A compounding system and a method of determining a feed to an extruder of such compounding system

Abstract

The present disclosure relates to a compounding system and a method of determining a feed to an extruder of such compounding system. The compounding system includes a hardware part and a virtual experimentation part. The hardware part includes one or more extruders to to obtain a mixture extrudate. The virtual experimentation part includes a trained machine learning model trained with training data of material properties of recycled polymer compositions, material properties of virgin polymer compositions and material properties of mixture extrudates obtained by mixing the recycled polymer compositions and the virgin polymer compositions. The machine learning model receives, during extrusion, material properties of the recycled polymer composition in the extruder and is arranged to generate, in response to processing said input properties, a recommendation of a virgin polymer composition to be fed to the extruder to obtain the mixture extrudate having desired material properties.

Description

[0001] 24POLY0036-WO-ORD 1

[0002] A compounding system and a method of determining a feed to an extruder of such compounding system

[0003] Technical field

[0004] The present disclosure relates to a method for determining feeds to an extruder, in particular for mixing a recycled polymer composition and a virgin polymer composition to obtain a mixture extrudate having desired material properties. The present disclosure further relates to a compounding system using such method.

[0005] Background

[0006] Demand for use of recycled materials as a feed for making polymer compositions is increasing. Recycled plastic materials can be recycled by mechanical recycling or chemical recycling. The most common method for the recycling of plastic materials is mechanical recycling (Al-Salem et al., 2009a). The process of mechanical recycling typically includes collection, sorting, washing and grinding of the material. Steps may occur in a different order, multiple times or not at all, depending on the origins and composition of the waste. The mechanically recycled polymer can be a post-consumer recycled (PCR) polymer or a post-industrial recycled (PIR) polymer. A post-consumer recycled (PCR) material refers to material that is made from items recycled by consumers. Usually, recyclable items such as plastics, metals and cardboards / paper are collected by the local recycling program and transported to facilities for sorting based on material type. Recycled bales will then be purchased and sent to different recyclers to be used for a variety of finished products. Using PCR / PIR content, namely recycled materials, for new products results in environmental benefits of carbon savings and resource efficiency. Also, it reduces the demand for virgin raw materials and improves the end-of-life of the materials. The inclusion of PCR / PIR into production supports the development of a circular economy.

[0007] Comparatively, a virgin polymer composition is a new, direct resin composition produced using natural gas or crude oil and without any recycled materials. A virgin polymer composition has not been subjected to any processing other than for its production.

[0008] In general, recycled polymer compositions are mixed with virgin polymer compositions for making polymer compositions having desired properties. However, due to relatively large variations in the material properties of recycled polymer compositions, mainly due to different time and location to obtain the recycled polymer 24POLY0036-WO-ORD 2 compositions, use of the same virgin polymer compositions for the same recycled polymer compositions can result in undesirable variations in the material properties of the obtained polymer compositions. The variations may be too large and the resulting mixtures may not have the desired properties. In order to overcome such undesirable variations, adjustments have to be made as to which virgin polymer compositions should be fed as well as their feeding ratio with respect to the recycled compositions by trial-and-error for obtaining mixtures with the desired properties.

[0009] Summary

[0010] A summary of aspects of certain examples disclosed herein is set forth below. It should be understood that these aspects are presented merely to provide the reader with a brief summary of these certain embodiments and that these aspects are not intended to limit the scope of this disclosure. Indeed, this disclosure may encompass a variety of aspects and / or a combination of aspects that may not be set forth.

[0011] It is an objective of the present disclosure to provide an improved method for determining feeds to an extruder for mixing a recycled polymer composition and a virgin polymer composition to obtain a mixture extrudate having desired material properties.

[0012] According to an aspect of the present disclosure, a method of determining a feed to an extruder to obtain a mixture extrudate having desired material properties is presented. The method may include i) providing a trained machine learning (ML) model trained with training data comprising one or more of material properties of recycled polymer compositions, material properties of virgin polymer compositions and material properties of mixture extrudates obtained by mixing the recycled polymer compositions and the virgin polymer compositions. The method may further include ii) feeding a recycled polymer composition and a virgin polymer composition to the extruder. The method may further include measuring, during extrusion, material properties of the recycled polymer composition in the extruder to obtain an input set of material properties of the recycled polymer composition. The method may further include iv) receiving, by the trained ML model, the input set of material properties of the recycled polymer composition obtained by step iii) and an input set of the desired material properties of the mixture extrudate. The method may further include v) processing the input sets using the trained ML model. The method may further include vi) generating, by the trained ML model in response to said processing of the input sets, a recommendation of the virgin polymer composition to be fed to the extruder to obtain the mixture extrudate having the desired material properties. 24POLY0036-WO-ORD 3

[0013] According to the method of the present disclosure, a recommendation may be made by the trained ML model as to which virgin polymer compositions should be fed to the extruder and / or their feeding ratio with respect to the recycled compositions depending on the desired material properties of the mixture extrudate.

[0014] Prior to performing the method of the present disclosure, training data for training a machine learning model may be created. Various combinations of different types of recycled polymer compositions and different types of virgin polymer compositions may be fed to an extruder to obtain mixture extrudates. Material properties of these mixture extrudates may be measured. Training data comprising data of the material properties of the mixture extrudates in relation to the material properties of the recycled polymer compositions and the virgin polymer compositions for making the mixture extrudates may be obtained.

[0015] The ML model may be trained by the training data to predict the material properties of mixture extrudates based on inputs of material properties of recycled polymer compositions and virgin polymer compositions.

[0016] A certain level of information about material properties of the recycled polymer composition may be provided, e.g., by product certificate of the recycled polymer composition from a supplier. Virgin polymer composition to be mixed with the recycled polymer composition may be decided (e.g., by human) based on the desired material properties of the mixture extrudate.

[0017] The recycled polymer composition may be fed to an extruder as well as the virgin polymer composition so decided. Material properties of the recycled polymer composition in the extruder may be measured during the extrusion (in-line measurement).

[0018] The ML model may receive an input on the desired material properties of mixture extrudates and the measured material properties of the recycled polymer composition in the extruder. The ML model may give a recommendation on which virgin polymer composition should be fed to the extruder based on the inputs to meet the desired final properties. This recommendation can change during extrusion depending on the measured material properties of the recycled polymer composition in the extruder.

[0019] In an embodiment, the recycled polymer composition is substantively a postconsumer recycled (PCR) polymer.

[0020] In an embodiment, the recycled polymer composition may include a polypropylene(PP)-based polymer, e.g., in an amount of at least 80 wt%, preferably at least 85 wt%, more preferably at least 90 wt%, more preferably at least 95 wt%, with respect to the recycled polymer composition. 24POLY0036-WO-ORD 4

[0021] In an embodiment, the virgin polymer composition may include a polypropylene- based polymer and optionally a filler, an impact modifier and / or additives. Examples of the filler include (but not limited to) talc, long glass fibers, short glass fibers, mica and CaCO3.

[0022] In an embodiment, the recycled polymer composition and the virgin polymer composition can be differentiated by a different ash content. For example, the recycled polymer composition may have an ash content, measured according to ISO 3451- 1 :2019 at 550°C, of 200-6000 ppm, preferably 300-1 OOOppm, more preferably 400- 600ppm. The virgin polymer composition may have an ash content, measured according to ISO 3451-1 :2019 at 550°C, of 100ppm or less, preferably 0-50ppm, more preferably at most 1ppm.

[0023] In the context of the present invention, a polypropylene(PP)-based polymer, whether a recycled one or a virgin one, is preferably a crystalline polypropylene, like a propylene homopolymer, a random copolymer, or a so-called heterophasic copolymer of propylene and ethylene and / or another alpha-olefin, as is well known in the field.

[0024] In an embodiment, the material properties in the training data used for training the ML model may include at least one of the following properties of recycled polymer compositions: melt flow index; intrinsic viscosity; ash content; ethylene content; impact strength; modulus and shrinkage.

[0025] In an embodiment, the material properties in the training data used for training the ML model may include at least one of the following properties of virgin polymer compositions: melt flow index; intrinsic viscosity; ethylene content; impact strength; modulus and shrinkage.

[0026] In an embodiment, the material properties in the training data used for training the ML model may include at least one of the following properties of mixture extrudates: melt flow index; intrinsic viscosity; ash content; ethylene content; impact strength; modulus and shrinkage.

[0027] In an embodiment, the material properties of the recycled polymer composition measured in step iii) may include at least one of the following properties of the recycled polymer composition: melt flow index; intrinsic viscosity; ash content; ethylene content; impact strength; modulus and shrinkage.

[0028] In an embodiment, the method may further include iiib) measuring material properties of the mixture extrudate to obtain an input set of material properties of the mixture extrudate. The method may further include ivb) receiving, by the trained ML model, the input set of material properties of the mixture extrudate obtained by step iiib). 24POLY0036-WO-ORD 5

[0029] This allows the machine learning model to give a more accurate recommendation. The additional measurement of material properties in step iiib) may be performed in-line or off-line.

[0030] In an embodiment, the material properties of the mixture extrudate measured in step iiib) may include at least one of the following properties of the mixture extrudate: melt flow index; intrinsic viscosity; ash content; ethylene content; impact strength, modulus and shrinkage.

[0031] According to an aspect of the present disclosure, a process for mixing a recycled polymer composition and a virgin polymer composition in a first extruder is presented. The process may include I) feeding the recycled polymer composition and the virgin polymer composition to the first extruder to obtain a mixture extrudate. The virgin polymer composition may be based on the recommendation of the virgin polymer composition as generated by the ML model of the method having one or more of the above-described features.

[0032] Since the virgin polymer composition may be selected according to the recommendation by the ML model, the resulting mixture extrudate is more likely to have desired properties even though there may be variations in material properties of the recycled polymer composition. The in-line measurement of material properties of the recycled polymer composition allows feeding a virgin polymer composition better suited for obtaining mixture extrudate having the desired material properties. Advantageously, adjustments can be made to the type and / or amount of the virgin polymer composition quickly in response to the fluctuations in the properties of the recycled polymer composition, reducing the amount of mixture extrudate not having the desired material properties.

[0033] In an embodiment, the process may further include II) mixing the mixture extrudate and a further virgin composition in a second extruder to obtain a further mixture extrudate.

[0034] In an embodiment, melt of the mixture extrudate may be obtained by step I). The melt of the mixture extrudate obtained by step I) may be mixed with the further virgin composition in step II) without an intermediate solidification step.

[0035] The mixture extrudate obtained in step I) may be used in step II) as a masterbatch, hereinafter also referred to as RECYCLE ingredient, for obtaining the further mixture extrudate as the final product. Such two-steps process for obtaining the final product allows making any necessary adjustments in the feed easier. Adjustments in the virgin polymer composition for step I) may be made without adjusting feeds and extrusion conditions for step II). This reduces process complexity significantly and increases the design freedom of the final product. 24POLY0036-WO-ORD 6

[0036] Preferably, the mixture extrudate is obtained by step I) as a melt. In step II), said melt may be mixed without an intermediate solidifying step with the further virgin composition to obtain the further mixture extrudate. This advantageously requires lower energy consumption.

[0037] In another embodiment, the mixture extrudate obtained by step I) may be solidified and used for further treatments and processing steps at the same or different time and / or location of step I). The solidified mixture extrudate may for example be cut into pellets, tested for some of physical properties, and reserved for or transported to a step II) processing. This advantageously produces flexibility and a standardized way of using the recycled polymer composition in a ‘standard’ compounding process.

[0038] According to an aspect of the present disclosure, a compounding system is presented. The compounding system may include a hardware part. The compounding system may further include a virtual experimentation part. The hardware part may include one or more extruders to to obtain a mixture extrudate. The virtual experimentation part may include a trained ML model. The ML model may have been trained with training data comprising one or more of material properties of recycled polymer compositions, material properties of virgin polymer compositions and material properties of mixture extrudates obtained by mixing the recycled polymer compositions and the virgin polymer compositions. The ML model may be arranged to receive, during extrusion, material properties of the recycled polymer composition in the extruder. The ML model may be arranged to generate, in response to processing said input properties, a recommendation of a virgin polymer composition to be fed to the extruder to obtain the mixture extrudate having desired material properties.

[0039] In an embodiment, the compounding system may be arranged to perform a process of mixing a recycled polymer composition and a virgin polymer composition, with the process having one or more of the above described features. The compounding system may comprise a first extruder and a second extruder that are arranged to continuously perform steps I) and II) of the process. The first extruder may include ai) a first elongated cylindrical tube. The first elongated cylindrical tube may have an end portion having a first inlet port configured to receive, in operation, the recycled polymer composition. The first elongated cylindrical tube may further have an end portion having a first outlet port. The first elongated cylindrical tube may further have a first side inlet port between the first inlet port and the first outlet port configured to, in operation, receive the virgin polymer composition. The first elongated cylindrical tube may further have a first side outlet port between the first inlet port and the side inlet port configured to discharge, in operation, a sample of the recycled polymer 24POLY0036-WO-ORD 7 composition. The first elongated cylindrical tube may further have a measuring means for measuring material properties of the sample of the recycled polymer composition from the first side outlet port. The first outlet port may be configured to discharge, in operation, a melt of the mixture extrudate. The first extruder may further include aii) a first screw arranged in the first elongated cylindrical tube configured to, in operation, convey the recycled polymer composition and the virgin polymer composition to the first outlet port. The second extruder may include bi) a second elongated cylindrical tube. The second elongated cylindrical tube may have an end portion having a second inlet port configured to receive, in operation, the melt of the mixture extrudate from the first outlet port. The second elongated cylindrical tube may further have an end portion having a second outlet port. The second elongated cylindrical tube may further have a second side inlet port between the second inlet port and the second outlet port configured to, in operation, receive a further virgin composition. The second outlet port may be configured to discharge, in operation, a melt of the further mixture extrudate. The second extruder may further include bii) a second screw arranged in the second elongated cylindrical tube configured to, in operation, convey the melt of the mixture extrudate and the further virgin composition to the second outlet. The first screw and the second screw may be operable at different screw speeds. Step I) may be performed in the first extruder. Step II) may be performed in the second extruder. The material properties measured by the measurement means may be used as input properties to the ML model.

[0040] In an embodiment, the compounding system may comprise a third extruder, which has a configuration similar to the first and the second extruder, to receive the melt of the further mixture extrudate, mix with a third virgin composition, and discharge, in operation, a melt of the third mixture extrudate.

[0041] In an embodiment, the further virgin composition used in step II) may include a polypropylene-based polymer. The further virgin composition used in step II) may include a filler, an impact modifier and / or additives.

[0042] For the sake of clarity, the properties of the recycled polymer compositions or the virgin polymer compositions are normally measured by the methods described below. Melt flow index (MFI) is measured according to ISO 1133 (2005) (2.16 kg, 230°C). The unit of MFI is g / 10 min. It should be noted that the Melt Flow Index (MFI) and the Melt Flow Rate (MFR) may be used interchangeably.

[0043] Impact strength may refer to Charpy impact strength or Izod impact strength, both are well known in the field. As an example, Izod impact strength is measured according to ISO 180-1 A (2000). The unit of the Izod impact strength is kJ / m2. 24POLY0036-WO-ORD 8

[0044] Flexural modulus is measured according to ISO 178 (2010), the unit is N / mm2 or MPa. Tensile modulus is measured according to ISO 527 2(1A) (2012), the unit is N / mm2 or MPa.

[0045] Intrinsic viscosity of the matrix phase of a heterophasic polypropylene and of the dispersed phase of a heterophasic polypropylene are determined according to ISO- 1628-1 (2009) and ISO-1628-3 (2010) based on the amounts of xylene-insoluble matter (CXI) and xylene-soluble matter (CXS) measured according to ISO 16152 (2005).

[0046] Shrinkage is measured according to ISO 294-4 (2001). The unit of shrinkage is percentage (%).

[0047] Ash content is measured according to ISO 3451-1 :2019 at 550°C, the unit is %.

[0048] Brief description of the Drawings

[0049] Embodiments of the present disclosure will now be described, by way of example only, with reference to the accompanying schematic drawings in which corresponding reference symbol indicate corresponding parts, in which:

[0050] Fig. 1 is an abstract representation of an example embodiment of a compounding system 100 of the present disclosure;

[0051] Fig. 2 is an example compounding process using two extruders in series;

[0052] Fig. 3 illustrates an example of a virtual experimentation part including a machine learning model for use in a compounding system according to the present disclosure;

[0053] Fig. 4 is an example of a compounding extruder used in an embodiment of the present disclosure; and

[0054] Fig. 5 shows an example embodiment of a computing system for implementing certain aspects of the present technology.

[0055] The figures are intended for illustrative purposes only, and do not serve as restriction of the scope of the protection as laid down by the claims.

[0056] Detailed description

[0057] It will be readily understood that the components of the embodiments as generally described herein and illustrated in the appended figures could be arranged and designed in a wide variety of different configurations. Thus, the following more detailed description of various embodiments, as represented in the figures, is not intended to limit the scope of the present disclosure but is merely representative of 24POLY0036-WO-ORD 9 various embodiments. While the various aspects of the embodiments are presented in drawings, the drawings are not necessarily drawn to scale unless specifically indicated.

[0058] The described embodiments are to be considered in all respects only as illustrative and not restrictive. The scope of the present disclosure is, therefore, indicated by the appended claims rather than by this detailed description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.

[0059] Reference throughout this specification to features, advantages, or similar language does not imply that all of the features and advantages that may be realized with the present disclosure should be or are in any single example of the present disclosure. Rather, language referring to the features and advantages is understood to mean that a specific feature, advantage, or characteristic described in connection with an embodiment is included in at least one embodiment of the present disclosure. Thus, discussions of the features and advantages, and similar language, throughout this specification may, but do not necessarily, refer to the same example.

[0060] Furthermore, the described features, advantages, and characteristics of the present disclosure may be combined in any suitable manner in one or more embodiments. One skilled in the relevant art will recognize, in light of the description herein, that the present disclosure may be practiced without one or more of the specific features or advantages of a particular embodiment. In other instances, additional features and advantages may be recognized in certain embodiments that may not be present in all embodiments of the present disclosure. Reference throughout this specification to "one embodiment," "an embodiment," or similar language means that a particular feature, structure, or characteristic described in connection with the indicated embodiment is included in at least one embodiment of the present disclosure. Thus, the phrases "in one embodiment," "in an embodiment," and similar language throughout this specification may, but do not necessarily, all refer to the same embodiment.

[0061] The solution of the present disclosure may be applied to compounding technology for producing compounds, e.g., to customer specifications, preferably based on Post-Consumer Recycled plastics (PCR) of broad composition and performance. A compounding system of the present disclosure may include a hardware part and a virtual experimentation part, which operate together to achieve optimal compounding.

[0062] Feedstock to the compounding process may include PCR with broad quality variations (e.g., in melt flow index, ash content, color, impact strength, modulus, shrinkage), impurities as well as different physical shapes (e.g., granules, flakes). In 24POLY0036-WO-ORD 10 addition, the quality of feedstock may be unknown and potentially varies even in one batch (e.g., mixed in silo). The solution of the present disclosure improves the compounding processing of such quality fluctuating feedstock.

[0063] The hardware part of the compounding technology of the present disclosure may include a tandem-processing unit consisting of two extruders in series, typically including process mass customization control units (FMCC). In a first extruder, the PCR with high variation in composition and specs may be extruded into a masterbatch (hereinafter also referred to as RECYCLE ingredient) with a known, predefined, (more) constant performance. In a second extruder, ‘standard’ compounding may be executed in which the masterbatch RECYCLE ingredient may be used as a ‘standard’ ingredient in a compounding recipe, yielding a compound with desired customer specified performance. Both extruders may be connected to each other to allow for lower energy consumption and costs, following the FMCC concept.

[0064] The virtual experimentation part of the compounding technology may include a pre-trained machine-learning model used to select the amount and possibly type of virgin ingredients to be dosed at the hardware part to produce a RECYCLE ingredient with desired specification.

[0065] In following example embodiments, the solution of the present disclosure will be described in more details. In the example embodiments, polypropylene (PP) compounds are described, but is to be understood that the solution of the present disclosure may be used for other compounded plastics products as well, such as (a mix of) base-resin(s) (e.g. polyethylene, engineering thermoplastics products).

[0066] Fig. 1 is an abstract representation of an example embodiment of a compounding system 100 of the present disclosure. The compounding system 100 may include one or more extruders, possibly working in series, to perform a compounding process in one compounding process or in multiple sub-compounding processes.

[0067] In the example of Fig. 1 , two extruders operate in series. A hardware part 110 is shown including a first extruder 112 that is coupled with a second extruder 114 to perform the compounding process.

[0068] Also shown is a virtual experimentation part 120, which may be coupled with one or more of the extruders of the compounding system 100 for ML optimization of the compounding process.

[0069] In the example of Fig. 1 , the virtual experimentation part is coupled with the first extruder 112 for ML optimization of the compounding process in the first extruder 112 for producing the RECYCLE ingredient.

[0070] An example compounding process 200 as may be performed by the hardware part 110 is shown in Fig. 2. A first extruder process is performed in the first extruder 24POLY0036-WO-ORD 11

[0071] 112, where the RECYCLE ingredient is created to give a quality boost based on spread reduction and known performance. A second extruder process is performed in the second extruder 114, where the final product 5 is created based on ‘standard’ extrusion. The following input feeds to the process 200 are shown: feed 1 into the first extruder 112; optional feed 2 into the first extruder 112; feed 3a into the second extruder 114; feed 3b into the second extruder 114 and including the RECYCLE ingredient obtained from the first extruder 112; and feed 4 into the second extruder 114.

[0072] The first extruder 112 may compound a masterbatch (i.e., RECYCLE ingredient 3b) with sufficiently narrow and defined performance specifications based on PCR product (feed 1) with varying (upfront unknown) composition / specifications and optionally performance boosting ingredients (feed 2), like virgin PP, elastomers, mineral filler. Feed 1 is fed into the first extruder 112. This stream (i.e., feed 1) may be obtained from external sources and can have a large spread in composition and properties. Feed 2 may include performance boosting ingredients that are optionally added to boost performance and compensate for the variability of the feed 1 products, like virgin PP, elastomer, additives, color masterbatch. Feed 3b is the output of the first extruder 112 and typically includes masterbatch RECYCLE ingredient, i.e., upgraded recycle PP with narrow defined specifications.

[0073] The second extruder 114 may compound final products 5, e.g., according to high demanding customer speciation, similar in approach as for ‘standard’, e.g., Polypropylene Copolymer (PPc), extrusion. Herein, the masterbatch 3b serves as an ingredient in the recipe. Feed 3b includes the output of the first extruder 112 including the masterbatch RECYCLE ingredient. Feed 3a may include added ingredients, e.g., to meet customer specifications, like virgin PP, elastomers, additives, color master batch. Feed 4 may include fillers like talcum, mica, fibers (e.g. short glass, long glass, carbon, natural, ....), and etcetera that may be added to the second extruder process. Output 5 is a final product, e.g., meeting the customers specification.

[0074] The compounding of the RECYCLE ingredient 3b by the first extruder 112 prior to the compounding of the final product 5 by the second extruder 114 greatly improves the quality of the final product 5. To even further improve the compounding, virtual experimentation may be added to the first extrusion process in the first extruder 112. This virtual experimentation may be performed by the virtual experimentation part 120 as shown in Fig. 1. Such virtual experimentation may also be applied to a compounding system including just one extruder, which extrusion process may be improved by virtual experimentation in a similar manner. 24POLY0036-WO-ORD 12

[0075] In the present disclosure, virtual experimentation is a computer implemented method involving ML to optimize a compounding process. For example, in the compounding system 100, a pre-trained ML model may be used to select an amount and / or type of virgin ingredients to be dosed at feed 2 to produce a RECYCLE ingredient with desired and / or known specification. The ML model may be obtained using material informatics tools such as show in Fig. 3, which will be further described below. To be able to anticipate batch-to- batch variations in PCR, an in-line measurement of PCR quality may be applied. This allows for real time adjustments of the feed 2 composition to further drive down spread in performance of the compound product, such as the RECYCLE ingredient 3b.

[0076] To train the ML model, training data may be generated based on the key identifiers of the PCR, related to the ‘correcting’ virgin ingredients. This data may be split in a set of training data and validation data, preferably in a ratio of 75:25. The training data may be used to train the ML model using, e.g., a folding scheme (k-fold cross validation). The validation data (or unseen data) is not used for training and may serve to validate the ML model.

[0077] To train the model off-line measured performance data from the compound product, such as the RECYCLE ingredient 3b, may be used, e.g., via design of experiments. Having PCR supplier certificate of analysis available may help in improving model performance as well. The exact features (i.e., measured quantities) which have to be used for obtaining optimal model performance may be obtained by investigating feature importance.

[0078] Once the ML model is trained, the in-line measured properties (e.g., PCR quality, processing data) together with desired compound ingredient performance, e.g., the performance of RECYCLE ingredient 3b, may be used to tune the composition of feed 2.

[0079] A more detailed example of a virtual experimentation part including such ML model will be described with Fig. 3. In the example of Fig. 3, material informatics tools are used with the virtual experimentation part 120.

[0080] Fig. 3 illustrates an exemplary process flow 300 for virtual experimentation. An exemplary compounding system may include one or more extruders. In the example of Fig. 3, a first extruder 112 of compounding system 100 is shown. A feed 1 may enter the one or more extruders 112 at a first inlet, and may be, for example, a PP recycle feed stream. The feed 1 may be provided by one or more suppliers 302 and may be associated with various feed data 304 including, but not limited to, PCR materials’ quality data. The PCR quality data may include, but is not limited to, a certificate of analysis, containing information like (but not limited to) melt flow index, ash content, 24POLY0036-WO-ORD 13 modulus (stiffness), and / or variability characteristics. In certain embodiments, the PCR materials may have high variability from batch to batch, due to different origins of the batches at the same time of the year, or the different times of the year from the same location, or both with differences in time and location.

[0081] A feed 2 composition may be fed inline into the extruder 112. Feed 2 may include, for example, virgin polypropylene, additives, fillers, and / or elastomers. Inline processes 312 may provide various inline PCR quality data 314 including, but not limited to, viscosity (such as for the purposes of measuring for manufacturers), ethylene content, and / or ash content. Other inline processes 322 may provide various processing data 324 including, but not limited to, barrel torque, melt pressure, temperature (in some embodiments based on zones within the extruder 112), feed 1 quantities, and / or feed 2 quantities. In various embodiments, additional feeds may be added as needed to the extruder 112.

[0082] A final product, in this example RECYCLE ingredient 3b, may exit the extruder 112. Offline processes 332 may provide offline final product quality data 334 including, but not limited to, ash content, modulus (stiffness), impact strength, shrinkage, melt flow rate, and / or quality / assurance testing.

[0083] A pre-trained ML model 350 may be used. The ML model 350 may be used to determine what to add to the extruder 112 as a feed 2 composition based on the unknown and / or variable qualities of feed 1 compositions. In certain embodiments, the model 350 may provide input on the amounts and / or types of ingredients to add as a feed 2 composition to produce a final product, such as RECYCLE ingredient 3b, with desired and / or known specifications from the extruder 112. The model may be obtained using known material informatics tools. To be able to anticipate batch-to- batch variations in PCR materials, however, an in-line measurement of PCR quality may be applied. This may allow for real time adjustments of feed 2 compositions to ensure meeting desired performance of a final product of the extruder 112, such as the RECYCLE ingredient 3b.

[0084] To train the ML model 350, training data / input 340 may be generated based on key identifiers of the PCR in feed 1 .

[0085] This training data may be split into a set of training data and validation data. The ratio of training data may vary from approximately 99:1 to approximately 1 :99, but preferably is in a ratio of approximately 75:25. The training data may be used to train the ML model using, e.g., a folding scheme (k-fold cross validation) or similar process. The validation data (or unseen data) may be not used for training and may be used to validate the ML model. 24POLY0036-WO-ORD 14

[0086] To train the model 350 it may be necessary to have off-line measured performance data 344 from the final product of the extruder 112, e.g., the RECYCLE ingredient 3b, available (e.g., via design of experiments). Utilizing PCR supplier certificates of analysis may improve model performance. The exact features (i.e., measured quantities) that may be used for obtaining optimal model performance may be obtained by investigating feature importance and may depend on the measurement capabilities of the exact equipment used.

[0087] Once the ML model 350 is trained, the in-line measured properties including, but not limited to inline PCR quality 314 and inline processing data 324, together with desired final product performance, e.g., the performance of the RECYCLE ingredient 3b, may be used to tune the composition of feed 2.

[0088] Fig. 4 illustrates an example of a compounding extruder system 400 used in an example embodiment of the present disclosure. The system 400 may include a first extruder 410 and a second extruder 450.

[0089] The first extruder 410 may include a first elongated cylindrical tube and a first screw (not shown) arranged in the first elongated cylindrical tube.

[0090] The first elongated cylindrical tube may have an end portion having a first inlet port 412 and an end portion having a first outlet port 414. The first inlet port 412 may be configured to receive, when in operation, a recycled polymer composition.

[0091] Downstream of the first inlet port 412, the first elongated cylindrical tube may have a side outlet port 418 configured to discharge, when in operation, a sample of the recycled polymer composition. The sample may be received by a measuring means 420 for measuring material properties of the sample from the side outlet port 418.

[0092] Downstream of the side outlet port 418, the first elongated cylindrical tube may further have a side inlet port 422 configured to, when in operation, receive a virgin polymer composition. The first outlet port 414 may be configured to discharge, when in operation, a first melt composition comprising the recycled polymer composition and the virgin polymer composition.

[0093] In this example embodiment, the first elongated cylindrical tube may further be provided with a vacuum degassing section 416 between the side inlet port 422 and the first outlet port 414.

[0094] The first melt composition discharged from the first outlet port 414 may be transferred to the second extruder 450 by a heated transition piece.

[0095] The second extruder 450 may include a second elongated cylindrical tube and a second screw (not shown) arranged in the second elongated cylindrical tube.

[0096] The second elongated cylindrical tube may have an end portion having a second inlet port 452 and an end portion having a second outlet port 454. The second inlet port 24POLY0036-WO-ORD 15

[0097] 452 may be configured to receive, when in operation, the first melt composition from the first outlet port 414. The second elongated cylindrical tube may further have an inlet port 456 before the second inlet port 452, which may be configured to receive, when in operation, additives, for example color masterbatch.

[0098] The second elongated cylindrical tube may further have side inlet ports, e.g., three side inlet ports 458, 460, 462, between the second inlet port 452 and the second outlet port 454. For example, the inlet ports 458 and 460 may be configured to receive, when in operation, fillers such as talc and a flame retardant, and the inlet port 462 may be configured to receive, when in operation, glass fibers.

[0099] In this example embodiment, the second elongated cylindrical tube may further be provided with a vacuum degassing section 468 between the side inlet port 462 and the second outlet port 454. The second outlet port 454 may be configured to discharge, when in operation, a second melt composition comprising components added from the inlet ports 456, 452 , 458, 460 and 462.

[0100] The second melt composition extruded from the second outlet port 454 may be solidified and cut into pellets.

[0101] The composition of the material added through any one of the inlet ports 422, 456, 458, 460, 462 may be optimized by a virtual experimentation part, such as the virtual experimentation part 120, implementing a ML model, e.g., utilizing an information flow such as the process flow 300. Feed 2 shown in, e.g., Fig. 3, is an example of an inlet port 422, 456, 458, 460, 462.

[0102] Fig. 5 shows an example embodiment of a computing system 500 for implementing certain aspects of the present technology. In various examples, the computing system 500 may be any computing device making up the virtual experimentation part 120, any part of the system architecture of Fig. 1 , or any other computing system described herein.

[0103] In some implementations, a computing system 500 may implement the methods described herein, such as the method of the process flow 300 of the present disclosure.

[0104] The computing system 500 may include any component of a computing system described herein, which components may be in communication with each other using connection 505. The connection 505 may be a physical connection via a bus, or a direct connection into processor 510, such as in a chipset architecture. The connection 505 may also be a virtual connection, networked connection, or logical connection.

[0105] In some implementations, the computing system 500 may be a distributed system in which the functions described in this disclosure may be distributed within a datacenter, multiple datacenters, a peer network, etc. In some embodiments, one or 24POLY0036-WO-ORD 16 more of the described system components represents many such components each performing some or all of the functions for which the component is described. In some embodiments, the components may be physical or virtual devices.

[0106] The example system 500 includes at least one processing unit (CPU or processor) 510 and a connection 505 that couples various system components including system memory 515, such as read-only memory (ROM) 520 and randomaccess memory (RAM) 525 to processor 510. The computing system 500 may include a cache of high-speed memory 512 connected directly with, in close proximity to, or integrated as part of the processor 510.

[0107] The processor 510 may include any general-purpose processor and a hardware service or software service, such as services 532, 534, and 536 stored in storage device 530, configured to control the processor 510 as well as a special-purpose processor where software instructions are incorporated into the actual processor design. The processor 510 may essentially be a completely self-contained computing system, containing multiple cores or processors, a bus, memory controller, cache, etc. A multi-core processor may be symmetric or asymmetric.

[0108] To enable user interaction, the computing system 500 may include an input device 545, which may represent any number of input mechanisms, such as a microphone for speech, a touch-sensitive screen for gesture or graphical input, keyboard, mouse, motion input, speech, etc. The computing system 500 may also include an output device 535, which may be one or more of a number of output mechanisms known to those of skill in the art. In some instances, multimodal systems may enable a user to provide multiple types of input / output to communicate with the computing system 500. The computing system 500 may include a communications interface 540, which may generally govern and manage the user input and system output. There is no restriction on operating on any particular hardware arrangement, and therefore the basic features here may easily be substituted for improved hardware or firmware arrangements as they are developed.

[0109] A storage device 530 may be a non-volatile memory device and may be a hard disk or other types of computer readable media which may store data that are accessible by a computer, such as magnetic cassettes, flash memory cards, solid state memory devices, digital versatile disks, cartridges, random access memories (RAMs), read-only memory (ROM), and / or some combination of these devices.

[0110] The storage device 530 may include software services, servers, services, etc., that, when the code that defines such software is executed by the processor 510, causes the system to perform a function. In some embodiments, a hardware service that performs a particular function may include a software component stored in a 24POLY0036-WO-ORD 17 computer-readable medium in connection with the necessary hardware components, such as processor 510, connection 505, output device 535, etc., to carry out the function.

[0111] Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure and the appended claims. In the claims, the word “comprising” does not exclude other elements or steps, and the indefinite article “a” or “an” does not exclude a plurality. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope thereof.

Claims

24POLY0036-WO-ORD 18CLAIMS1 . A method (300) of determining a feed (2) to an extruder (112) to obtain a mixture extrudate (3b) having desired material properties, the method comprising: i) providing a trained machine learning, ML, model (350) trained with training data comprising one or more of material properties (304) of recycled polymer compositions, material properties of virgin polymer compositions and material properties (334) of mixture extrudates obtained by mixing the recycled polymer compositions and the virgin polymer compositions; ii) feeding (1 , 2) a recycled polymer composition and a virgin polymer composition to the extruder (112); iii) measuring (312, 322), during extrusion, material properties (314, 324) of the recycled polymer composition in the extruder (112) to obtain an input set (342) of material properties of the recycled polymer composition; iv) receiving, by the trained ML model (350), the input set (342) of material properties of the recycled polymer composition obtained by step iii) and an input set (340, 344) of the desired material properties of the mixture extrudate (3b); v) processing the input sets using the trained ML model (350); and vi) generating, by the trained ML model (350) in response to said processing of the input sets, a recommendation of the virgin polymer composition to be fed (2) to the extruder (112) to obtain the mixture extrudate (3b) having the desired material properties.

2. The method according to claim 1 , wherein the recycled polymer composition comprises a polypropylene, PP, -based polymer, preferably in an amount of at least 80 wt%, preferably at least 85 wt%, more preferably at least 90 wt%, more preferably at least 95 wt%, with respect to the recycled polymer composition.

3. The method according to any one of preceding claims, wherein the virgin polymer composition comprises a polypropylene-based polymer and optionally a filler, an impact modifier and / or additives.

4. The method according to any one of the preceding claims, wherein the material properties in the training data used for training the ML model comprise: at least one of melt flow index, intrinsic viscosity, ash content, ethylene content, impact strength, modulus and shrinkage of recycled polymer compositions; and / or24POLY0036-WO-ORD 19 at least one of melt flow index, intrinsic viscosity, ethylene content, impact strength, modulus and shrinkage of virgin polymer compositions; and / or at least one of melt flow index, intrinsic viscosity, ash content, ethylene content, impact strength, modulus and shrinkage of mixture extrudates.

5. The method according to any one of the preceding claims, wherein the material properties of the recycled polymer composition measured in step iii) comprise: at least one of melt flow index, intrinsic viscosity, ash content, ethylene content, impact strength, modulus and shrinkage of the recycled polymer composition.

6. The method according to claim any one of the preceding claims, wherein the method further comprises: iiib) measuring (332) material properties (334) of the mixture extrudate (3b) to obtain an input set (344) of material properties of the mixture extrudate (3b); and ivb) receiving, by the trained ML model, the input set (344) of material properties of the mixture extrudate obtained by step iiib).

7. The method according to claim 6, wherein the material properties of the mixture extrudate measured in step iiib) comprise: at least one of melt flow index, intrinsic viscosity, ash content, ethylene content, impact strength, modulus and shrinkage of the mixture extrudate.

8. A process for mixing a recycled polymer composition and a virgin polymer composition in a first extruder (112, 410), the process comprising:I) feeding (1 , 2) the recycled polymer composition and the virgin polymer composition to the first extruder (112, 410) to obtain a mixture extrudate (3b), wherein the virgin polymer composition is based on the recommendation of the virgin polymer composition as generated by the ML model of the method according to any one of the claims 1-7.

9. The process according to claim 8, further comprising:II) mixing the mixture extrudate (3b) and a further virgin composition (4) in a second extruder (114, 450) to obtain a further mixture extrudate (5).

10. The process according to claim 9, wherein a melt of the mixture extrudate is obtained by step I), and wherein the melt of the mixture extrudate obtained by step I) is24POLY0036-WO-ORD 20 mixed with the further virgin composition in step II) without an intermediate solidification step.

11. A compounding system (100) comprising a hardware part (110) and a virtual experimentation part (120), wherein the hardware part (110) comprises one or more extruders (112, 114) to to obtain a mixture extrudate (3b, 454), wherein the virtual experimentation part comprises a trained machine learning, ML, model (350), wherein the ML model has been trained with training data comprising one or more of material properties (304) of recycled polymer compositions, material properties of virgin polymer compositions and material properties (334) of mixture extrudates obtained by mixing the recycled polymer compositions and the virgin polymer compositions, wherein the ML model is arranged to receive, during extrusion, material properties (314, 324) of the recycled polymer composition in the extruder (112), wherein the ML model is arranged to generate, in response to processing said input properties (314, 324), a recommendation of a virgin polymer composition to be fed (2) to the extruder (112) to obtain the mixture extrudate (3b, 454) having desired material properties.

12. The compounding system (100, 400) according to claim 11 , arranged to perform the process according claim 9 or claim 10, wherein the compounding system (100, 400) comprises: a) a first extruder (112, 410) comprising: ai) a first elongated cylindrical tube having- an end portion having a first inlet port configured to receive, in operation, the recycled polymer composition,- an end portion having a first outlet port,- a first side inlet port between the first inlet port and the first outlet port configured to, in operation, receive the virgin polymer composition,- a first side outlet port between the first inlet port and the side inlet port configured to discharge, in operation, a sample of the recycled polymer composition, and- a measuring means (420) for measuring material properties of the sample of the recycled polymer composition from the first side outlet port,24POLY0036-WO-ORD 21 wherein the first outlet port is configured to discharge, in operation, a melt of the mixture extrudate, and aii) a first screw arranged in the first elongated cylindrical tube configured to, in operation, convey the recycled polymer composition and the virgin polymer composition to the first outlet port; and b) a second extruder (114, 450) comprising: bi) a second elongated cylindrical tube having- an end portion having a second inlet port configured to receive, in operation, the melt of the mixture extrudate from the first outlet port,- an end portion having a second outlet port, and- a second side inlet port between the second inlet port and the second outlet port configured to, in operation, receive a further virgin composition, wherein the second outlet port is configured to discharge, in operation, a melt of the further mixture extrudate, and bii) a second screw arranged in the second elongated cylindrical tube configured to, in operation, convey the melt of the mixture extrudate and the further virgin composition to the second outlet, wherein the first screw and the second screw are operable at different screw speeds, wherein step I) is performed in the first extruder (410) and step II) is performed in the second extruder (450), and wherein the material properties measured by the measurement means (420) are used as input properties (314, 324) to the ML model (350).

13. The compounding system (100) according to claim 11 or claim 12, wherein the further virgin composition used in step II) comprises a polypropylene-based polymer and optionally a filler, an impact modifier and / or additives.

Images