. Application 510a may eliminate a need to install and execute the software applications on these devices. In some embodiments, one application 510a may send/receive information to/from one or more other applications 510a, via virtual machine 510b. In some embodiments, application 510a may be a recycling management application and/or an application directed toward recommending a commodity to recover using materials of a composite member. The recycling management application may include one or more user interfaces that, when displayed on the user device 502, permit a user to submit a request for a recycling recommendation.

Virtual machine 510b may include a software implementation of a machine (e.g., a computer) that executes programs like a physical machine. Virtual machine 510b may be either a system virtual machine or a process virtual machine, depending upon use and degree of correspondence to any real machine by virtual machine 510b. A system virtual machine may provide a complete system platform that supports execution of a complete operating system (“OS”). A process virtual machine may execute a single program and may support a single process. In some embodiments, virtual machine 510b may execute on behalf of another device (e.g., user device 502), and may manage infrastructure of the cloud computing environment 508, such as data management, synchronization, or long-duration data transfers.

Virtualized storage 510c may include one or more storage systems and/or one or more devices that use virtualization techniques within the storage systems or devices of the computing resource 510. In some embodiments, within the context of a storage system, types of virtualizations may include block virtualization and file virtualization. Block virtualization may refer to abstraction (or separation) of logical storage from physical storage so that the storage system may be accessed without regard to physical storage or heterogeneous structure. The separation may permit administrators of the storage system flexibility in how the administrators manage storage for end users. File virtualization may eliminate dependencies between data accessed at a file level and a location where files are physically stored. This may enable optimization of storage use, server consolidation, and/or performance of non-disruptive file migrations.

Hypervisor 51 Od may provide hardware virtualization techniques that allow multiple operating systems (e.g., “guest operating systems”) to execute concurrently on a host computer, such as computing resource 510. Hypervisor 51 Od may present a virtual operating platform to the guest operating systems and may manage the execution of the guest operating systems.

Network 512 may include one or more wired and/or wireless networks. For example, network 512 may include a cellular network (e.g., a fifth generation (5G) network, a fourth generation (4G) network, such as a long-term evolution (LTE) network, a third generation (3G) network, and/or a code division multiple access (CDMA) network), a public land mobile network (PLMN), a local area network (LAN), a wide area network (WAN), a metropolitan area network (MAN), a telephone network (e.g., the Public Switched Telephone Network (PSTN)), a private network, an ad hoc network, an intranet, the Internet, a fiber optic-based network, a cloud computing network, or the like, and/or a combination of these or other types of networks.

The number and arrangement of devices and networks shown in Fig. 5 are provided as an example. In practice, there may be additional devices and/or networks, fewer devices and/or networks, different devices and/or networks, or differently arranged devices and/or networks than those shown in Fig. 5. Furthermore, two or more devices shown in Fig. 5 may be implemented within a single device, or a single device shown in Fig. 5 may be implemented as multiple, distributed devices. Additionally, or alternatively, a set of devices (e.g., one or more devices) of environment 500 may perform one or more functions described as being performed by another set of devices of environment 500.

Fig. 6 is a diagram of example components of a device 600. Device 600 may correspond to the user device 502, the data storage device 504, and/or the recycling management platform 506. In some embodiments, the user device 502, the data storage device 504, and/or the recycling management platform 506 may include one or more devices 600 and/or one or more components of device 600. As shown in Fig. 6, device 600 may include a bus 602, a processor 604, a memory 606, a storage component 608, an input component 610, an output component 612, and/or a communication interface 614. Bus 602 includes a component that permits communication among multiple components of device 600. Processor 604 is implemented in hardware, firmware, and/or a combination of hardware and software. Processor 604 includes a central processing unit (CPU), a graphics processing unit (GPU), an accelerated processing unit (APU), a microprocessor, a microcontroller, a digital signal processor (DSP), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), and/or another type of processing component. In some embodiments, processor 604 includes one or more processors capable of being programmed to perform a function. Memory 606 includes a random-access memory (RAM), a read only memory (ROM), and/or another type of dynamic or static storage device (e.g., a flash memory, a magnetic memory, and/or an optical memory) that stores information and/or instructions for use by processor 604.

Storage component 608 stores information and/or software related to the operation and use of device 600. For example, storage component 608 may include a hard disk (e.g., a magnetic disk, an optical disk, a magneto-optic disk, and/or a solid-state disk), a compact disc (CD), a digital versatile disc (DVD), a floppy disk, a cartridge, a magnetic tape, and/or another type of non-transitory computer-readable medium, along with a corresponding drive.

Input component 610 includes a component that permits device 600 to receive information, such as via user input (e.g., a touch screen display, a keyboard, a keypad, a mouse, a button, a switch, and/or a microphone). Additionally, or alternatively, input component 610 may include a sensor for sensing information (e.g., a global positioning system (GPS) component, an accelerometer, a gyroscope, and/or an actuator). Output component 612 includes a component that provides output information from device 600 (e.g., a display, a speaker, and/or one or more light-emitting diodes (LEDs)).

Communication interface 614 includes a transceiver-like component (e.g., a transceiver and/or a separate receiver and transmitter) that enables device 600 to communicate with other devices, such as via a wired connection, a wireless connection, or a combination of wired and wireless connections. Communication interface 614 may permit device 600 to receive information from another device and/or provide information to another device. For example, communication interface 614 may include an Ethernet interface, an optical interface, a coaxial interface, an infrared interface, a radio frequency (RF) interface, a universal serial bus (USB) interface, a WiFi interface, a cellular network interface, or the like. Device 600 may perform one or more processes described herein. Device 600 may perform these processes based on processor 604 executing software instructions stored by a non- transitory computer-readable medium, such as memory 606 and/or storage component 608. A computer-readable medium is defined herein as a non-transitory memory device. A memory device includes memory space within a single physical storage device or memory space spread across multiple physical storage devices.

Software instructions may be read into memory 606 and/or storage component 608 from another computer-readable medium or from another device via communication interface 614. When executed, software instructions stored in memory 606 and/or storage component 608 may cause processor 604 to perform one or more processes described herein. Additionally, or alternatively, hardwired circuitry may be used in place of or in combination with software instructions to perform one or more processes described herein. Thus, embodiments described herein are not limited to any specific combination of hardware circuitry and software.

The number and arrangement of components shown in Fig. 6 are provided as an example. In practice, device 600 may include additional components, fewer components, different components, or differently arranged components than those shown in Fig. 6. Additionally, or alternatively, a set of components (e.g., one or more components) of device 600 may perform one or more functions described as being performed by another set of components of device 600.

The foregoing disclosure provides illustration and description but is not intended to be exhaustive or to limit the embodiments to the precise form disclosed. Modifications and variations may be made in light of the above disclosure or may be acquired from practice of the embodiments.

Some embodiments are described herein in connection with thresholds. As used herein, satisfying a threshold may refer to a value being greater than the threshold, more than the threshold, higher than the threshold, greater than or equal to the threshold, less than the threshold, fewer than the threshold, lower than the threshold, less than or equal to the threshold, equal to the threshold, etc., depending on the context.

Certain user interfaces have been described herein and/or shown in the figures. A user interface may include a graphical user interface, a non-graphical user interface, a text-based user interface, etc. A user interface may provide information for display. In some embodiments, a user may interact with the information, such as by providing input via an input component of a device that provides the user interface for display. In some embodiments, a user interface may be configurable by a device and/or a user (e.g., a user may change the size of the user interface, information provided via the user interface, a position of information provided via the user interface, etc.). Additionally, or alternatively, a user interface may be pre-configured to a standard configuration, a specific configuration based on a type of device on which the user interface is displayed, and/or a set of configurations based on capabilities and/or specifications associated with a device on which the user interface is displayed.

It will be apparent that systems and/or methods, described herein, may be implemented in different forms of hardware, firmware, and/or a combination of hardware and software. The actual specialized control hardware or software code used to implement these systems and/or methods is not limiting of the embodiments. Thus, the operation and behavior of the systems and/or methods were described herein without reference to specific software code - it being understood that software and hardware can be used to implement the systems and/or methods based on the description herein.

No element, act, or instruction used herein should be construed as critical or essential unless explicitly described as such. Also, as used herein, the articles “a” and “an” are intended to include one or more items and may be used interchangeably with “one or more.” Furthermore, as used herein, the term “set” is intended to include one or more items (e.g., related items, unrelated items, a combination of related and unrelated items, etc.), and may be used interchangeably with “one or more.” Where only one item is intended, the phrase “only one” or similar language is used. Also, as used herein, the terms “has,” “have,” “having,” or the like are intended to be open-ended terms. Further, the phrase “based on” is intended to mean “based, at least in part, on” unless explicitly stated otherwise.

While all the invention has been illustrated by a description of various embodiments, and while these embodiments have been described in considerable detail, it is not the intention of the Applicant to restrict or in any way limit the scope of the appended claims to such detail. Additional advantages and modifications will readily appear to those skilled in the art. The invention in its broader aspects is therefore not limited to the specific details, representative apparatus and method, and illustrative examples shown and described. Accordingly, departures may be made from such details without departing from the spirit or scope of the Applicant’s general inventive concept.

Claims

1. A method of recovering one or more target commodities from materials included within a composite member of a wind turbine, comprising: receiving, by a computing device, materials data that includes quality grade data indicating quality grades of respective materials included within the composite member; receiving, by the computing device, commodity data associated with a plurality of candidate commodities that are capable of being recovered from the one or more materials of the composite member; determining, by the computing device, expected profitability for each of said plurality of respective candidate commodities, wherein the quality grade data for the respective materials and the commodity data for the plurality of candidate commodities are provided as inputs to a data model to cause the data model to determine the expected profitability for each of said plurality of respective candidate commodities, and wherein the data model has been trained using machine learning based on: historical quality grade data indicating quality grades of materials included in composite members, and historical commodity data for candidate commodities that were made from the materials in those composite members; determining, by the computing device and based on the expected profitability, a recommendation indicating, for respective materials of the composite member, the one or more target commodities to recover using the respective materials; delivering, from the computing device, the recommendation to another device or recipient; and recovering the target commodities from the materials within the wind turbine composite member in a manner consistent with the recommendation.

2. The method of claim 1 , wherein the historical commodity data for the candidate commodities includes sales price and sales quantity data from an online marketplace in which one or more of the candidate commodities was sold, and wherein determining the expected profitability comprises: determining, using the data model, the expected profitability for a candidate commodity, of the plurality of candidate commodities, based at least in part on the sales price and sales quantity data from the online marketplace.

3. The method of claims 1 or 2, wherein determining the expected profitability for a candidate commodity, of the plurality of candidate commodities, comprises: determining, using the data model, an expected cost associated with recovering the candidate commodity using a material, of the one or more materials of the composite member, determining, using the data model, an expected sales price of the candidate commodity, and determining the expected profitability for the candidate commodity based on the expected cost and the expected sales price.

4. The method of claim 3, wherein determining the expected sales price comprises: determining the expected sales price based on historical sales prices of the candidate commodity, and historical sales prices of non-recycled commodities that share one or more characteristics with the candidate commodity.

5. The method of claim 3 or 4, wherein determining the expected sales price comprises: processing the quality grade data to identify a quality grade of the material, and determining the expected sales price based at least in part on the quality grade of the material.

6. The method of claim 3 or 4, wherein determining the expected sales price comprises: determining a pricing adjustment based on the candidate commodity being a recycled product, and determining the expected sales price based at least in part on the pricing adjustment.

7. The method of any of the preceding claims, wherein the data model has been trained using historical market data for a market corresponding to a candidate commodity, of the plurality of candidate commodities, and wherein determining the expected profitability for the candidate commodity comprises: determining one or more indicators of change in the market of the candidate commodity based on the historical market data, and determining the expected profitability for the candidate commodity based at least in part on the one or more indicators of change.

8. The method of claim 7, wherein the one or more indicators of change include at least one of: an indicator associated with a change to a law or regulation for a geographic region corresponding to a location of the composite member, an indicator associated with a reduction in carbon dioxide emission caused by replacing non-recycled commodities with the candidate commodity, where the non-recycled commodities share one or more characteristics with the candidate commodity, an indicator associated with a change in a transportation cost of the candidate commodity, an indicator associated with a location of production of the candidate commodity, or an indicator associated with a location of sale of the candidate commodity.

9. The method of any of the preceding claims, wherein the data model has been trained using historical market data for one or more markets corresponding to the respective candidate commodities, and wherein determining the profitability for a candidate commodity, of the candidate commodities, comprises: processing the historical market data to determine a maturity level of a market to which the candidate commodity belongs, selecting generalized machine learning features or specific machine learning features based on the maturity level of the market, and determining the profitability for the candidate commodity using selected features.

10. The method of any of the preceding claims, wherein determining the recommendation comprises: determining a set of instructions that specify a manner in which to alter a material, of the one or more materials of the composite member, such that a target commodity is recovered from the material.

11 . The method of any of the preceding claims, wherein the materials of the composite member include cured epoxy resin, and wherein the target commodities include one or more of: a particulate resin matrix, one or more chemically depolymerized monomers, and one or more chemically depolymerized oligomers; and wherein determining the recommendation comprises: determining to recommend recovery of at least one of the particulate resin matrix, the one or more chemically depolymerized monomers, and the one or more chemically depolymerized oligomers.

12. The method of any of the preceding claims, wherein determining the recommendation comprises: determining a set of chemical treatment instructions that specify a manner in which to alter a material, of the one or more materials of the composite member, such that a target commodity is recovered from the material.

13. A device, comprising: one or more memories; and one or more processors, communicatively coupled to the one or more memories, to: receive materials data that includes quality grade data indicating quality grades of respective materials included within a composite member of a wind turbine; receive commodity data for a plurality of candidate commodities that are capable of being recovered from the one or more materials of the composite member; determine expected profitability for each of said plurality of respective candidate commodities, wherein the quality grade data for the respective materials and the commodity data for the plurality of candidate commodities are provided as inputs to a data model to cause the data model to determine the expected profitability for each of said plurality of respective candidate commodities, and wherein the data model has been trained using machine learning based on: historical quality grade data indicating quality grades of materials included in composite members, and historical commodity data for candidate commodities that were made from the materials in the composite members; determine, based on the expected profitability, a recommendation indicating, for respective materials of the composite member, one or more target commodities to recover using the respective materials; and deliver the recommendation to another device or recipient.

14. A non-transitory computer-readable medium storing instructions, the instructions comprising: one or more instructions that, when executed by one or more processors, cause the one or more processors to: receive materials data that includes quality grade data indicating quality grades of respective materials included within a composite member of a wind turbine; receive commodity data for a plurality of candidate commodities that are capable of being recovered from the one or more materials of the composite member; determine expected profitability for each of said plurality of respective candidate commodities, wherein the quality grade data for the respective materials and the commodity data for the plurality of candidate commodities are provided as inputs to a data model to cause the data model to determine the expected profitability for each of said plurality of respective candidate commodities, and wherein the data model has been trained using machine learning based on: historical quality grade data indicating quality grades of materials included in composite members, and historical commodity data for candidate commodities that were made from the materials in the composite members; determine, based on the expected profitability, a recommendation indicating, for respective materials of the composite member, one or more target commodities to recover using the respective materials; and deliver the recommendation to another device or recipient.