Characterization of recycled textile fiber material

Abstract

A method for characterizing a sample of recycled textile fiber material that includes at least one of fibers, neps, and hard-ends is proposed The sample of recycled textile fiber material is provided. The sample is individualized (21) to produce individualized entities. The individualized entities are transported (22) in an air flow. Measured parameters of the entities are determined (23) while they are transported in the air flow. The measured parameters include time-of-flight, area-under-curve, peak, average signal height, and speed, and at least one of median signal height and sum of squared signal height. The entities are classified (25) as at least one of fibers, neps, and hard-ends based at least in part on comparing (24) their measured parameters with thresholds. The sample is characterized (26) using at least one result of the classification. (Fig. 2)

Description

CHARACTERIZATION OF RECYCLED TEXTILE FIBER MATERIALTECHNICAL FIELD

[0001] The disclosure lies in the field of textile quality assessment, and in particular in the field of measurements of quality parameters of recycled textile materials. It relates to a method and an apparatus for characterizing a sample of recycled textile fiber material, according to the preambles of the independent claims.BACKGROUND ART

[0002] In recent decades, there has been ever-increasing attention on sustainability, with focus on many diverse areas such as social, environmental, and energy. One of the industries that has been addressing and contributing to this important cause is the textile industry. For example, natural fibers require large amounts of land, water, and potentially harmful pesticides. Manmade fibers have their own challenges, as they are manufactured chemically. Furthermore, with the popularity of fast fashion in recent decades, clothing fashions go in and out of style faster than ever, resulting in about 60% more garment consumption than fifteen years ago. Overall, 90+ million tons of textile waste is produced each year globally, and fast fashion alone is responsible for about 10% of all global carbon emissions. The textile industry has been focusing on recycling as one of the major contributions to sustainability in recent times.

[0003] Textile recycling may be the world's oldest recycling industry, with records of a rag collector association formed by Roman slaves in 50 AD. In general, textile recycling intends to recover fiber, yarn, or fabric from the textile material and to create new and useful products. Textile recycling processes are typically based on either mechanical or chemical methods. The mechanical recycling processes involve physically deconstructing fabrics through shredding, crushing, or melting of a garment into fibers, and then reprocessing them back into yarn. In general, this process can damage the original fibers and create a lower quality textile with limited applications. Therefore, it is desirable to enhance the recycling process to not only achieve the deconstruction / shredding (generally referred to as shreddingherein) results, but also to minimize damage to the fibers. This optimization requires accurate measurements of 1) the quality of the recycled fibers, and 2) the quality of the deconstructing / shredding process itself.

[0004] Typical fiber quality parameters include, but are not limited to, fiber length and fiber neps, which are well defined, standardized, and understood in the textile industry. However, the shredding process quality parameters are less known due to the recent emerging recycling methods and machinery. Recycling parameters should indicate how well the shredding process deconstructs the garments and fabrics to the fibers. This includes detecting and quantifying the pieces of yam, fabric, and other non-fibrous materials that are left over in the recycled fibers at the end of recycling process.

[0005] Fig. 1 shows two examples of recycled fibers created through a mechanical recycling process. Both examples A and B show recycled fibers with left-over pieces of yarn.

[0006] As used in this disclosure, the term “hard-ends” refers to any leftover materials in recycled fiber that is not classified as a fiber or a nep. This includes yarns, fabrics, and any non-fibrous materials (referred to as contamination in this document).

[0007] Recycled fiber typically includes fibers, neps, and hard-ends. The fibers, neps, and various elements of the hard-ends are generally referred to as “entities”.

[0008] An instrument manufactured by Uster Technologies, Inc., known as AFIS (see, e.g., the brochure “USTER® AFIS PRO 2 - The fiber process control system”, Uster Technologies AG, 2016, and patents US-4,512,060, US-5,270,787, and US-5,469,253) measures the fiber quality of virgin (not recycled) fiber bundles by separating fibers and neps into one or two airstreams and measuring the fiber length and nep size via optical sensing. For recycled fibers, however, AFIS measurement accuracy is degraded due to the presence of hard-ends.SUMMARY OF THE DISCLOSURE

[0009] It is an object of the present invention to provide a method and an apparatus for characterizing a sample of recycled textile fiber material and, simultaneously, accurately- 2 - P0651PF-WOmeasuring fiber properties. The invention shall enable inline and offline measurements as required in different processes within recycling and spinning mills as well as quality measurement methods and apparatuses elsewhere, including, but not limited to, recycling fiber suppliers, fiber merchants, and traders. It is a further object of the invention to measure the quality of a textile recycling process.

[0010] These and other objects are solved by the method and the apparatus according to the invention, as defined in the independent claims. The dependent claims define preferred embodiments of the invention.

[0011] The method according to the invention is for characterizing a sample of recycled textile fiber material that includes at least one of fibers, neps, and hard-ends. The method comprises the steps ofa. individualizing the sample to produce individualized entities;b. transporting the entities in an air flow;c. determining measured parameters of the entities while they are transported in the air flow;d. the measured parameters including time-of-flight, area-under-curve, peak, average signal height, and speed, and at least one of median signal height and sum of squared signal height;e. classifying the entities as at least one of fibers, neps, and hard-ends based at least in part on comparing their measured parameters with thresholds; andf. characterizing the sample using at least one result of the classification.

[0012] In one embodiment, the measured parameters additionally include acceleration.

[0013] In one embodiment, entity parameters including length, diameter, area, and / or weight of the entities are determined from the measured parameters. Hard-end dimensions may be determined from hard-end time-of-flight, speed, and average signal height. Hard-end weight may be determined from the hard-end dimensions and at least one of an assumed bulk density and an experimentally determined scaling factor.

[0014] In one embodiment, statistical parameters related to at least one of fibers, neps, and hard-ends are determined from the entity parameters, and the sample is characterized using the statistical parameters. For example, a recycling opening index may be determined from- 3 - P0651PF-WOa total weight of classified hard-ends and a total weight of the sample. The statistical parameters may include minimum, maximum, average, standard deviation, and / or histogram distribution.

[0015] The thresholds can be predetermined or variable.

[0016] One embodiment further comprises the step of separating hard-ends from the fibers and neps based on entity weight, after the step of individualizing the sample.

[0017] In one embodiment, the step of classifying the entities includes further classifying the hard-ends as at least one of yarn, entangled yam, woven fabric, knit fabric, and contamination.

[0018] One embodiment further comprises the step of computing from the measured parameters entity parameters comprising at least one of entity length, entity size, entity diameter, entity shape, entity weight, and entity texture.

[0019] In one embodiment, the method is applied to at least one of offline laboratory test equipment and inline process monitoring equipment.

[0020] In one embodiment, the step of measuring parameters of the entities is accomplished by side-by-side split photodiode transmissive sensors providing two time-resolved extinction signals for each entity.

[0021] In one embodiment, the step of measuring parameters of the entities comprises imaging of the entities via a one-dimensional or a two-dimensional image sensor.

[0022] The apparatus according to the invention is for characterizing a sample of recycled textile fiber material that includes at least one of fibers, neps, and hard-ends. The apparatus comprises:a. an individualizer for individualizing the sample to produce individualized entities; b. an aerodynamic transporter for transporting the entities in an air flow;c. at least one sensor for determining measured parameters of the entities while they traverse the aerodynamic transporter;- 4 - P0651PF-WOd. the at least one sensor being configured to determine measured parameters including time-of-flight, area-under-curve, peak, average signal height, and speed, and at least one of median signal height and sum of squared signal height;e. a signal processor configured to classify the entities as at least one of fibers, neps, and hard-ends based at least in part on comparing their measured parameters with thresholds; andf. the signal processor being further configured to characterize the sample using at least one result of the classification.

[0023] In one embodiment, the at least one sensor is additionally configured to determine acceleration as a measured parameter.

[0024] In one embodiment, the signal processor is further configured to determine entity parameters including length, diameter, area, and / or weight of the entities from the measured parameters. The signal processor may be further configured to determine hard-end dimensions from hard-end time-of-flight, speed, and average signal height. The signal processor may be further configured to determine hard-end weight from the hard-end dimensions and at least one of an assumed bulk density and an experimentally determined scaling factor.

[0025] In one embodiment, the signal processor is further configured to determine statistical parameters related to at least one of fibers, neps, and hard-ends form the entity parameters, and to characterize the sample using the statistical parameters. For example, the signal processor may be further configured to determine a recycling opening index from a total weight of classified hard-ends and a total weight of the sample. The signal processor may be configured to compute statistical parameters including minimum, maximum, average, standard deviation, and / or histogram distribution.

[0026] The signal processor may be configured to compute thresholds that are predetermined or variable.

[0027] One embodiment further comprises a separator for separating the hard-ends from the fibers and neps based on entity weight, the separator being arranged after the individualizer.

[0028] In one embodiment, the signal processor is further configured to further classify hard-ends as at least one of yam, entangled yarn, woven fabric, knit fabric, and contamination.- 5 - P0651PF-WO

[0029] In one embodiment, the signal processor is further configured to compute from the measured parameters entity parameters comprising at least one of entity length, entity size, entity diameter, entity shape, entity weight, and entity texture.

[0030] In one embodiment, the at least one sensor comprises at least one of an optical reflective sensor and an optical transmissive sensor.

[0031] In one embodiment, the at least one sensor includes side-by-side split photodiode transmissive sensors that provide two time-resolved extinction signals for each entity.

[0032] In one embodiment, the at least one sensor includes a one-dimensional or a two-dimensional image sensor.

[0033] A textile offline laboratory test equipment or a textile inline process monitoring equipment according to the invention comprises the apparatus as described above.

[0034] The invention improves the current AFIS technology such that it can inspect and analyze recycled fibers by measuring the following quality parameters:1. Fiber Qualities:a. Fiber length: the length of individual recycled fibers.b. Nep size: a small knot or cluster of entangled recycled fibers.2. Recycling Qualities:a. Hard-ends: pieces of un-opened recycled garments or fabrics:i. Yarn: total number, size, size distribution, diameter, and weight estimate of yarn in the recycled fiber samples being tested;ii. Fabric: total number, size, size distribution, type, and weight estimate of fabric in the recycled fiber samples being tested;iii. Contamination: total number, size, size distribution, and weight estimate of any non -fibrous objects in the recycled fiber samples being tested;b. Recycling Opening Index by weight (roiw): an index that signifies how well the recycling process shredded (i.e opened) the fabrics, considering the- 6 - P0651PF-WOweight of the hard-ends and weight of the recycled fibers. This index may include individual or collective assessment of yarn hard-ends, fabric hard- ends, and contamination hard-ends.c. Recycling Opening Index by count (roic): an index that signifies how well the recycling process shredded (i.e opened) the fabrics, considering the weight of the hard-ends per number of recycled fibers. This index may include individual or collective assessment of yarn hard-ends, fabric hard- ends, and contamination hard-ends.

[0035] The present invention improves the current AFIS technology to measure the recycling quality and fiber quality of recycled textile fiber material. This is accomplished by a further development of the current AFIS technology. While AFIS only measures the individual fiber length and nep size, this invention also measures the hard-ends type and size using optical sensors coupled with intelligent classification and measurement algorithms. The main differences to AFIS include:1. Ability to separate individual fibers, neps, and hard-ends for a recycled textile fiber material.2. Use of an optimized two-phase flow separation scheme to route fibers, neps, and hard-ends into one of two air streams based on their weight.3. Differentiation between individual fibers, neps, and hard-ends through classification algorithms.4. Improvement of fiber length and nep measurement accuracy by reducing interferences due to the presence of hard-ends in the recycled fiber bundle influencing the fiber length and nep sensing and measurement.5. Provision of an optional optical reflection method for all of the measurements.- 7 - P0651PF-WOBRIEF DESCRIPTION OF THE DRAWINGS

[0036] In the following, the invention is explained in detail based on the attached drawings.

[0037] Fig. 1 shows photographs of two samples of recycled fibers with hard-ends present. (A) Mixed color recycled fibers with color hard-ends; (B) White recycled fibers with white hard-ends, wherein areas with hard-ends are marked by rectangles.

[0038] Fig. 2 shows a flowchart of the method according to the invention.

[0039] Fig. 3 schematically shows an apparatus according to the invention.

[0040] Fig. 4 schematically shows the architecture of a measurement system according to the invention.

[0041] Fig. 5 schematically shows typical shapes and characteristics of fibers, neps, and various hard-ends.

[0042] Fig. 6 is an illustration of transmissive sensor outputs for (A) a fiber, (B) a nep, and (C) a hard-end as functions of time.

[0043] Fig. 7 is a diagram illustrating a variable threshold for the method according to the invention.

[0044] Fig. 8 shows two-dimensional transmissive sensor images for (A) a first nep, (B) a second nep, and (C) a single piece of yam traveling at 30 m / s. Left-hand side: unprocessed images; right-hand side: processed images.

[0045] Fig. 9 illustrates a computer vision algorithm for processing two-dimensional images of fibers, neps, and various hard-ends.- 8 - P0651PF-WODESCRIPTION OF EMBODIMENTS1. Overview of the method

[0046] The method according to the invention is first explained with reference to the flowchart of Fig. 2. The method is for characterizing a sample of recycled textile fiber material (cf. Fig. 1) that includes at least one of fibers, neps, and hard-ends.

[0047] The sample of recycled textile fiber material is provided. The sample is individualized or singulated 21 by means of an individualizer or a singulator to produce individualized entities.

[0048] The individualized entities are transported 22 in an air flow by means of an aerodynamic transporter.

[0049] Measured parameters of the entities are determined 23 by means of at least one sensor while they are transported in the air flow. The measured parameters include time-of-flight, area-under-curve, peak, average signal height, and speed, and at least one of median signal height and sum of squared signal height.

[0050] The entities are classified 25 by means of a signal processor as at least one of fibers, neps, and hard-ends based at least in part on comparing 24 their measured parameters with thresholds.

[0051] Finally, the sample is characterized 26 by means of the signal processor using at least one result of the classification.2. Measurement system

[0052] Fig. 3 schematically shows a measurement system or apparatus that can be used for inline and offline measurements. The measurement system contains the following major functional blocks:100 Recycled fiber sample: recycled textile fiber sample, e.g., in the form of a fiber bundle, to be characterized.- 9 - P0651PF-WO200 Sample preparation module: prepares the recycled fiber sample 100 by separating fibers, neps, and hard-ends, and optionally routes them into one of two air streams based on their weights.101 Entities: single or multiple individual recycled fibers, neps, or hard-ends.300 Aerodynamic transporter: transporting mechanism, e.g., an air-entrained transfer shuttle, utilizing air flow for transporting fibers, neps, and hard-ends.400 Sensing module: responsible for sensing and computing various characteristics of fibers, neps, and hard-ends; comprises one or more sensors.500 Waste collection module: collects and keeps all the fibers, neps, and hard- ends.600 Controller module: controls all the components, processes all sensor data, distinguishes various fibers, neps, and hard-ends, computes fibers, neps, and hard-ends parameters, provides all the results to users through a user interface, and communicates various information to external equipment.700 External equipment: external information system and / or other textile equipment that utilizes the fibers, neps, and hard-ends measurements for process optimization.

[0053] The measurement system as shown in Fig. 3 operates as follows:1. The recycled fiber sample 100 is presented to the measurement system.2. The measurement system starts processing by first preparing the recycled fiber sample 100 via the sample preparation module 200, which opens the recycled fiber sample 100 and then creates single or groups of individualized entities 101.3. The entities 101 are transported to sensing module 400 by the aerodynamic transporter 300.4. The sensing module 400 determines various measured parameters such as time-of- flight, area-under-curve, peak, and average signal height for each single entity 101.- 10 - P0651PF-WO5. The controller module 600 processes the measured parameters to compute entity parameters such as length, diameter, weight, etc. of each single entity 101.6. The entities 101 are transported from the output of the sensing module 400 to the waste collection module 500 by the aerodynamic transporter 300.7. The controller module 600 not only processes the measured parameters but also controls all the operations and provides the results to users and the external equipment 700. The external equipment 700 may utilize these results to optimize its own operation based on the measurements and application.3. Measurement system architecture

[0054] Fig. 4 shows a more detailed architecture of the measurement system or apparatus and contains subsystems or modules described in the following.3.1 Sample preparation module 200

[0055] The sample preparation module 200 is responsible for bringing the fiber sample 100 into the measurement system and preparing the fibers for presentation to the sensing module 400. The fiber sample 100 may be found in various formats, such as fiber bale, carding mat, sliver, etc.

[0056] The sample preparation module 200 opens the sample 100 and individualizes each entity 101 via a fiber individualizer 201. There are multiple methods to accomplish this task. For example, one method is to use a drafting technique where the bundled fibers are stretched and thus opened via two rollers operating at different speeds. Yet another method is to use a roller and a pin cylinder to take a more aggressive approach to open the sample 100.

[0057] In either opening method, the aerodynamic transporter 300 is utilized to support the fiber individualization process as well as transport the entities 101 into the sensing module 400.

[0058] The sample preparation module 200 may use a two-phase flow separation scheme to route fibers, neps, and hard-ends into one of two air streams based on their weights (cf. Fig.- 11 - P0651PF-WO3). The dual air streams, coupled with dual sensors, improve the accuracy of all the measurements.3.2 Sensing module 400

[0059] The sensing module 400 senses all the quality parameters (fiber and recycling parameters as listed in the summary of the disclosure section) for each entitiy 101.

[0060] The aerodynamic transporter 300 continuously transports all the entities 101 from the fiber individualizer 201 into the sensing module 400.

[0061] The sensing is performed as each individual entity 101 or groups of individuals entities 101 continue to travel throughout the sensing module 400. The sensing module 400 comprises at least one sensor; in the embodiment of Fig. 4, an optical reflective sensor 402 and an optical transmissive sensor 404 are drawn. An illumination source 401 for illuminating the entities 101 may be present. The sensors 402 and 404 operate either continuously or intermittently, as triggered by a trigger and speed sensor 403 when the entities 101 are present. In either continuous or triggered mode, the sensors 402 and 404 may capture an image of the entities 101 based on the sensor type. For example, for a two-dimensional image sensor, an area image is captured. In another example, for a onedimensional line-scan sensor, a series of line images are captured. In yet another example, for a discrete photodiode, a series of single-point images is captured. In this disclosure, the word “image” is meant to signify a data vector in any number of dimensions. The images are then processed by the controller module 600.

[0062] There are several optical methods by which measured parameters of the entities 101 can be determined. For example, one is to use a reflection optical method for imaging the entities 101. Another method uses transmission optics. Yet another method uses both transmission and reflection optical methods.

[0063] The illumination source 401, reflective sensor 402, and transmissive sensor 404 have spectral characteristics that may include, but are not limited to, ultraviolet, visible, or infrared regions. Examples of electromagnetic illumination sources 401 include, but are not limited to, LEDs, halogen lamps, mercury vapor lamps, incandescent lamps, or xenon lamps. Examples of electromagnetic sensors 402, 404 include, but are not limited to, line-scan imagers, area-scan imagera, or photodiode type sensors. Yet other examples mix different- 12 - P0651PF-WOsensing technologies, for example, a reflection area-scan image sensor and a transmission photodiode sensor, or a reflection photodiode sensor and a transmission area-scan image sensor, or reflection area-scan sensor and a transmission area-scan image sensor, or reflection photodiode sensor and a transmission photodiode sensor.

[0064] In addition or as an alternative to optical sensors 402, 404, the sensor module 400 may comprise other types of sensors, e.g. capacitive sensors.

[0065] The trigger and speed sensor 403 senses the presence of the entities 101, measures their travel speed, and triggers the signal acquisitions from reflective sensor 402 and / or transmissive sensor 404. One example of a trigger and speed sensor 403 is two side-by-side split photodiodes at a fixed and known distance apart from one another, where the speed is computed by measuring the time taken for the entities 101 to travel across the two split photodiodes. Another example uses an area-scan imager, where speed is computed by measuring the time taken for the entities 101 to travel across the field-of-view of the imager. In yet other examples, the trigger and speed sensor 403 serves not only as the speed sensor but also as the reflective sensor 402 and / or the transmissive sensor 404.4. Controller module 600 and external equipment 700

[0066] The controller module 600 is responsible for at least the following three functions:1. Controlling;2. Processing; and3. Interfacing.

[0067] A controller 601 within the controller module 600 controls the operation of the measurement system, which includes, but is not limited to, controlling the fiber individualizer 201 operation, illumination source 401, sensors 402, 403, and 404, as well as the air-flow operation of the aerodynamic transporter 300.

[0068] A signal processor 602 within the controller module 600 processes the outputs (e.g., a series of images) of sensors 402 and 404 based on a processing algorithm 900. The interface 603 provides interfaces to various external equipment 700.- 13 - P0651PF-WO

[0069] The external equipment 700 includes, but is not limited to, a human user interface 701, an information system 703, and equipment 702. The equipment 702 may include, but is not limited to, textile machinery, which can use the information to optimize its own operation and / or the operation of other textile equipment.5. Waste collection module 500

[0070] The waste collection module 500 collects and houses all the entities 101 until they are removed either manually or automatically.6. Signal processing flow and measurement concepts

[0071] This section describes the signal processing flow and the measurement concepts. The processing algorithm 900 processes the signals (e.g., captured images) from sensors 402 and 404, and provides at least the following main results:1. Classification of each entity 101,2. Fiber quality measurements, and3. Hard-ends measurements,6.1 Classification

[0072] Fig. 5 schematically shows typical presentations of a single fiber 810, nep 811, and hard-ends 812-816. The processing algorithm 900 uses a classification algorithm to differentiate between the fibers 810, neps 811, and hard-ends 812-816 that are present in the entities 101. One example of a classification algorithm 900 is a traditional rule-based method that uses different measured parameters to differentiate between fiber 810, nep 811, and hard-ends 812-816. Another example of a classification algorithm 900 is an artificialintelligence (Al) method in which the algorithm is trained with various signals or images of fibers 810, neps 811, and hard-ends 812-816. The entities 101 are classified as at least one of fibers 810, neps 811, and hard-ends 812-816 based at least in part on comparing their measured parameters with threshold values.- 14 - P0651PF-WO

[0073] The classification algorithm 900 classifies each image of entities 101 into one of the following classes:1. Fiber: a single fiber 810.2. Nep: a single nep 811, which is an entanglement of multiple fibers.3. Hard-ends:a. A single yarn 812 that is left over from an unsuccessful fabric shredding during the recycling process.b. An entangled-yarn 813 that is left over from an unsuccessful fabric shredding during the recycling process.c. A woven fabric 814 that is left over from an unsuccessful fabric shredding during the recycling process.d. A knitted fabric 815 that is left over from an unsuccessful fabric shredding during the recycling process.e. A contamination 816 containing any nonfibrous materials left over in the recycled fiber from either an unsuccessful fabric shredding during the recycling process or contamination from any other processes.

[0074] Fig. 6 shows examples of outputs of a side-by-side split photodiode transmissive sensor 404 for typical entities 101 as they traverse across a field of view of the transmissive sensor 404. Examples of a rule-based classification use measured parameters such as time-of-flight, area-under-curve, peak, rise-time, and / or fall-time, which are distinctively different for typical fibers 810, neps 811, and hard-ends 812-816. The measured parameters time-of-flight and peak are plotted on the diagrams of Fig. 6. Such measured parameters are compared with thresholds to classify the entities 101 as fibers 810, neps 811, and hard-ends 812-816.

[0075] In another example, speed information is used to provide and / or improve the classification. Entity speed is typically different for fibers 810, neps 811, and hard-ends 812-816 due to their inherent size differences. The speed information on entities 101 can be- 15 - P0651PF-WOprovided, e.g., by the side-by-side split photodiode transmissive sensor 404 or by the trigger and speed sensor 403.

[0076] As further illustration of a classification, magnitude of an extinction signal of a transmission photodiode may be compared with a threshold, above which the entity 101 may no longer be considered to be a fiber 810, but may be a nep 811 or a hard-end 812-816. Further thresholds may be applied individually or to combinations of entity time-of-flight, peak, area-under-curve, average signal height, and velocity. It may further be advantageous to employ a variable threshold to at least one parameter, said threshold depending on the value of a second parameter, as is described below with reference to Fig. 7.

[0077] In other embodiments, the reflective sensor 402 uses an area-scan or line-scan imager to create a two-dimensional image of the entities 101 as they traverse across the field of view of the reflective sensor 402. In such a configuration, digital image processing techniques may be used to differentiate between fibers, neps, and hard-ends based on their distinct geometrical attributes and surface structures.

[0078] The classification is at least partially accomplished by comparing the measured parameters with threshold values. Examples of measured parameters are entity time-of-flight, area-under-curve, peak, average signal height, median signal height, sum of squared signal height, velocity, and acceleration. As an example, an entity 101 may be classified as a hard-end 812-816 in part by comparing the area-under-curve calculated from a sensor signal (cf. Fig. 6) with a minimum area-under-curve threshold. In one embodiment, the threshold is predetermined. In another embodiment, the threshold is variable.

[0079] The classification may involve multiple such attributes and threshold criteria. Criteria may be applied to combinations of attributes. It may be further advantageous to employ thresholds that vary with the value of certain measured parameters. For example, as part of the classification, an entity’s area-under-curve may be checked against a threshold that varies with the entity’s speed.

[0080] An example of a variable threshold is illustrated in Fig. 7. The example relates to a case where sensing is accomplished by side-by-side split photodiode transmission sensors providing two time-resolved extinction signals for each entity 101. A measured parameter Te may be determined for each entity 101, Te being the time between the photodiode signals- 16 - P0651PF-WOas the entity 101 exits the sensor module 400. Thus, Te has an inverse relationship with the entity’s exit speed. A varying threshold may be applied to the entity’s area-under-curve, requiring that the area-under-curve be greater than a minimum threshold value that varies linearly with the ratio of Te and a second, fixed threshold on Te. Fig. 7 shows a diagram in a coordinate system spanned by a horizontal axis along which the measured parameter Te is plotted and by a vertical axis along which the area-under-curve is plotted. A straight line 70 in the diagram represents the variable threshold. Thus, for an entity 101 to be classified as a hard-end 812-816, it must in part have an area-under-curve greater than the threshold 70 that itself depends on the entity’s speed.

[0081] A further step in classifying an entity 101 as a hard-end 812-816 may invoke a final threshold applied to a combination of entity time-of-flight, area-under-curve, peak, average signal height, entry velocity, exit velocity, median signal height, and sum of squared signal height. Such criteria can be applied to entities 101 that survived all previously applied thresholds. The threshold check may consist of a machine learning classification algorithm trained on a combination of entity time-of-flight, area-under-curve, peak, average signal height, entry velocity, exit velocity, median signal height, and sum of squared signal height.

[0082] Fig. 8 shows actual two-dimensional photographs (left-hand side) of two different neps 811 (Figs. 6(A) and 6(B)) and a single piece of yam 812 (Fig. 8(C)) traveling in the aerodynamic transporter 300 at 30 m / s. Fig. 8 also illustrates how image processing techniques (right-hand side) can segment the object to classify an object as a fiber 810, a nep 811, or a hard-end 812-816 based on its geometrical attributes, including, but not limited to, length, area, grey-levels, circumference, speed, and orientation. Additionally, the surface structure of the hard-ends 812-816 can be used to differentiate between woven fabric 814 and knitted fabric 815 (see Fig. 5).6.2 Fiber length measurement

[0083] Fiber length is an important fiber characteristic for both natural and manmade fibers. The expression “length” for a single fiber 810 (see Fig. 5) usually refers to a fully stretched length of the single fiber. In one embodiment, where sensors 402 and 404 (cf. Fig. 4) use a side-by-side split photodiode, the fiber length is computed by measuring the speed and time-of-flight of single fibers 810 as they traverse across the reflective sensor 402 and / or transmissive sensor 404, as shown Fig. 6(A). The speed of a single fiber 810 is measured by- 17 - P0651PF-WOthe trigger and speed sensor 403, whereas time-of-flight is measured by the reflective sensor 402 and / or the transmissive sensor 404. The sensors 402-404 can be calibrated with a reference fiber sample of known length, such that the results of the fiber speed measurement are outputted in a certain measurement unit, such as meters / second.

[0084] In other embodiments, reflective sensor 402 may utilize an area-scan or line-scan imager to create a two-dimensional image of fibers as they traverse across the imager. In such a configuration, digital image processing techniques may be used to measure the length of a single fiber 810 in terms of imager’ s pixel as shown in Fig.9(A), which can be converted to a measurement unit such as millimeters via a calibration process.

[0085] As a fiber sample 100 contains many single fibers 810, a statistical method is used to characterize the fiber lengths within the entire fiber sample 100. Examples of results of such statistical methods include minimum, maximum, average, standard deviation, and histogram distribution of fiber lengths in the fiber sample 100. Another example of statistical parameters is to compute the percentage of certain fiber lengths in the fiber sample 100.6.3 Neps measurement

[0086] A nep 811 (see Fig. 5) is defined as entanglement of multiple fibers and quantified based on its size, which is a measure of the entanglement portion. In one embodiment where sensors 402 and 404 use a side-by-side split photodiode, the nep size is computed by measuring the speed, peak, and time-of-flight of a single nep 811 as it traverses across the reflective sensor 402 and / or the transmissive sensor 404, as shown in Fig. 6(B). A reference sample of known nep size distribution can be used to calibrate the nep size to a certain measurement unit, such as micrometers.

[0087] In other embodiments, the reflective sensor 402 uses area-scan or line-scan imagers to create a two-dimensional image of the neps 811 as they traverse across the imager. In such a configuration, digital image processing techniques are used to measure the size of a single nep 811 in terms of the imager’s pixels, as shown in Fig. 9(B). The nep size is computed as either the area of the entanglement in square pixels, or the square root of the area, and converted to a certain measurement unit such as square micrometers or micrometers, respectively, via a calibration process.- 18 - P0651PF-WO

[0088] If the fiber sample 100 contains many neps 811, a statistical method can be used to measure the neps quality of the entire fiber sample 100. For example, minimum, maximum, average, standard deviation, and / or histogram distribution of all nep 811 size measurements are computed to characterize the neps quality for a given fiber sample 100. Another example of statistical parameters is to compute the number of fiber neps per unit weight of the fiber sample 100.6.4 Hard-ends measurement

[0089] In one embodiment, where the sensors 402 and 404 use a side-by-side split photodiode, the hard-ends parameters are computed as follows:1. Yarn: for a single yarn 812 or entangled yam 813 (see Fig. 5), length, diameter, and weight are computed. The length is computed by measuring the speed and time-of- flight of the single yam 812 or entangled yarn 813 as it traverses across the reflective sensor 402 and / or the transmissive sensor 404, as shown in Fig. 6(C). The diameter is computed from the average signal height of yarn 812 or entangled yam 813. The weight is estimated based on length and diameter, and converted to a certain measurement unit such as milligrams, using a calibration process.2. Fabric: for a woven fabric 814 or knitted fabric 815 (see Fig. 5), size and weight are computed. Size is computed based on the speed, time-of-flight, and average signal height of woven fabric 814 or knitted fabric 815 as it traverses across the reflective sensor 402 and / or the transmissive sensor 404, as shown in Fig. 6(C). The weight is computed by converting size to a certain measurement unit such as milligrams, using a calibration process.3. Contamination: for a contamination 816 (see Fig. 5), size and weight are measured.Size is computed based on the speed, time-of-flight, and average signal height of contamination 816 as it traverses across the reflective sensor 402 and / or the transmissive sensor 404, as shown in Fig. 6(C). The weight is computed by converting the size to a certain measurement unit such as milligrams, using a calibration process.

[0090] In other embodiments, the reflective sensor 402 uses area-scan or line-scan imagers to create a two-dimensional image of hard-ends 812-816 (see Fig. 5) as they traverse across- 19 - P0651PF-WOthe imager. In such a configuration, digital image processing techniques may be used to compute the hard-ends parameters as follows:1. Yarn: for a single yarn 812 (see Fig. 5), length, diameter, and weight are computed.Digital image processing techniques may be used to measure the length and diameter in terms of the imager’s pixels, as shown in Fig. 9(C). The length measurement can be converted to a certain measurement unit such as millimeters. The calibration is the ratio of the imager field of view (in millimeters) and the imager resolution. The weight is computed by length times diameter squared, and converted to a certain measurement unit such as milligrams, using a calibration process.2. Entangled-yarn: for an entangled-yam 813 (see Fig. 5), size and weight are computed.Digital image processing techniques may be used to measure the area in terms of the imager’s pixels, as shown in Fig. 9(D). The size is computed as either the area in pixels squared or the square root of the area, and converted to a certain measurement unit such as millimeters squared or millimeters, respectively, and the weight is computed in a certain measurement unit such as milligrams, using a calibration process.3. Fabric: for a woven fabric 814 or knitted fabric 815 (see Fig. 5), size and weight are computed. Digital image processing techniques may be used to measure the area in terms of the imager’s pixels, as shown in Figs. 9(E) and 9(F). The size is computed as either an area in pixels squared or the square root of area, and converted to a certain measurement unit such as millimeters squared or millimeter, respectively, and weight is computed in a certain measurement unit such as milligrams, using a calibration process.4. Contamination: for a contamination 816 (see Fig. 5), size and weight are measured.Digital image processing techniques may be used to measure the area in terms of the imager’s pixels, as shown in Fig. 9(G). The size is computed as either the area in pixels squared or the square root of the area, and converted to a certain measurement unit such as millimeters squared or millimeters, respectively, and the weight is computed in a certain measurement unit such as milligrams, using a calibration process.- 20 - P0651PF-WO

[0091] As the fiber sample 100 may contain many hard-ends 812-816, a statistical method can be used to characterize the entire fiber sample 100. For example, minimum, maximum, average, standard deviation, and / or histogram distribution of hard-ends length and size are computed to characterize the hard-ends 812-816 for a given fiber sample 100.

[0092] In addition to the hard-ends parameters mentioned above, we introduce two new additional statistical recycling quality parameters defined and computed as follows:1. Recycling opening index by weight (roiw), computed as:total weights of hard-endsroiw= 100 * (1 - total weight of the - - recycled fiber sample 100 .The roiw may be computed per type of hard-ends 812-816 or summation of all types.2. Recycling opening index by count (roic), computed as:, total weights of hard-endsroic= m * ( -2- ) ,total weight of the recycled fiber sample 100where m is a relative constant. For example, m = 1,000 provides recycling opening index for every 1,000 fibers. The roicmay be computed per type of hard- ends 812-816 or summation of all types.P0651PF-WOREFERENCE SIGNS21 Individualizing (step)22 Transporting (step)23 Determining measured parameters (step) 24 Comparing with thresholds (step)25 Classifying (step)26 Characterizing (step)70 Variable threshold (line in diagram) 100 Recycled fiber sample101 Entities (fibers, neps, and / or hard-ends) 200 Sample preparation module201 Fiber individualizer300 Aerodynamic transporter400 Sensing module401 Illumination source402 Optical reflective sensor403 Trigger and speed sensor404 Optical transmissive sensor500 Waste collection module600 Controller module601 Controller602 Signal processor603 Interface700 External equipment701 User interface702 Equipment703 Information system810 Fiber811 Nep812 Yarn (hard-end)813 Entangled yam (hard-end)814 Woven fabric (hard-end)815 Knitted fabric (hard-end)816 Contamination (hard-end)900 Processing algorithm- 22 - P0651PF-WO

Claims

CLAIMS1. A method for characterizing a sample (100) of recycled textile fiber material that includes at least one of fibers (810), neps (811), and hard-ends (812-816), the method comprising the steps of:a. individualizing (21) the sample (100) to produce individualized entities (101); b. transporting (22) the entities (101) in an air flow; andc. determining (23) measured parameters of the entities (101) while they are transported in the air flow;characterized in thatd. the measured parameters include time-of-flight, area-under-curve, peak, average signal height, and speed, and at least one of median signal height and sum of squared signal height;e. the entities are classified (25) as at least one of fibers (810), neps (811), and hard- ends (812-816) based at least in part on comparing (24) their measured parameters with thresholds; andf. the sample (100) is characterized using at least one result of the classification.

2. The method of claim 1, wherein the measured parameters additionally include acceleration.

3. The method of any one of the preceding claims, wherein entity parameters including length, diameter, area, and / or weight of the entities (101) are determined from the measured parameters.

4. The method of claim 3, wherein hard-end dimensions are determined from hard-end time-of-flight, speed, and average signal height.

5. The method of claim 4, wherein hard-end weight is determined from the hard-end dimensions and at least one of an assumed bulk density and an experimentally determined scaling factor.

6. The method of claim 3, wherein statistical parameters related to at least one of fibers (810), neps (811), and hard-ends (812-816) are determined from the entity parameters, and the sample (100) is characterized using the statistical parameters.- 23 - P0651PF-WO7. The method of claims 5 and 6, wherein a recycling opening index is determined from a total weight of classified hard-ends (812-816) and a total weight of the sample (100).

8. The method of claim 6, wherein the statistical parameters include minimum, maximum, average, standard deviation, and / or histogram distribution.

9. The method of any one of the preceding claims, wherein the thresholds are predetermined or variable.

10. The method of any one of the preceding claims, further comprising the step of separating hard-ends (812-816) from the fibers (810) and neps (811) based on entity weight, after the step of individualizing (21) the sample (100).

11. The method of any one of the preceding claims, wherein the step of classifying (25) the entities (101) includes further classifying the hard-ends (812-816) as at least one of yam (812), entangled yarn (813), woven fabric (814), knit fabric (815), and contamination (816).

12. The method of any one of the preceding claims, further comprising the step of computing from the measured parameters entity parameters comprising at least one of entity length, entity size, entity diameter, entity shape, entity weight, and entity texture.

13. The method of any one of the preceding claims, wherein the method is applied to at least one of offline laboratory test equipment and inline process monitoring equipment.

14. The method of any one of the preceding claims, wherein the step of determining (23) measured parameters of the entities (101) is accomplished by side-by-side split photodiode transmissive sensors providing two time-resolved extinction signals for each entity (101).

15. The method of any one of the preceding claims, wherein the step of determining (23) measured parameters of the entities (101) comprises imaging of the entities (101) via a one-dimensional or a two-dimensional image sensor.

16. An apparatus for characterizing a sample (100) of recycled textile fiber material that includes at least one of fibers (810), neps (811), and hard-ends (812-816), the apparatus comprising:- 24 - P0651PF-WOa. an individualizer (201) for individualizing the sample (100) to produce individualized entities (101);b. an aerodynamic transporter (300) for transporting the entities (101) in an air flow;andc. at least one sensor (402, 404) for determining measured parameters of the entities (101) while they traverse the aerodynamic transporter (300);characterized in thatd. the at least one sensor (402, 404) is configured to determine measured parameters including time-of-flight, area-under-curve, peak, average signal height, and speed, and at least one of median signal height and sum of squared signal height;e. the apparatus comprises a signal processor (602) configured to classify the entities (101) as at least one of fibers (810), neps (811), and hard-ends (812-816) based at least in part on comparing their measured parameters with thresholds; and f. the signal processor (602) is further configured to characterize the sample (100) using at least one result of the classification.

17. The apparatus of claim 16, wherein the at least one sensor (402, 404) is additionally configured to determine acceleration as a measured parameter.

18. The apparatus of claim 16 or 17, wherein the signal processor (602) is further configured to determine entity parameters including length, diameter, area, and / or weight of the entities (101) from the measured parameters.

19. The apparatus of claim 18, wherein the signal processor (602) is further configured to determine hard-end dimensions from hard-end time-of-flight, speed, and average signal height.

20. The apparatus of claim 19, wherein the signal processor (602) is further configured to determine hard-end weight from the hard-end dimensions and at least one of an assumed bulk density and an experimentally determined scaling factor.

21. The apparatus of claim 18, wherein the signal processor (602) is further configured to determine statistical parameters related to at least one of fibers (810), neps (811), and hard-ends (812-816) form the entity parameters and to characterize the sample (100) using the statistical parameters.- 25 - P0651PF-WO22. The apparatus of claims 20 and 21, wherein the signal processor (602) is further configured to determine a recycling opening index from a total weight of classified hard- ends (812-816) and a total weight of the sample (100).

23. The apparatus of claim 21, wherein the signal processor (602) is configured to compute statistical parameters including minimum, maximum, average, standard deviation, and / or histogram distribution.

24. The apparatus of any one of the claims 16-23, wherein the signal processor (602) is configured to compute thresholds that are predetermined or variable.

25. The apparatus of any one of the claims 16-24, further comprising a separator for separating the hard-ends (812-816) from the fibers (810) and neps (812) based on entity weight, the separator being arranged after the individualizer (201).

26. The apparatus of any one of the claims 16-25, wherein the signal processor (602) is further configured to further classify hard-ends (812-816) as at least one of yarn (812), entangled yarn (813), woven fabric (814), knit fabric (815), and contamination (816).

27. The apparatus of claim any one of the claims 16-26, wherein the signal processor (602) is further configured to compute from the measured parameters entity parameters comprising at least one of entity length, entity size, entity diameter, entity shape, entity weight, and entity texture.

28. The apparatus of any one of the claims 16-27, wherein the at least one sensor (402, 404) comprises at least one of an optical reflective sensor (402) and an optical transmissive sensor (404).

29. The apparatus of claim any one of the claims 16-28, wherein the at least one sensor (402, 404) includes side-by-side split photodiode transmissive sensors that provide two time- resolved extinction signals for each entity (101).

30. The apparatus of any one of the claims 16-29, wherein the at least one sensor (402, 404) includes a one-dimensional or a two-dimensional image sensor.

31. A textile offline laboratory test equipment or a textile inline process monitoring equipment comprising the apparatus of any one of the claims 16-30.- 26 - P0651PF-WO

Images