Control of an abrading process

Abstract

A method for abrading comprises: providing an abrasive article (ART1), which comprises a machine-readable code (QR1), selecting a target result (RM1,T), retrieving a performance model (MODEL31,GEN1, MODEL41,GEN1) from a database (DBM) based on the code (QR1), determining one or more usage parameters (GM1,GEN1) from the target result (RM1,T) by using the performance model (MODEL31,GEN1, MODEL41,GEN1), and abrading with the first abrasive article (ART1) according to the determined usage parameters (GM1,GEN1)

Description

[0001] CONTROL OF AN ABRADING PROCESS

[0002] FIELD

[0003] The present invention relates to controlling abrading.

[0004] BACKGROUND

[0005] It is known that an abrasive article, e.g. a grinding disk, may carry printed information about suitable usage parameters. For example, a grinding disk may have a text, which indicates that the maximum allowed rotation speed is 10000 revolutions per minute. A user desiring to reach maximum efficiency may subsequently grind a workpiece with the grinding disk such that the actual rotation speed is at the maximum allowed rotation speed. However, operating at the maximum rotation speed does not always provide the desired abrading result.

[0006] SUMMARY

[0007] An object is to provide a method for controlling abrading. An object is to provide an abrading system.

[0008] According to an aspect, there is provided a method, comprising:

[0009] - providing a first abrasive article (ARTi), which comprises a machine-readable code (QRi),

[0010] - selecting a target result (RMI,T),

[0011] - retrieving a performance model (MODEL3I,GENI , MODEL4I,GENI) from a database (DBM) based on the code (QRi),

[0012] - determining one or more usage parameters (GMI.GENI) from the target result (RMI.T) by using the performance model (MODEL3I,GENI , MODEL4I,GENI), and

[0013] - abrading with the first abrasive article (ARTi) according to the determined usage parameters (GMI.GENI).

[0014] According to an aspect, there is provided a method of claim 1 . According to an aspect, there is provided an abrading system of claim 16.

[0015] The scope of protection sought for various embodiments of the invention is set out by the independent claims. The embodiments, if any, described in this specification that do not fall under the scope of the independent claims are to be interpreted as examples useful for understanding various embodiments of the invention.

[0016] A performance model associated with an abrasive article may be accessed by using a machine-readable code of the abrasive article. The performance model may represent the abrasive performance of the abrasive article. Usage parameters suitable for providing a desired abrading result may be determined by using the performance model. A workpiece may be abraded with the abrading article according to the determined usage parameters. Abrading according to the determined usage parameters may increase the likelihood of achieving the desired abrading result.

[0017] The method may also optimize usage of the abrasive article through machine learning. The abrasive article may be attached to an abrading apparatus, which comprises one or more sensors for measuring usage parameters. The abrading apparatus may comprise one or more sensors for measuring abrading results. For example, the abrading apparatus may comprise a dust sensor for measuring the rate of removing material from a workpiece. A performance model representing the abrasive performance of the abrasive article may be trained by using the observed abrading results. The trained performance model may subsequently be used for determining optimum updated usage parameters for the abrasive article.

[0018] Training the performance model by using measured results and updating the usage parameters by using the trained performance model may enable optimum usage of the abrasive article.

[0019] The abrading effect of an abrasive article is typically not constant but may decrease over time due to clogging and / or wear. Ultimately, the abrasive article may need to be replaced with a new abrasive article. To achieve a constant surface finish, the abrading process may also need to compensate the change of the abrading performance of the abrasive article throughout a finishing operation. In an embodiment, the surface finish attained by the abrading may be substantially constant over the processed surface of the workpiece. For example, variation of the gloss value may be reduced or minimized over the processed surface of the workpiece.

[0020] The abrading process may be optimized by using the live data stream of measured abrading results from the abrading apparatus together with the performance model of the abrasive article. Results which may be measured during abrading may include e.g. dust concentration in a dust extractor, gloss value and / or torque applied to a rotating abrasive article. Usage parameters which can be adjusted based on the optimization may be e.g. rotation speed, pressing force, and / or transverse feed rate.

[0021] The method may comprise measuring one or more abrading results, and controlling abrading based on the measured abrading results.

[0022] The machine-learning performance model may be trained by using the measured results. The usage parameters may be determined by using the trained machinelearning model such that abrading with the abrasive article may provide desired abrading results.

[0023] The method may comprise determining one or more updated usage parameters from the target result by using the trained performance model, and abrading with an abrasive article according to the updated usage parameters.

[0024] The method may comprise storing the trained performance model in the database.

[0025] The abrading apparatus may rotate or oscillate an abrasive article by using an electric motor, which rotates at a rotation speed, and which provides a torque. The magnitude of an electric current of the motor may be an indication of the magnitude of the torque. The apparatus may measure the torque by measuring the current. A reduced torque at constant pressing force may be an indication that the abrasive article is e.g. clogged or has worn abrasive grains. The performance model of the abrasive article may be trained also by using the new information about the measured torque. Depending on the desired result, the abrading system may use the trained performance model e.g. to set an initial pressing force, set an initial rotation speed, change pressing force, change rotation speed, change transverse feed rate and / or replace the abrasive article with a new abrasive article, in order to achieve the desired result.

[0026] The desired result may be selected e.g. from the following list: provide a predetermined rate of removing material away from a workpiece, maximize the rate of removing material away from a workpiece, maximize an amount of material removed from a workpiece by using a single abrasive article, minimize costs for removing a given amount of material from a workpiece, minimize time for reaching a given gloss value of a workpiece, maximize a rate of change of gloss value of a workpiece.

[0027] An initial performance model for an abrasive article, or for a batch of abrasive articles may be determined e.g. based on experimental laboratory tests. The initial performance model may be stored in a model database such that the model is associated with the machine-readable code of the abrasive article (or with a batch of abrasive articles). The machine-readable code allows reliable identification of the abrasive article also after storage and transportation to an abrading site. The machine-readable code of the abrasive article may be read at the abrading site, and the initial performance model may be retrieved from the database based on the machine-readable code of the abrasive article. The abrasing system may determine suitable initial usage parameters from desired target results by using the initial performance model. The abrading process may be started by applying the initial usage parameters. The abrading apparatus may measure abrading results, the performance model may be trained by using the measured abrading results, and the usage parameters may be updated by using the trained performance model.

[0028] In an embodiment, the abrasive article may be attached to an abrading apparatus, which comprises a robot or a collaborative robot (cobot). The robot may enable use of additional usage parameters and / or may enable more accurate control of usage parameters. The robot may e.g. enable more accurate control of transverse feed rate and / or pressing force. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In the following examples, several variations will be described in more detail with reference to the appended drawings, in which

[0030] Fig. 1a shows, by way of example, a first abrasive article, which comprises a first machine-readable code,

[0031] Fig. 1 b shows, by way of example, a second abrasive article, which comprises a second machine-readable code,

[0032] Fig. 2a shows, by way of example, reading the code of the first abrasive article, and abrading with the first abrasive article,

[0033] Fig. 2b shows, by way of example, reading the code of the second abrasive article, and abrading with the second abrasive article,

[0034] Fig. 2c shows, by way of example, storing abrading results of the first abrasive article in a database,

[0035] Fig. 2d shows, by way of example, storing abrading results of the second abrasive article in a database,

[0036] Fig. 3a shows, by way of example, estimating abrading results from selected usage parameters by using a performance model,

[0037] Fig. 3b shows, by way of example, estimating suitable usage parameters from desired target results by using an inverse performance model,

[0038] Fig. 4a shows, by way of example, determining suitable usage parameters from desired target results by using an inverse performance model,

[0039] Fig. 4b shows, by way of example, training the inverse performance model by using measured abrading results, Fig. 4c shows, by way of example, determining updated usage parameters from selected target results by using the trained inverse performance model,

[0040] Fig. 5a shows, by way of example, providing an abrasive article to an abrading site, and determining suitable usage parameters for the abrasive article from selected target results by using an inverse performance model,

[0041] Fig. 5b shows, by way of example, training the inverse performance model by using measured abrading results,

[0042] Fig. 5c shows, by way of example, determining suitable usage parameters from selected target results by using the trained inverse performance model, and storing the trained inverse performance model in a database,

[0043] Fig. 6 shows, by way of example, method steps for optimizing usage parameters,

[0044] Fig. 7a shows, by way of example, determining suitable usage parameters from selected target results by using a performance model,

[0045] Fig. 7b shows, by way of example, training the performance model by using measured abrading results,

[0046] Fig. 7c shows, by way of example, determining suitable usage parameters from selected target results by using the trained performance model,

[0047] Fig. 8a shows, by way of example, controlling abrading by using usage parameters,

[0048] Fig. 8b shows, by way of example, controlling abrading by using usage parameters, Fig. 8c shows, by way of example, method steps for determining and using usage parameters,

[0049] Fig. 8d shows, by way of example, determining a usage parameter from a target result by using the model,

[0050] Fig. 8e shows, by way of example, determining a usage parameter from a target result by using the model,

[0051] Fig. 9a shows, by way of example, temporal evolution of parameters during abrading, in a situation where wear of the abrasive article is not compensated,

[0052] Fig. 9b shows, by way of example, temporal evolution of parameters during abrading, in a situation where wear of the abrasive article is compensated,

[0053] Fig. 9c shows, by way of example, temporal evolution of parameters during abrading, in a situation where wear of the abrasive article is compensated,

[0054] Fig. 10 shows, by way of example, an abrading system.

[0055] DETAILED DESCRIPTION

[0056] Referring to Fig. 1 a, a first abrasive article ARTi may comprise a machine- readable code QRi, which allows identification of the abrasive article ARTi, or which allows unique identification of a batch of abrasive articles, wherein the batch comprises the first abrasive article ART 1. The machine-readable code QRi may be unique. The machine-readable code QRi may be e.g. an optically readable matrix code.

[0057] Referring to Fig. 1 b, a second abrasive article ART2 may comprise a different machine-readable code QR2, which allows identification of the abrasive article ART2, or which allows unique identification of a batch of abrasive articles, wherein the batch comprises the second abrasive article ART2. The machine-readable code QR2 may be e.g. an optically readable matrix code.

[0058] Referring to Figs. 2a and 2b, abrasive articles ART1, ART2 may be attached to an abrading apparatus MAC1 , MAC2, and the abrasive articles ART1, ART2 may be used for abrading surfaces SRF1 , SRF2 of workpieces OBJ1 , OBJ2.

[0059] A first abrading apparatus MAC1 at a first abrading site SITE1 may comprise a reader READ1 for reading the machine-readable code QR1, QR2. The abrading may be performed according to usage parameters GM1, GM2. The abrading may provide abrading results RM1, RM2. The abrading apparatus MAC1 may comprise one or more sensors GSEN1 for measuring usage parameters, e.g. a pressing force. The abrading apparatus MAC1 may comprise one or more sensors RSEN1 for measuring abrading results, e.g. for measuring concentration of dust particles in a dust-laden airflow, which is drawn from the abrasive article to a dust extractor VAC1.

[0060] The abrading apparatus MAC1 may comprise a dust extractor VAC1 for extracting particles, which are removed from a workpiece OBJ1 during abrading. The dust extractor VAC1 may draw the particles together with an air flow via a duct DLIC1 . The dust extractor VAC1 may draw a dust-laden air flow via a duct DLIC1. The apparatus may comprise the dust sensor RSEN1 for measuring the concentration of the particles in the dust-laden air flow. The dust sensor RSEN1 may provide a dust concentration signal DRMEAS, which is proportional to the concentration of particles in the air flow. The dust sensor RSEN1 may provide a dust concentration signal DRMEAS, which is proportional to the rate of removing material from a workpiece OBJ1. The rate of removing material may be expressed e.g. in grams per second (g / s).

[0061] A second abrading apparatus MAC2 at a second abrading site SITE2 may comprise a reader READ2 for reading the machine-readable code QR1, QR2. The abrading may be performed according to usage parameters GM1, GM2. The abrading may provide abrading results RM1, RM2. The abrading apparatus MAC2 may comprise a sensor GSEN2 for measuring a usage parameter. The abrading apparatus MAC2 may comprise a sensor RSEN2 for measuring an abrading result. The apparatus MAC2 may comprise a dust extractor VAC2.The dust extractor VAC2 may draw a dust-laden air flow via a duct DLIC2.

[0062] Referring to Fig. 2c, the usage parameters GMi and / or the measured abrading results RMi may be stored in an abrading result database DBR such that the usage parameters GMi and the abrading results RMi of the first abrasive article ARTi may be retrieved from the database DBR based on the machine-readable code QRi of the first abrasive article ARTi.

[0063] Referring to Fig. 2d, the usage parameters GM2 and / or the measured abrading results RM2 may be stored in the abrading result database DBR such that the usage parameters GM2 and the abrading results RM2 of the second abrasive article ART2 may be retrieved from the database DBR based on the machine- readable code QR2 of the second abrasive article ART2.

[0064] Referring to Fig. 3a, the abrading performance of the first abrasive article ARTi may be represented by a performance model MODEL3i. The performance model MODEL3i may be used for estimating abrading results RM I .E from selected usage parameters GMI ,T.

[0065] Referring to Fig. 3b, the abrading performance of the first abrasive article ARTi may also be represented by an inverse performance model MODEL4i. The inverse performance model MODEL4i may be used for determining suitable usage parameters GMi.E from desired abrading results RMI ,T. The subscript "1" may refer to the "first" abrasive article. The subscript "E" may refer to an "estimate". The subscript "T" may refer to a "target".

[0066] The inverse performance model MODEL4i represents the abrading performance of the first abrasive article ARTi. Also the inverse performance model MODEL4i may be called as a performance model.

[0067] Referring to Fig. 4a, an initial inverse performance model MODEL4I ,GENI may be used for determining suitable usage parameters GMI.GENI from one or more desired abrading results RMI ,T. The subscript GEN1 may refer to the first generation. The desired abrading results RMI ,T may comprise one or more desired target values ri,i,T, ri,2j. The target value ri,i,T may specify e.g. a target material removal rate. The usage parameters GMI.GENI may comprise usage parameters gi,i,GENi, gi,2,GENi, gi,3,GENi. The usage parameter gi.i.GENi may specify e.g. a pressing force. The usage parameter gi,2,GENi may specify e.g. a rotation speed.

[0068] Abrading with the first abrasive article ARTi according to the usage parameters GM I .GENI may provide abrading results RMI.GENI .

[0069] Referring to Fig. 4b, the inverse performance model MODEL4I ,GENI may be trained by using measured abrading results RMI.GENI , and by using the actual usage parameters, so as to form a trained inverse performance model MODEL4 I ,GEN2- The subscript GEN2 may refer to the second generation. The usage parameters GMI.GENI may comprise e.g. usage parameters gi.i.GENi, gi,2,GENi, gi,3,GENi . The measured abrading results RM I .GENI may comprise e.g. one or more measured values H .I .MEAS, H ,2,MEAS. The measured result ri.i.MEAs may specify e.g. a measured material removal rate.

[0070] Referring to Fig. 4c, the trained inverse performance model MODEL4I ,GEN2 may be used for determining updated usage parameters GMi,GEN2 from the desired abrading results RM I ,T. The updated usage parameters GM I ,GEN2 may comprise e.g. updated usage parameters gi,i,GEN2, gi,2,GEN2, gi,3,GEN2.

[0071] Abrading according to the updated usage parameters GMI ,GEN2 is likely to provide updated abrading results RMI ,GEN2, which may closely match with the desired abrading results RMI ,T. A difference between the updated abrading results RM I ,GEN2 and the desired abrading results RMI ,T may be smaller than a difference between the first abrading results RMI .GENI and the desired abrading results RMi T.

[0072] An iteration cycle may comprise determining usage parameters by using the model, obtaining abrading results by abrading according to the usage parameters, and training the model by using the abrading results. The iteration cycle may be optionally repeated several times in order to further reduce the difference between the abrading results and the desired target results.

[0073] Fig. 5a shows an abrading system GSYS1 , and a manufacturing site FCT1. Abrasive articles ARTi, ART2, ARTk are produced at one or more manufacturing sites FCT1 by production processes PROC1, PROC2, PROCk. The abrasive articles ART1, ART2, ARTk may be produced according to different production parameters, and the abrasive articles ART1, ART2, ARTk may have different abrasive performance. A model MODEL4i may represent the abrasive performance of the article ART1. A model MODEL42 may represent the abrasive performance of the article ART2. A model MODEL4k may represent the abrasive performance of the article ARTk. The abrasive article ART1 may have a machine- readable code QRi.The abrasive article ART2 may have a machine-readable code QR2. The abrasive article ARTk may have a machine-readable code QRk.

[0074] Initial performance models MODEL4I,GENI, MODEL42,GENI, MODEL4k,GENi for the abrasive articles ART1, ART2, ARTk may be determined e.g. by using results obtained in a test laboratory and / or by theoretical calculations.

[0075] The models may be stored in a model database DBM such that the model MODEL4i of each abrasive article ART 1 may be retrieved from the database DBM based on the machine-readable code QR1 of said abrasive article ART1.

[0076] The produced abrasive articles ART1, ART2, ARTk may be transported and / or stored in one or more storages STO1. The machine-readable codes QR1, QR2, QRk allow identification of the abrasive articles ART1, ART2, ARTk after the storage and the transportation. The identity of an abrasive article or the identity of a batch of the abrasive articles may be reliably determined by reading the machine-readable code after the storage and the transportation.

[0077] The abrasive articles ART1, ART2, ARTk may be transported to one or more abrading sites SITE1 . The abrasive article ART1, ART2, ARTk may be attached to an abrading apparatus MAC1 , which is located at an abrading site SITE1. The abrading apparatus MAC1 comprises a reader READ1 for reading the machine- readable codes QR1, QR2, QRk of the abrasive articles ART1, ART2, ARTk. For example, the first abrasive article ART1 may be attached to abrading apparatus MAC1 , and the reader READ1 may read the machine-readable code QR1 of the first abrasive article ART1. The initial performance model MODEL4I,GENI of the first abrasive article ART 1 may be retrieved from the database DBM based on the machine-readable code QRi.

[0078] A user of the abrading apparatus MAC1 may select desired target results RM I .T e.g. by using a user interface LIIF1 . The abrading system GSYS1 may determine initial usage parameters GMi.GENi from the target results RMi by using the initial (inverse) performance model MODEL4I ,GENI . A workpiece 0BJ1 may be abraded by using the abrasive article ARTi according to the determined initial usage parameters GMI.GENI .

[0079] Referring to Fig. 5b, the abrading apparatus MAC1 may comprise one or more sensors for obtaining abrading results RMI .GENI in a situation where the workpiece OBJ1 is abraded with the abrasive article ARTi according to the usage parameters GMI.GENI . The model MODEL4I ,GENI may subsequently be trained by using the abrading results RM I .GENI and by using the usage parameters GMI.GENI, so as to provide a trained model MODEL4I ,GEN2.

[0080] The model may be trained by supervised learning by using measured abrading results. The model may be used for estimating optimum updated usage parameters for the abrasive article.

[0081] Referring to Fig. 5c, the abrading system GSYS1 may determine updated usage parameters GMi,GEN2from the target results RMi by using the trained (inverse) performance model MODEL4I ,GEN2. The abrasive article ARTi may subsequently be used for abrading according to the updated usage parameters GMI ,GEN2.

[0082] In an embodiment, the updated usage parameters GM I ,GEN2 may also be used when abrading with another abrasive article of the same manufacturing batch.

[0083] Fig. 6 shows, by way of example, method steps for optimizing abrading according to measured abrading results.

[0084] A target result RMijfor abrading may be selected in step 1010.

[0085] The machine-readable code QRi of an abrasive article ARTi may be read at an abrading site SITE1 in step 1015. An initial performance model MODEL3I ,GENI , MODEL4I ,GENI representing the performance of the abrasive article ARTi may be retrieved from a model database DBM based on the machine-readable code QRi of the abrasive article ARTi in step 1020.

[0086] Initial usage parameters GMI.GENI for using the abrasive article ARTi may be determined from the target result RMi by using the initial model MODEL3I ,GENI , MODEL4 I ,GENI , which represents the performance of said abrasive article ARTi (step 1025).

[0087] The abrasive article ARTi may be used for abrading according to the determined initial usage parameters GMI.GENI in step 1030.

[0088] Abrading results RMI .GENI may be measured in the situation where the abrasive article ARTi is used for abrading according to the usage parameters GMI.GENI (step 1035).

[0089] The model MODEL31.GEN1, MODEL4 I .GENI representing the performance of the abrasive article ARTi may be trained by using the obtained abrading results RMI .GENI in step 1040.

[0090] Updated usage parameters GMI ,GEN2 for using the abrasive article ARTi may be determined from the target result RM I ,T by using the trained model MODEL3I ,GEN2, MODEL4I ,GEN2, which represents the performance of said abrasive article ARTi (step 1045).

[0091] The abrasive article ARTi or another abrasive article ARTi of the same manufacturing batch may be used for abrading according to the updated usage parameters GMI ,GEN2 in step (1050).

[0092] An iteration cycle may comprise e.g. the steps 1035, 1040, 1045, 1050. If needed, the iteration cycle may be repeated one or more times in order to further reduce the difference between the measured abrading results RMi and the desired target results RMI ,T. The trained model MODEL3I ,GEN2, MODEL4I ,GEN2 may be optionally stored in a database DBM in step 1061. The trained model may subsequently be retrieved from the database DBM based on the machine-readable code QRi of the abrasive article ART 1.

[0093] The measured abrading results RMI.GENI , RMI ,GEN2 may be optionally stored in a database DBR in step 1O62.The results may subsequently be retrieved from the database DBM based on the machine-readable code QRi of the abrasive article ARTi.

[0094] The usage parameters GMI.GENI , GMi,GEN2may be optionally stored in a database DBR in step 1063. The usage parameters may subsequently be retrieved from the database DBM based on the machine-readable code QRi of the abrasive article ARTi.

[0095] The performance model MODEL3 determines results from usage parameters, whereas the inverse performance model MODEL4 determines usage parameters from results. However, also the performance model MODEL3 may be used in an iterative manner for providing usage parameters from results.

[0096] Fig. 7a shows how the usage parameters GMI.GENI may be determined from the desired target result RM I ,T by using the performance model MODEL3I ,GENI . The performance model MODEL3I ,GENI may determine estimated result RMI .E from usage parameters GMi. The usage parameters GMi may be iteratively varied until the estimated result RMI .E matches the desired target result RMI ,T. The usage parameters GMi which provide the matching result may be outputted as the initial usage parameters GMI.GENI .

[0097] Referring to Fig. 7b, the performance model MODEL3I ,GENI may be trained by using the measured abrading results and by using the usage parameters, so as to form a trained performance model MODEL3I ,GEN2.

[0098] Referring to Fig. 7c, updated usage parameters GM I ,GEN2 may be determined from the desired target result RM I ,T by using the trained performance model MODEL3I ,GEN2. The performance model MODEL3I ,GEN2 may determine estimated result RMI .E from usage parameters GMi. The usage parameters GMi may be iteratively varied until the estimated result RMI .E matches the desired target result RMI ,T. The usage parameters GMi which provide the matching result may be outputted as the updated usage parameters GMI ,GEN2.

[0099] In general, the abrasive articles may be e.g. selected from the following list: coated abrasive, flexible abrasive disk, flexible abrasive belt, endless flexible abrasive belt, abrasive mesh article, perforated coated abrasive article, bonded abrasive, abrading wheel, honing stone.

[0100] The term abrading may herein refer e.g. to grinding, polishing, lapping, honing, or superfinishing, when performed by using an abrasive article.

[0101] The abrasive articles may comprise abrasive grains. The grit size of the abrasive grains may be selected e.g. from the range of 40 to 5000 (FEPA standard). The abrasive grains may comprise e.g. aluminum oxide, silicon carbide, cubic boron nitride, and / or diamond. The abrasive grains may comprise ceramic material.

[0102] The machine-readable code may be read by a reader device. The machine- readable code may be an optically readable matrix code. The machine-readable matrix code may be a two-dimensional barcode. The matrix code may be e.g. selected from the following list: QR code, Aztec code, multi-row barcode, stacked linear barcode, circular barcode, Semacode, Dotcode.

[0103] The machine-readable code may comprise a unique identifier, which allows unique identification of an abrasive article, or which allows unique identification of a batch of abrasive articles. The machine-readable code may optionally further comprise additional data, e.g. production parameters (e.g. adhesive composition), limit values of abrading parameters (e.g. maximum allowable rotation speed) and / or performance parameters.

[0104] An optically readable matrix code may be visible on a surface of the abrasive article. The matrix code may be e.g. printed directly on the abrasive article. The matrix code may be printed on an adhesive label, which is attached to the abrasive article. A single abrasive article may have one or more identical matrix codes. The abrasive article may have one or more identical matrix codes on an abrading side of the abrasive article and / or the abrasive article may have one or more identical matrix codes on a rear fastening side of the abrasive article. A matrix code located on the wearable surface of the abrasive article may be read also when the abrasive article is mounted on an abrading apparatus. A matrix code located on a wearable surface of the abrasive article may be read by the reader also when the abrasive article is rotated and / or oscillated by the abrading apparatus. A matrix code located on a wearable surface of the abrasive article may be advantageously read before the abrasive article is used for abrading. In an embodiment, the matrix code located on the wearable surface may be rapidly damaged when the abrasive article is used for abrading.

[0105] The abrading apparatus may comprise a reader for reading the machine- readable code. The abrading apparatus may comprise an optical reader for reading the matrix code. The reader may comprise e.g. an illuminating unit to illuminate the matrix code, a camera to capture an image of the illuminated matrix code, and a signal processor to convert the captured image into a digital code. The illuminating unit may comprise e.g. a flashlamp to enable fast reading of a matrix code, which may be located e.g. on a moving abrasive article.

[0106] The machine-readable code may be read before the abrasive article is attached to an abrading apparatus at an abrading site. The machine-readable code may be read after the abrasive article is attached to an abrading apparatus but before the abrasive article is used for abrading at an abrading site. The machine- readable code may be read during abrading at an abrading site. The machine- readable code may be read after abrading but before the abrasive article is separated from the abrading apparatus at an abrading site. The machine- readable code may be read after the abrasive article has been separated from the abrading apparatus at an abrading site.

[0107] A matrix code located on a rear surface of the abrasive article may be read before the abrasive article is mounted on an abrading apparatus, and / or the matrix code located on a rear surface of the abrasive article may be read after the abrasive article has been separated from the abrading apparatus.

[0108] In an embodiment, the abrading apparatus may be arranged to allow reading a matrix code located on a rear surface of the abrasive article also in a situation where the rear surface of the abrasive article is in contact with the abrading apparatus. For example, the rear surface of the abrasive article may be attached to a backing pad of the abrading apparatus, wherein the backing pad may be transparent or may have a slot so that the reader of the apparatus may read the matrix code through the backing pad.

[0109] The performance model MODEL3I ,GENI , MODEL3I ,GEN2 may comprise e.g. a regression function, a neural network and / or a decision tree. The inverse performance model MODEL4I ,GENI , MODEL4I ,GEN2 may comprise e.g. a regression function, a neural network and / or a decision tree.

[0110] The performance model MODEL3I ,GENI , MODEL3I ,GEN2 may comprise a function that maps usage parameters to results. The machine-learning model MODEL3I ,GENI , MODEL3I ,GEN2 may be trained e.g. by supervised learning by using input vectors and output values. An input vector may comprise several usage parameters. Actual usage parameters may be used as input vectors. Measured results may be used as the corresponding output values.

[0111] The inverse performance model MODEL4I ,GENI , MODEL4I ,GEN2 may comprise a function that maps results to usage parameters. The machine-learning model MODEL4 1.GEN1, MODEL4 I ,GEN2 may be trained e.g. by supervised learning by input values and output vectors. An output vector may comprise several usage parameters. Measured results may be used as input values. Actual usage parameters may be used as the corresponding output vectors.

[0112] The machine-learning model MODEL3I ,GENI , MODEL3I ,GEN2, MODEL4I ,GENI , MODEL4 I ,GEN2 may be implemented also as a neural network. A neural network is a structure comprising several layers of successive computation. A layer may comprise one or more units or neurons performing an elementary computation. A unit may be connected to one or more units, and the connection may have associated a weight. The weight, also called as a “weight parameter”, may be used for scaling the signal passing through the associated connection. Weight may be a learnable parameter, i.e., values, which can be learned from training data.

[0113] The architectures for the neural network may be e.g. feed-forward architecture and / or recurrent architecture. A feed-forward neural networks is such that there is no feedback loop: each layer takes input from one or more of the layers before and provides its output as the input for one or more of the subsequent layers. Also, units inside a certain layer take input from units in one or more of preceding layers and provide output to one or more of following layers. In a recurrent neural network, there is a feedback loop, so that the network becomes stateful, i.e., it is able to memorize information or a state.

[0114] The neural network may learn properties from input data in supervised way or in unsupervised way. Such learning is a result of a training algorithm, or of a metalevel neural network providing a training signal. In general, the training algorithm comprises changing some properties of the model so that its calculated output is as close as possible to an observed output, as specified by training data.

[0115] Training can be performed in several ways. The main ones are supervised, unsupervised, and reinforcement training. In supervised training, the model is provided with input-output pairs, where the output may be a label. In unsupervised training, the model is provided only with input data (and also with output raw data in case of self-supervised training). In reinforcement learning, the supervision is sparser and less precise; instead of input-output pairs, the model gets input data and, sometimes, delayed rewards in the form of scores (e.g., -1 , 0, or +1 ).

[0116] In order to enable a neural network to perform a task, an untrained model has to go through a training phase. The training phase is the development phase, where the neural network learns to perform the final task. The training algorithm comprises changing some properties of the neural network so that its computed output (estimate) is as close as possible to the real observed output (observed results). A training data set that is used to train the neural network is supposed to be the representative of the data on which the neural network will be used. During training, the neural network uses the examples in the training data set to modify its learnable parameters (e.g., its connections’ weights) in order to achieve the desired task. Input to the neural network is the data, and the output of the neural network represents the desired task. The training may happen by minimizing or decreasing the output’s error. For example, in the case of classification of results, the output of the neural network can be used to derive a class or category index which indicates the class or category that the result belongs to. Training usually happens by minimizing or decreasing the output’s error. Examples of losses are mean squared error, cross-entropy, etc. In a deep learning technique, training is an iterative process, where at each iteration, the algorithm modifies the weights of the model to make a gradual improvement of the network’s output, i.e., to gradually decrease the error.

[0117] Training a neural network may be an optimization process. The goal of the optimization or training process is to make the model learn the properties of the data distribution from a training dataset. The goal may be to learn to use a training dataset in order to learn to generalize to previously unseen data, i.e., data which was not used for training the model. This is usually referred to as generalization. In practice, data may also be split into at least two sets, namely a training set and a validation set. The training set may be used for training the network, i.e., to modify its learnable parameters in order to minimize the output’s error. The validation set may be used for checking the performance of the network on data, which was not used to minimize the error, as an indication of the final performance of the model.

[0118] After training, the trained model is applied to new data during an inference phase, in which the model performs the desired task to which it has been trained for. As a result of the inference phase, the model provides an output which is a result of the inference on the new data.

[0119] Referring to Fig. 8a, a user may select one or more target results RMI ,T for abrading with an abrasive article ART. The target result RMI ,T may specify e.g. a desired mass removal rate (e.g. in grams per second, g / s).

[0120] Reference values for usage parameters may be determined by using the performance model MODELS or the inverse performance model MODEL4. A suitable mode MODELS or MODEL4 may be retrieved from a memory of the abrading system GSYS1 based on a code of the abrasive article ART.

[0121] A reference value may specify e.g. a reference value NRPMREF for the rotation speed of the abrasive article ART during abrading. A reference value may specify e.g. a reference value FPREF for the pressing force, which is applied to the abrasive article ART during abrading. The reference values may be communicated to a control unit of the abrading apparatus MAC1 . The control unit may control operation of the abrading apparatus MAC1 by using the reference values NRPMREF, FPREF as the initial usage values for the abrading. In particular, the control unit may control the rotation speed of the motor of the abrading apparatus MAC1 such that the initial rotation speed of the abrasive article ART at least during the first few seconds of the abrading is equal to the reference rotation speed value NRPMREF. The control unit may control a robot such that the initial pressing force at least during the first few seconds of the abrading is equal to the reference value FPREF of the pressing force.

[0122] The abrading apparatus MAC1 may measure one or more abrading results. The results may be measured continuously in real time, or after an abrading operation has been completed. For example, the abrading apparatus MAC1 may comprise a sensor for measuring the torque applied to the abrasive article ART. To the first approximation, the torque applied to the abrasive article ART may be determined from an electric current IC of the motor, which rotates the abrasive article ART. The abrading apparatus MAC1 may comprise a sensor RSEN1 for measuring concentration of dust in a dust-laden air flow DAIR1 , which is drawn via the abrading apparatus MAC1 to a dust extractor VAC1 . To the first approximation, the measured dust concentration DRMEAS may be proportional to the mass removal rate caused by abrading with the abrasive article ART. To the first approximation, the dust concentration may be proportional to the rate of removing material from the workpiece OBJ1 .

[0123] The control unit of the abrading apparatus MAC1 may adjust the pressing force FP e.g. such that the measured electric current ICMEAS of the motor is kept substantially equal to the initial value of the electric current.

[0124] The control unit of the abrading apparatus MAC1 may adjust the rotation speed NRPM of the motor and / or the pressing force FP e.g. such that the measured dust concentration DRMEAS is kept substantially equal to the initial value of the measured dust concentration DRMEAS.

[0125] The measured abrading results ICMEAS, DRMEAS may be communicated for training the model MODEL3 and / or MODEL4. The total amount of material removed from a workpiece OBJ1 , and / or the rate of removing material from the workpiece 0BJ1 may be measured e.g. by using weighing device WSEN1 (weight sensor). The measured weighing results RMMEAS may be communicated for training the model MODEL3 and / or MODEL4.

[0126] The measured dust concentration DRMEAS and the measured rate RMMEAS of removing material from the workpiece 0BJ1 may allow determining a calibration coefficient, which establishes a relation between the measured dust concentration DRMEAS and the measured rate of removing material from the workpiece 0BJ1 . The calibration coefficient may be stored in a database as such and / or the calibration coefficient may be included in the trained model MODEL3 and / or MODEL4. The calibration coefficient may be retrieved from the database e.g. based on the code of an abrasive article. The same calibration coefficient may be used e.g. for all abrasive articles of a manufacturing batch.

[0127] Referring to Fig. 8b, the MODEL3 or MODEL4 may be optionally used to provide reference values DRREF, ICREF for one or more measured results DRMEAS, ICMEAS.

[0128] The reference values DRREF, ICREF may be used e.g. for checking reliability of the control system CSYS1 . The control system of the abrading apparatus MAC1 may compare the initial measured results DRMEAS, ICMEAS with the determined reference values DRREF, ICREF. The control system may e.g. provide an indication ("alarm") to a user if the difference between the initial measured results DRMEAS, ICMEAS and the reference values DRREF, ICREF is greater than a predetermined limit. The control system may e.g. start operation in fail safe mode if the difference between the initial measured results DRMEAS, ICMEAS and the reference values DRREF, ICREF is greater than the predetermined limit.

[0129] Fig. 8c shows method steps:

[0130] - for determining suitable usage parameter values FPREF, NRPMREF by using a model MODEL3, MODEL4,

[0131] - for abrading according to the determined usage values,

[0132] - for training the model MODEL3, MODEL4 based on measured abrading results RMMEAS.

[0133] One or more target results RMI ,T may be selected (selecting step 1110). A user of the abrading system GSYS1 may input a target result RM I ,T by using a user interface LIIF1. A target result RM I .T may specify e.g. a target rate for material removal. A target result RMI ,T may specify e.g. the total mass of material which should be removed from one or more workpieces OBJ1 during the whole operating lifetime of an abrasive article ARTi.

[0134] The code QRi of the abrasive article ART 1 may be read to determine the identity and / or manufacturing batch of the abrasive article (identification step 1115).

[0135] A model MODELS 1.GEN1, MODEL4 I ,GENI , which is relevant for the abrasive article ARTi, may be retrieved from a database DBM based on the code QRi (retrieving step 1 120).

[0136] Suitable reference values FPREF, NRPMREF of usage parameters may be determined from the one or more target results RM I ,T by using the retrieved model MODEL3I ,GENI , MODEL4I ,GENI (determining step 1125).

[0137] The abrasive article ARTi may be used for abrading by using the determined reference values FPREF, NRPMREF (Abrading step 1130). One or more workpieces OBJ1 may be abraded by using the abrasive article ARTi according to the determined reference values FPREF, NRPMREF.

[0138] In particular, the pressing force FP and / or the rotation speed NRPM may be controlled based on measured dust concentration DRMEAS and / or based on measured electric current ICMEAS of the motor, by using the determined reference values FPREF, NRPMREF as initial values.

[0139] One or more abrading results RMMEAS may be measured (measuring step 1135). A measured abrading result RMMEAS may be e.g. the measured rate of removing material from the workpiece OBJ1 . A measured abrading result RMMEAS may be e.g. the total mass of material which is removed from one or more workpieces OBJ1 during the whole operating lifetime of the abrasive article ARTi. The abrading result may be measured e.g. by measuring a change of weight of the workpiece OBJ1 and / or by weighing the material removed from the workpiece OBJ1. The model MODEL3I ,GENI , MODEL4I ,GENI may be trained by using the one or more abrading results RMMEAS (training step 1140). The unit of the measured abrading result (e.g. g / s) may correspond to the unit of the target result, so as to facilitate training of the model. The training may provide a model MODEL3I ,GEN2, MODEL4 I ,GEN2 of the next generation.

[0140] The previously determined reference values FPREF, NRPMREF may be associated with the corresponding measured abrading results RMMEAS to provide input values and the corresponding output values. The model MODEL3I ,GENI , MODEL4I ,GENI may be trained by using the input values and the corresponding output values.

[0141] Updated reference values FPREF, NRPMREF of usage parameters may be optionally determined from one or more target results RMI ,T by using the trained model MODEL3I ,GEN2, MODEL4I ,GEN2 (determining step 1145).

[0142] One or more workpieces OBJ1 may be optionally abraded by using the abrasive article ARTi according to the updated reference values FPREF, NRPMREF (abrading step 1150). One or more workpieces OBJ1 may be optionally abraded by using another abrasive article ARTi of the manufacturing batch of the first abrasive article ARTi according to the updated reference values FPREF, NRPMREF. The abrading may be controlled by using the updated reference values FPREF, NRPMREF.

[0143] The measuring step 1135, the training step 1140, the determining step 1145, and the abrading step 1150 may be repeated one or more times to iteratively improve the model MODEL3i, MODEL4i.

[0144] Improving the model may enable reducing a difference between a target result and a measured abrading result, when abrading is performed according to the reference values, which are determined by using the model.

[0145] The trained model MODEL3I ,GEN2, MODEL4I ,GEN2 may be optionally stored in the database DBM (storing step 1161 ). Measured results RMMEAS, DRMEAS, ICMEAS may be optionally stored in the database DBM (storing step 1162).

[0146] Usage parameters FP, NRPM recorded during the abrading may be optionally stored in the database DBM (storing step 1163).

[0147] The reference values FPREF, NRPMREF of the usage parameters FP, NRPM may be optionally stored in the database DBM (storing step 1164).

[0148] Referring to Fig. 8d, one or more reference values of one or more usage parameters may be determined from one or more desired abrading results by using the model. In particular, a reference value FPREF of pressing force and a reference value N RPMREF of rotation speed may be determined from a desired abrading result. The desired abrading result may be e.g. the desired material removal rate RMI ,T. The reference value may be used as the best available estimate of an optimum usage parameter.

[0149] The model MODEL3I ,GEN2, MODEL4I ,GEN2 may comprise e.g. a part, which represents a relation between the pressing force FP applied to the abrasive article ARTi and the corresponding material removal rate RM. The relation may be approximated e.g. by a linear function or by a polynomial function. For example, a pressing force FPi may correspond to a rate RMi, a pressing force FP2 may correspond to a rate RM2, a pressing force FP3 may correspond to a rate RM3, a pressing force FP4 may correspond to a rate RM1. The symbols RM1, RM2, RM3, RM4 denote different values of the target result. The symbols FP1, FP2, FP3, FP4 denote different values of the pressing force.

[0150] A reference pressing force FPREF which is likely to provide a target material removal rate RM I ,T may be determined by using the model MODEL3I ,GEN2, MODEL4 I ,GEN2- The determined reference pressing force FPREF may be as an initial value for controlling the abrading.

[0151] Referring to Fig. 8e, also the reference value NRPMREF of rotation speed may be determined from the desired abrading result. The symbols NRPM1, NRPM2, NRPM3, NRPM4 denote different values of the rotation speed. The measured dust concentration DR and / or the measured electric current IC of a motor of the abrading apparatus MACi may be used as auxiliary data for controlling abrading the workpiece OBJ1 with the abrasive article ARTi.

[0152] Fig. 9a shows, byway of example, controlling the abrading without compensating the wear of the abrasive article. If the wear of the abrasive article is not compensated, then the dust concentration, electric current, and material removal rate may decrease with increasing degree of wear.

[0153] A reference value FPREF of the pressing force and a reference value NRPMREF of the rotation speed may be determined from a selected target result by using the MODEL3, MODEL4.

[0154] A workpiece OBJ1 may be abraded by pressing the abrasive article ART 1 with a pressing force FPsET(t), which may be kept equal to the reference value FPREF, and the rotation speed NRPMsET(t) may be kept equal to the reference value NRPMREF.

[0155] Fig. 9a shows, by way of example, temporal evolution of rotation speed, temporal evolution of pressing force, temporal evolution of measured electric current, temporal evolution of measured dust sensor signal, and temporal evolution of measured rate of removing material.

[0156] The uppermost curve of Fig. 9a shows the rotation speed NRPMsET(t) as a function of time t. The abrasive article ARTi may be brought into contact with the workpiece at the time to.

[0157] The second curve from the top of Fig. 9a shows the pressing force FPsET(t) as a function of time t.

[0158] The third curve from the top of Fig. 9a shows the measured electric current ICMEASG) as a function of time t. The measured electric current ICMEASG) may be indicative of a torque, which is transmitted via the abrasive article ARTi to the workpiece OBJ1. To the first approximation, the measured electric current ICMEAs(t) may be proportional to the torque, which is transmitted via the abrasive article ARTi to the workpiece OBJ1 . The measured electric current ICMEASG) may attain an initial value ICo at the time to. The measured electric current ICMEAs(t) may decrease after the time to, as the abrasive article ARTi wears and requires a smaller torque for causing the relative movement between the abrasive article ARTi and the workpiece OBJ1. The degree of wear of the abrasive article ARTi may be estimated e.g. from the ratio ICMEAs(t) / ICo.

[0159] The fourth curve from the top of Fig. 9a shows the measured dust concentration DRMEAS as a function of time t. The measured dust concentration DRMEAS may be indicative of the rate of removing material from the workpiece OBJ1 . To the first approximation, the measured dust concentration DRMEAS may be proportional to the rate of removing material from the workpiece OBJ1. The measured dust concentration DRMEAs(t) may attain an initial value DRo at the time to. The measured dust concentration DRMEAs(t) may decrease after the time to, as the abrasive article ARTi wears (or becomes clogged) and becomes less effective for removing material from the workpiece OBJ1 .

[0160] The lowermost curve of Fig. 9a shows the rate RMMEAs(t) of removing material from the workpiece OBJ1 , as a function of time t. The material removal rate RMMEAs(t) may have an initial value RMo at the time to. The measured material removal rate RMMEAs(t) may be measured e.g. by measuring the rate of change of the weight of the workpiece OBJ1. The measured material removal rate RMMEAs(t) may be measured e.g. by measuring the rate of change of the weight of dust, which is separated from the workpiece OBJ1 .

[0161] The abrasive article ARTi has an operating lifetime TLIFEI . The material removal rate may be equal to a value RMEND at the end of the operating lifetime TLIFEI . The operating lifetime TLIFEI at a constant pressing force FPREF and at a constant rotation speed NRPMREF may be defined e.g. so that the material removal rate RMMEAs(t) is equal to 33% of the initial material removal rate RMo at the end of the operating lifetime TLIFEI .

[0162] The material removal rate RMMEAs(t) may decrease during operation from the initial value RMo, in the situation where the pressing force and the rotation speed are kept at the constant values FPREF, NRPMREF. The material removal rate RMMEAs(t) at the times ti , t2, ts may be lower than the initial value RMo at the start time to. Keeping the pressing force and the rotation speed constant does not necessarily represent an optimum condition e.g. regarding the time needed for removing a certain amount of material from the workpiece, the amount of material removed from the workpiece during the operating lifetime of the abrasive article, and / or the rate of change of the gloss value of the workpiece.

[0163] Wear of the abrasive article ART 1 may be detected e.g. based on a change of the dust sensor signal DRMEAS and / or based on a change of the measured electric current ICMEAS of the motor of the abrading apparatus MAC1 , in a situation where the pressing force and the rotation speed are kept constant. Worn abrasive grains of the worn abrasive article ART 1 may be less efficient for removing material from the workpiece than the sharp abrasive grains of a new unused abrasive article ARTi, in a situation where the new abrasive article and the worn abrasive article are pressed with the same pressing force and rotated at the same rotation speed. The worn abrasive grains of the worn abrasive article ARTi may move more easily on the surface of the workpiece than the abrasive grains of the new unused abrasive article ARTi, in a situation where the new abrasive article and the worn abrasive article are pressed with the same pressing force and rotated at the same rotation speed.

[0164] Referring to Fig. 9b, the control system of the abrading apparatus MAC1 may be arranged to compensate wear of the abrasive article ARTi.

[0165] The control system of the abrading apparatus MAC1 may be arranged to adjust the pressing force FP and / or the rotation speed NRPM based on the measured dust sensor signal DRMEAS and / or based on the measured electric current ICMEAS. The control system may have a fast response to changes of the signal DRMEAS and the current ICMEAS. In particular, the control system may increase rotation speed NRPM and / or pressing force FP so as to keep the dust concentration, the mass removal rate and the electric current substantially constant. Keeping the electric current constant may mean that also the torque applied to the abrasive article ARTi is kept constant.

[0166] Abrading with the abrasive article ARTi may be started by using the reference values FPREF, NRPMREF as the initial values at the time to. The reference values FPREF, NRPMREF may be determined from the target RMI ,T by using the model MODEL3, MODEL4.

[0167] The control system of the abrading apparatus may be arranged to adjust the pressing force FP and / or the rotation speed NRPM, e.g. in order to keep the material removal rate RMMEAs(t) substantially constant. The control system of the abrading apparatus may be arranged to adjust the pressing force FP and / or the rotation speed NRPM based on the signal DRMEAS of the dust sensor RSEN1 and / or based on the electric current ICMEAS of the motor of the abrading apparatus, so as to keep the material removal rate RMMEAs(t) substantially constant.

[0168] To the first approximation, the electric current ICMEAS may be proportional to the torque needed to rotate the abrasive article ARTi against the workpiece OBJ1. To the first approximation, the torque needed to rotate the abrasive article ARTi may be proportional to the pressing force FP. To the first approximation, the electric current ICMEAS may be proportional to the pressing force FP. The electric current ICMEAS may have an initial value ICo at the start time to. The control system of the abrading apparatus may be arranged to adjust the pressing force FP, so as to keep the electric current ICMEAS substantially equal to the initial value ICo.

[0169] To the first approximation, the dust concentration signal DRMEAS may be proportional to the rotation speed NRPM and also proportional to the pressing force FP. To the first approximation, the dust signal DRMEAS may be proportional to the rotation speed NRPM and also proportional to the electric current ICMEAS. The dust signal DRMEAS may have an initial value DRo at the start time to. The control system of the abrading apparatus may be arranged to adjust the rotation speed NRPM, so as to keep the dust signal DRMEAS substantially equal to the initial value ICo.

[0170] The dust sensor RSEN1 may be e.g. an optical dust sensor or an acoustic dust sensor. The optical dust sensor may form the dust signal DRMEAS e.g. based on the intensity of light scattered by dust. The acoustic dust sensor may form the dust signal DRMEAS e.g. based on vibrations generated by dust particles, which hit the sensor. Fig. 9b shows, by way of example, temporal evolution of rotation speed, temporal evolution of pressing force, temporal evolution of measured electric current, temporal evolution of measured dust sensor signal, and temporal evolution of measured rate of removing material.

[0171] Referring to the uppermost curve of Fig. 9b, the control system may increase the rotation speed when the abrasive article wears, so as to keep the material removal rate substantially constant.

[0172] Referring to the second curve from the top of Fig. 9b, the control system may increase the pressing force when the abrasive article wears, so as to keep the torque substantially constant.

[0173] Referring to the third curve from the top of Fig. 9b, the electric current IC may be kept substantially equal to the initial value ICo during the abrading with the abrasive article ART 1.

[0174] Referring to the fourth curve from the top of Fig. 9b, the dust signal DRMEAS may be kept substantially equal to the initial value DRo during the abrading with the abrasive article ART 1.

[0175] Referring to the lowermost curve of Fig. 9b, the material removal rate RMMEAS may be kept substantially equal to the initial value RMo.The material removal rate RMMEAS may have an initial value RMo at the start time to.

[0176] The initial value RMo of the material removal rate may deviate from the desired target value RM I ,T. ARMto denotes the difference (RM I ,T - RMo) between the target value RMI ,T and the initial value RMo.

[0177] Referring to Fig. 9c, the model may be trained e.g. by using previously measured results, and the trained model may subsequently provide one or more updated reference values, which in turn reduce a difference between the target removal rate and the measured removal rate. The difference ARMto between a target value and a corresponding result may be reduced or minimized by using the model MODEL3, MODEL4. Fig . 9c shows, by way of example, temporal evolution of rotation speed, temporal evolution of pressing force, temporal evolution of measured electric current, temporal evolution of measured dust sensor signal, and temporal evolution of measured rate of removing material.

[0178] Referring to the uppermost curve of Fig. 9c, the control system may increase the rotation speed when the abrasive article wears, so as to keep the material removal rate substantially constant.

[0179] Referring to the second curve from the top of Fig. 9c, the control system may increase the pressing force when the abrasive article wears, so as to keep the torque substantially constant.

[0180] Referring to the third curve from the top of Fig. 9c, the electric current IC may be kept substantially equal to the initial value ICo during the abrading with the abrasive article ART 1.

[0181] Referring to the fourth curve from the top of Fig. 9c, the dust signal DRMEAS may be kept substantially equal to the initial value DRo during the abrading with the abrasive article ART 1.

[0182] Referring to the lowermost curve of Fig. 9c, the material removal rate RMMEAS may be kept substantially equal to the initial value RMo.

[0183] The method may comprise:

[0184] - providing a first abrasive article ARTi, which comprises a machine-readable code QRi,

[0185] - selecting a target result RMI ,T,

[0186] - retrieving a performance model MODEL3I ,GENI , MODEL4I ,GENI from a database DBM based on the code QRi,

[0187] - determining one or more usage parameters GMI.GENI from the target result RMI ,T by using the performance model MODEL3I ,GENI , MODEL4I ,GENI , and

[0188] - abrading with the first abrasive article ARTi according to the determined usage parameters GMI.GENI . The difference RMI .T - RMo between the target value RMI .T and the initial value RMo may be reduced or minimized by:

[0189] - determining the reference values FPREF, NRPMREF from the target value RMI .T by using the model MODEL3, MODEL4, and

[0190] - abrading the workpiece 0BJ1 with the abrasive article ARTi by using the determined reference values FPREF, NRPMREF as the initial values.

[0191] The model MODEL3, MODEL4 may also be trained during abrading with the abrasive article ART 1. The trained model MODEL3, MODEL4 may be utilized also during abrading with the abrasive article ARTi.

[0192] The method may comprise:

[0193] - measuring one or more abrading results RMI.GENI , RMMEAS, and

[0194] - controlling abrading based on the measured abrading results RMI.GENI, RMMEAS.

[0195] After abrading with the abrasive article ART 1 has been started at the start time to, the difference (RMI ,T - RMo) between the target value RMI ,T and the initial value RMo may optionally be further reduced or minimized by:

[0196] - measuring an abrading result RMMEAS,

[0197] - training the model MODEL3, MODEL4 by using the measured abrading result RMMEAS and other applicable parameters (e.g. target value RM I .T, pressing force FP, rotation speed NRPM) as the training data,

[0198] - determining updated reference values FPREF, NRPMREF from the target value RMI .T by using the trained model MODEL3GEN2, MODEL4GEN2,

[0199] - abrading the workpiece 0BJ1 (or another similar workpiece) with the abrasive article ARTi by using the updated reference values FPREF, NRPMREF as the updated initial values.

[0200] A proportionality constant between the signal DRMEAS of the dust sensor RSEN1 and the actual material removal rate RMMEASW may depend on various parameters, e.g. on the grit size of the abrasive article, on the material of the workpiece, on the air flow rate of the dust extractor, and / or on the efficiency of collecting dust. The proportionality constant may be unknown at the start time to, but may be determined after the time to e.g. by comparing the signal DRMEAS of the dust sensor RSEN1 with the measured material removal rate RMMEAs(t). The proportionality constant for calculating the material removal rate RMMEAS(I) from the dust concentration signal DRMEAS(I) may also be called as the calibration coefficient. The calibration coefficient may also be known. In that case the measured material removal rate RMMEAs(t) may be determined from the dust concentration signal DRMEAS by using the calibration coefficient, and it is not necessary to use e.g. a weighing device WSEN1 for measuring the material removal rate RMMEAs(t). Usage parameters may be determined from a desired value of the dust concentration signal DR, by using the model. The model may be subsequently trained by using the measured dust concentration signal DRMEAS.

[0201] Referring to Fig. 10, the abrading system GSYS1 may comprise:

[0202] - an abrading apparatus MAC1 to abrade a workpiece OBJ1 with an abrasive article ARTi,

[0203] - a reader READ1 to read a machine-readable code QRi of the abrasive article ARTi,

[0204] - a user interface UIF1 for selecting a target result RMI ,T,

[0205] - a database DMB,

[0206] - a control system CSYS1 .

[0207] The target result RMI ,T may be selected by using the user interface UIF1. The control system CSYS1 may be configured to retrieve a performance model MODEL3I ,GENI , MODEL4I ,GENI from the database DBM based on the code QRi.

[0208] The control system CSYS1 may be configured to:

[0209] - determine one or more usage parameters GMI.GENI from the target result RMI ,T by using the performance model MODEL3I ,GENI , MODEL4I ,GENI , and

[0210] - control abrading with the first abrasive article ARTi according to the determined usage parameters GMI.GENI .

[0211] The control system CSYS1 may comprise a control unit CNT1 , a communication unit RXTX1 , a user interface UIF1 , and sensors. The control system CSYS1 may comprise e.g. a force sensor GSEN1 , a current measuring sensor ISEN1 , a dust concentration sensor RSEN1 , and / or a weight sensor WSEN1. The control system CSYS1 may comprise a code reader READ1 for reading the code QRi of the abrasive article ARTi. One or more target results RMi may be selected by using the user interface UIF1.

[0212] The control system CSYS1 may comprise a memory MEM2 for storing the performance model MODEL3, MODEL4 associated with the code QRi of the abrasive article ARTi. The memory MEM2 may store e.g. the model MODEL3I ,GENI , MODEL3I ,GEN2, MODEL41.GEN1 , MODEL4I ,GEN2.

[0213] The control unit CNT1 may comprise one or more data processors. The control unit CNT1 may be configured to carry out the method by executing computer program code PRG1 . The control system CSYS1 may comprise a memory MEM3 for storing the computer program code PRG1 .

[0214] The control unit CNT1 may be configured to determine usage parameters GMI.GENI , GM I ,GEN2 from the one or more target results RMI ,T by using the performance model MODEL3, MODEL4. The control system CSYS1 may comprise a memory MEM1 for storing the usage parameters. The usage parameters may comprise e.g. a reference value of pressing force FPREF, and / or a reference value of rotation speed NRPM. The usage parameters may comprise e.g. an initial value of electric current ICo and / or an initial value of dust sensor signal DRo.

[0215] The abrading apparatus MAC1 may be located at an abrading site SITE1. The abrading apparatus MAC1 may comprise an abrading head HEAD1. The abrading head HEAD1 may comprise a motor MOTOR1. The motor MOTOR1 may be arranged to rotate and / or oscillate the abrasive article ARTi. The abrading head HEAD1 may comprise a fixing pad PAD1. The abrasive article ARTi may be attached to the fixing pad PAD1. The motor MOTOR1 may be arranged to rotate and / or oscillate the fixing pad PAD1 .

[0216] The motor MOTOR1 may be driven by a driving unit DRV1 . The motor MOTOR1 may be e.g. a brushless electric motor. The driving unit DRV1 may e.g. form one or more alternating currents IC for driving the motor MOTOR1 at a target rotation speed NRPMSET. The driving unit DRV1 may comprise a current sensor ISEN1 for measuring the magnitude ICMEAS of the one or more electric currents of the motor MOTOR1 . To the first approximation, the measured electric current ICMEAS may be proportional to the torque applied to the abrasive article ARTi.

[0217] The control system CSYS1 may be configured to adjust a pressing force FP of the abrasive article ARTi and / or to adjust a rotation speed NRPM of the motor MOTOR1 , so as to keep an electric current IC of the motor MOTOR1 of the abrading apparatus MAC1 substantially constant.

[0218] The control system CSYS1 may comprise a dust concentration sensor RSEN1 to measure a concentration of dust material particles D1 abraded from the workpiece OBJ1. The control system CSYS1 may be configured to adjust a pressing force FP of the abrasive article ARTi and / or to adjust a rotation speed NRPM of the motor MOTOR1 , so as to keep the signal DRMEAS of the dust concentration sensor RSEN1 substantially constant.

[0219] The abrading head HEAD1 may be e.g. an rotary abrading device. The motor MOTOR1 may comprise a shaft SHF1. The shaft SHF1 may be rotated by the motor MOTOR1 . The shaft SHF1 may have a first axis of rotation. The abrasive article ART1 may be attached to a fixing pad PAD1 . In case of the rotary abrading device, the axis of rotation of the pad PAD1 may coincide with the axis of rotation of the shaft SHF1.

[0220] The abrading head HEAD1 may also be e.g. an orbital abrading device. The shaft SHF1 may have a first axis of rotation. The pad PAD1 may have a second different axis of rotation. The second axis may be displaced with respect to the first axis. The pad PAD1 may be coupled to the shaft SHF1 eccentrically via a bearing BEAR1. The bearing BEAR1 may allow rotation of the pad PAD1 about the second rotation axis. The motor MOTOR1 may rotate the shaft SHF1 such that the second axis moves along a circular path about the first rotation axis, causing oscillation of the pad PAD1 and the abrasive article ARTi.

[0221] The abrasive article ARTi may be pressed against the workpiece OBJ1 by an actuator ROBO1 . The actuator ROBO1 may be arranged to control the pressing force of the abrasive article ARTi. The actuator ROBO1 may comprise e.g. a robot and / or a compliance device. The abrasive article ARTi may be pressed against the workpiece OBJ1 with a pressing force FP. The control unit CNT 1 may form a control signal Spp.sET for the actuator ROBO1 , so as to provide a target pressing force FPSET. The actual value of pressing force may temporarily deviate from the target value of the pressing force. The abrading head HEAD1 may comprise a force sensor GSEN1 for measuring the actual pressing force FP. The force sensor GSEN1 may provide a signal SFP.MEAS indicative of the measured pressing force FPMEAS. The control system CSYS1 may be arranged to control the actuator ROBO1 such that the measured pressing force FPMEAS is substantially equal to the target pressing force FPSET.

[0222] The abrasive article ART1 comprises a plurality of abrasive grains AG1. Oscillatory and / or rotational movement of the abrasive grains AG1 against the workpiece OBJ1 may separate and remove dust particles D1 from the workpiece. The apparatus MAC1 may be arranged to remove the dust particles D1 with an air flow. The abrasive article ART1 may comprise a plurality of openings OP1 for drawing dust-laden air DAIR1 via the openings OP1 to a dust extractor VAC1. The apparatus MACi may draw air and dust particles D1 from the abrasive article ART 1 via the openings OP1 , and via the suction duct DLIC1 to the dust extractor VAC1.

[0223] The abrading apparatus MAC1 may comprise the dust extractor VAC1 . The dust extractor VAC1 may comprise a fan FAN1 , which is rotated by a motor MOTOR2. The fan FAN1 may cause a partial vacuum p1 , which may draw dust laden air DAIR1 from the abrading head HEAD1 into the dust extractor VAC1 via a duct DLIC1 . The duct DLIC1 may be e.g. a flexible hose. The dust extractor VAC1 may comprise a dust separator FIL1. The dust separator FIL1 may separate dust particles D1 from the dust-laden air flow DAIR1. The dust separator FIL1 may comprise e.g. a particle filter and / or a cyclone. The fan FAN1 may draw the dustfree air flow AIR1 from the dust separator FIL1 to an outlet. The dust extractor VAC1 may provide a dust-free air flow AIR1 from an outlet. The dust extractor may comprise a container BIN1 for containing collected separated dust particles D1.

[0224] A weight sensor WSEN1 may be arranged to measure the weight of the collected dust particles D1. The weight sensor WSEN1 may be arranged to measure a change of weight of the collected dust particles D1 . The weight sensor WSEN1 may provide a measurement result RMMEAS for the control system CSYS1. The fan FAN1 may have an upstream pressure p1 and a downstream pressure p2. The dust extractor VAC1 may comprise a pressure sensor PSEN1 for measuring the pressures p1 , p2 and / or the pressure difference Ap=p2-p1 . The dust extractor VAC1 may provide a signal indicative of the rotation speed NRPMFAN of the fan FAN1.

[0225] The control system CSYS1 may form a control signal SFP.SET to the actuator ROBO1 . The control signal SFP.SET may be indicative of a target value of the pressing force. The control system CSYS1 may comprise a force sensor GSEN1 to provide a signal SFP.MEAS indicative of the measured value of the pressing force FP. The control system CSYS1 may form a control signal SNRPM to the driving unit DRV1 of the motor MOTOR1. The control signal SNRPM may be indicative of a target rotation speed of the motor MOTOR1 . The driving unit DRV1 may comprise a current sensor SEN1 . The current sensor SEN1 may form a signal SIC.MEAS indicative of the measured value of the electric current IC of the motor MOTOR1 . The dust sensor RSEN1 may form a signal SRSENI indicative of the measured dust concentration DRMEAS. The reader READ1 may form a signal SQRI indicative of the code QRi of the abrasive article ARTi. The weight sensor WSEN1 may form a signal SWSENI indicative of measured weight of collected material.

[0226] The pressure sensor PSEN1 may form a signal SPSENI indicative of a pressure difference Ap of the dust extractor VAC1 . The dust extractor VAC1 may form a signal SNRPM, FAN indicative of the rotation speed of the fan FAN1 of the dust extractor VAC1 . The control system CSYS1 may be arranged to calculate the air flow rate of the dust-laden air flow DAIR1 from the pressure difference Ap. The control system CSYS1 may be arranged to calculate the air flow rate of the dustladen airflow DAIR1 from the pressure difference Ap and from the rotation speed of the fan FAN1. The rate of removing material from the workpiece may be proportional to the dust sensor signal and to the air flow rate. The air flow rate may also be taken into consideration for determining a rate of removing material from the workpiece OBJ1 .

[0227] The control system CSYS1 may receive a model MODEL3I .GENI , MODEL4I ,GENI from the database DBM via the communication unit RXTX1 .The control system CSYS1 may send a trained model MODEL3I .GEN2, MODEL4I ,GEN2 to the database DBM via the communication unit RXTX1. The database DBM may send or receive data via a communication unit RXTX2. COM1 denotes data communication. The communication COM1 may take place e.g. via wired or wireless communication. The communication COM1 may take place e.g. via the Internet, via a mobile communications network, and / or via an industrial communication network.

[0228] The control system CSYS1 may be configured to:

[0229] - read the code QRi of the abrasive article ART1 at the abrading site SITE1 by using the reader READ1 ,

[0230] - retrieve the model MODEL3, MODEL4 from the database DBM based on the code QRi,

[0231] - receive one or more target results RMI ,T e.g. based on user input received via the user interface LIIF1 ,

[0232] - determine one or more reference values FPREF, NRPMREF of usage data from the one or more target results RM I ,T by using the retrieved model MODEL3, MODEL4,

[0233] - use the determined reference values FPREF, NRPMREF as target values FPSET, NRPMSET when abrading with the abrasive article ARTi,

[0234] - abrade the workpiece 0BJ1 with the abrasive article ARTi according to the target values FPSET, NRPMSET.

[0235] The control system CSYS1 may be configured to:

[0236] - measure the dust signal DRMEAS and / or electric current ICMEAS,

[0237] - adjust the pressing force FP and / or speed NRPM of the abrasive article based on the measured dust signal DRMEAS and / or electric current ICMEAS,

[0238] Usage parameters GMI.GENI determined by the performance model MODEL3I ,GENI , MODEL4I ,GENI may comprise e.g. a reference value FPREF of the pressing force of the first abrasive article ARTi, wherein the method may further comprise:

[0239] - setting the pressing force FP of the first abrasive article ART 1 according to the reference force value FPREF of the pressing force,

[0240] - abrading with the first abrasive article ARTi,

[0241] - measuring a torque parameter ICMEAS indicative of a torque of the first abrasive article ARTi, and - adjusting the pressing force FP of the first abrasive article ARTi based on the measured torque parameter ICMEAS, SO as to keep the measured torque parameter ICMEAS substantially constant during the abrading.

[0242] The speed of the abrasive article may be specified e.g. by the rotation speed of the abrasive article, by an oscillation frequency of the abrasive article, and / or by the rotation speed of the motor MOTOR1 .

[0243] Usage parameters GMI.GENI determined by the performance model MODEL3I,GENI , MODEL4I,GENI may comprise a reference speed value NRPMREF indicative of a speed of the first abrasive article ARTi, wherein the method may further comprise:

[0244] - setting the speed NRPM of the first abrasive article ARTi according to the reference speed value NRPMREF,

[0245] - abrading with the first abrasive article ARTi,

[0246] - measuring a concentration parameter DRMEAS indicative of a dust concentration generated by the abrading, and

[0247] - adjusting the speed NRPM of the first abrasive article ARTi based on the measured concentration parameter DRMEAS, SO as to keep the measured concentration parameter DRMEAS substantially constant during the abrading.

[0248] The performance model may be trained by using measured results. The trained performance model may be stored in the database DBM. Updated usage parameters may be determined by using the trained performance model. Abrading may be controlled according to the updated usage parameters.

[0249] The control system CSYS1 may be configured to:

[0250] - measure one or more abrading results RMMEAS, e.g. the measured material removal rate.

[0251] - train the model MODEL3, MODEL4 by using the measured abrading result RMMEAS,

[0252] - store the trained model MODEL3I,GEN2, MODEL4I,GEN2 in the database DBM and / or in a memory.

[0253] The abrasive article ARTi may comprise the machine-readable code QRi for identifying the abrasive article ARTi, or for identifying a manufacturing batch of mutually similar abrasive articles. The batch may comprise said abrasive article ARTi.

[0254] The machine-readable code QRi may be e.g. an optically readable matrix code, which may be implemented on the abrasive article ARTi at very low cost. The matrix code may be implemented on the abrasive article ARTi e.g. by printing. The matrix code may withstand the high acceleration (e.g. impacts and vibration) experienced by the abrasive article. The matrix code may be easily read e.g. by using a digital camera.

[0255] The machine-readable code QRi may also be implemented e.g. by using a radio frequency identification tag (RFID tag). The abrasive article ARTi may comprise an RFID tag, which comprises the machine-readable code QRi. The code may be read by using an RFID reader. The RFID tag may send the code QRi in response to an interrogation signal, which is transmitted from the reader. The RFID tag may comprise an antenna and an RFID chip. The RFID tag may be arranged to communicate e.g. by using the HF band, e.g. at the frequency 13.56 MHz. The RFID tag may be arranged to communicate e.g. by using the UHF band, e.g. at the frequency 433 MHz, 865-868 MHz, 902-928 MHz, or 2.4 GHz. The RFID tag may be arranged to communicate e.g. according to the standard EPC Gen2.

[0256] When using the RFID tag, the reader does not need to have direct line of sight to the RFID tag. When using the RFID tag, there may sometimes be a risk of reading the code of a wrong (second) article, in a situation where several different RFID tags are simultaneously within the reading range of the reader. When compared with RFID tag, the optically readable matrix code may allow a higher reliability for reading the code of the correct abrasive article.

[0257] In an embodiment, suitable usage parameters for using an abrasive article may have been already determined and stored in a memory of the control system CSYS1 . The control system CSYS1 may control abrading with abrasive articles by using the previously stored usage parameters, without reading the code of the abrasive article, without retrieving the model, and / or without using the model. In an embodiment, the control system CSYS1 may be arranged to keep the torque and / or dust concentration substantially constant also without using the model. Abrading with the abrasive article may be carried out by:

[0258] - providing a reference force value FPREF of the pressing force,

[0259] - setting the pressing force FP of an abrasive article ARTi according to the reference force value FPREF,

[0260] - abrading with the abrasive article ARTi,

[0261] - measuring a torque parameter ICMEAS indicative of a torque of the abrasive article ARTi, and

[0262] - adjusting the pressing force FP of the abrasive article ARTi based on the measured torque parameter ICMEAS, SO as to keep the measured torque parameter ICMEAS substantially constant during the abrading.

[0263] Abrading with the abrasive article may be carried out by:

[0264] - providing a reference speed value NRPMREF,

[0265] - setting the speed NRPM of an abrasive article ARTi according to the reference speed value NRPMREF,

[0266] - abrading with the abrasive article ARTi,

[0267] - measuring a concentration parameter DRMEAS indicative of a dust concentration generated by the abrading, and

[0268] - adjusting the speed NRPM of the abrasive article ARTi based on the measured concentration parameter DRMEAS, SO as to keep the measured concentration parameter DRMEAS substantially constant during the abrading.

[0269] The method may comprise determining one or more usage parameters GMI.GENI from the target result RM I ,T by using the performance model MODEL3I ,GENI , MODEL4 I ,GENI - The target result RM I ,T may specify e.g. a target material removal rate. Usage parameters GMI.GENI determined by using the performance model MODEL3I ,GENI , MODEL4I ,GENI may comprise e.g. a reference value FPREF of the pressing force of the first abrasive article ARTi and / or a reference speed value NRPMREF.

[0270] The pressing force FP of the first abrasive article ARTi may be controlled based on the measured torque parameter ICMEAS, e.g. by using a determined reference value FPREF of the pressing force as an initial value of the pressing force. The pressing force FP of the first abrasive article ARTi may be controlled based on the measured electric current ICMEAS of the motor MOTOR1 of the abrading apparatus MAC1 , e.g. by using a determined reference value FPREF of the pressing force as an initial value of the pressing force.

[0271] The usage parameters GMI.GENI determined by the performance model MODEL3I ,GENI , MODEL4I ,GENI may comprise a reference value FPREF of the pressing force of the first abrasive article ARTi, wherein the method may further comprise:

[0272] - setting the pressing force FP of the first abrasive article ART 1 according to the reference force value FPREF of the pressing force,

[0273] - abrading with the first abrasive article ARTi,

[0274] - measuring a torque parameter ICMEAS indicative of a torque of the first abrasive article ARTi, and

[0275] - controlling the pressing force FP of the first abrasive article ARTi based on the measured torque parameter ICMEAS.

[0276] The pressing force FP of the first abrasive article ARTi may be controlled based on the measured electric current ICMEAS of the motor MOTOR1 e.g. by using a determined reference value FPREF of the pressing force as an initial value of the pressing force.

[0277] The control system CSYS1 may be configured to adjust a pressing force FP of the abrasive article ARTi and / or to adjust a rotation speed of a motor MOTOR1 of the abrading apparatus MAC1 based on an electric current IC of the motor MOTOR1 of the abrading apparatus MAC1.

[0278] The speed NRPM of the first abrasive article ARTi may be controlled based on the measured dust concentration parameter DRMEAS, e.g. by using a determined speed value NRPMREF as an initial value of the speed.

[0279] The usage parameters GMI.GENI determined by the performance model MODEL3I ,GENI , MODEL4I ,GENI may comprise a reference speed value NRPMREF for the first abrasive article ARTi, wherein the method may further comprise:

[0280] - setting the speed NRPM of the first abrasive article ARTi according to the reference speed value NRPMREF, - abrading with the first abrasive article ARTi,

[0281] - measuring a dust concentration parameter DRMEAS indicative of a dust concentration generated by the abrading, and

[0282] - controlling the speed NRPM of the first abrasive article ARTi based on the measured dust concentration parameter DRMEAS.

[0283] The speed NRPM of the first abrasive article ARTi may be controlled based on the measured dust concentration DRMEAS, e.g. by using a determined speed value NRPMREF as an initial value of the speed.

[0284] The control system CSYS1 may be configured to adjust a pressing force FP of the abrasive article ARTi and / or to adjust a rotation speed NRPM of a motor MOTOR1 of the abrading apparatus (MAC1) based on a signal DRMEAS of the dust concentration sensor RSEN1 .

[0285] For the person skilled in the art, it will be clear that modifications and variations of the devices and methods according to the present invention are perceivable. The figures are schematic. The particular embodiments described above with reference to the accompanying drawings are illustrative only and not meant to limit the scope of the invention, which is defined by the appended claims.

Claims

CLAIMS1 . A method, comprising:- providing a first abrasive article (ARTi), which comprises a machine-readable code (QRi),- selecting a target result (RMI ,T),- retrieving a performance model (MODEL3I ,GENI , MODEL4I ,GENI ) from a database (DBM) based on the code (QRi),- determining one or more usage parameters (GMI .GENI ) from the target result (RMI .T) by using the performance model (MODEL3I ,GENI , MODEL4I ,GENI ), and- abrading with the first abrasive article (ARTi) according to the determined usage parameters (GMI .GENI ).

2. The method of claim 1 , further comprising:- measuring one or more abrading results (RMI.GENI , RMMEAS), and- controlling abrading based on the measured abrading results (RMI.GENI, RMMEAS).

3. The method of claim 1 or 2, wherein the usage parameters (GM I .GENI ) determined by the performance model (MODEL3I ,GENI , MODEL4I ,GENI ) comprise a reference value (FPREF) of the pressing force of the first abrasive article (ARTi), wherein the method further comprises:- setting the pressing force (FP) of the first abrasive article (ARTi) according to the reference force value (FPREF) of the pressing force,- abrading with the first abrasive article (ARTi),- measuring a torque parameter (ICMEAS) indicative of a torque of the first abrasive article (ARTi), and- controlling the pressing force (FP) of the first abrasive article (ARTi) based on the measured torque parameter (ICMEAS).

4. The method of claim 1 or 2, wherein the usage parameters (GM I .GENI ) determined by the performance model (MODEL3I ,GENI , MODEL4I ,GENI ) comprise a reference value (FPREF) of the pressing force of the first abrasive article (ARTi), wherein the method further comprises:- setting the pressing force (FP) of the first abrasive article (ARTi) according to the reference force value (FPREF) of the pressing force,- abrading with the first abrasive article (ARTi),- measuring a torque parameter (ICMEAS) indicative of a torque of the first abrasive article (ARTi), and- adjusting the pressing force (FP) of the first abrasive article (ARTi) based on the measured torque parameter (ICMEAS), SO as to keep the measured torque parameter (ICMEAS) substantially constant during the abrading.

5. The method according to any of the claims 1 to 4, wherein the usage parameters (GMI.GENI) determined by the performance model (MODEL3I ,GENI , MODEL4I ,GENI ) comprise a reference speed value (NRPMREF) for the first abrasive article (ARTi), wherein the method further comprises:- setting the speed (NRPM) of the first abrasive article (ARTi) according to the reference speed value (NRPMREF),- abrading with the first abrasive article (ARTi),- measuring a dust concentration parameter (DRMEAS) indicative of a dust concentration generated by the abrading, and- controlling the speed (NRPM) of the first abrasive article (ARTi) based on the measured dust concentration parameter (DRMEAS).

6. The method according to any of the claims 1 to 4, wherein the usage parameters (GMI.GENI) determined by the performance model (MODEL3I ,GENI , MODEL4I ,GENI ) comprise a reference speed value (NRPMREF) for the first abrasive article (ARTi), wherein the method further comprises:- setting the speed (NRPM) of the first abrasive article (ARTi) according to the reference speed value (NRPMREF),- abrading with the first abrasive article (ARTi),- measuring a dust concentration parameter (DRMEAS) indicative of a dust concentration generated by the abrading, and- adjusting the speed (NRPM) of the first abrasive article (ARTi) based on the measured dust concentration parameter (DRMEAS), SO as to keep the measured concentration parameter (DRMEAS) substantially constant during the abrading.

7. The method according to any of the claims 1 to 6, comprising:- measuring one or more abrading results (RMI.GENI , RMMEAS), and- forming a trained performance model (MODEL3I ,GEN2, MODEL4I ,GEN2) by using the one or more measured abrading results (RMI.GENI , RMMEAS).

8. The method of claim 7, comprising:- determining one or more updated usage parameters (GMI ,GEN2) from the target result (RMI .T) by using the trained performance model (MODEL3I ,GEN2, MODEL4I ,GEN2), and- abrading with an abrasive article (ARTi) according to the updated usage parameters (GMI ,GEN2).

9. The method of claim 7 or 8, comprising:- storing the trained performance model (MODEL3I ,GEN2, MODEL4I ,GEN2) in the database (DBM).

10. The method according to any of the claims 7 to 9, wherein the target result (RMI .T) specifies a target material removal rate, and a measured abrading result (RMMEAS) used for training the model (MODEL3I ,GENI , MODEL4I ,GENI ) specifies an actual measured material removal rate.

11. The method according to any of the claims 7 to 10, wherein the abrading results (RMI .GENI ) comprise one or more results selected from the following list: generated dust concentration (DR), amount of removed material, rate of removing material, change of rate of removing material, change of thickness of a workpiece (OBJ1 ), rate of change of thickness of a workpiece (OBJ1 ), gloss value, rate of change of gloss value.

12. The method according to any of the claims 1 to 11 , wherein the first abrasive article (ARTi) is attached to an abrading apparatus (MAC1 ), which comprises at least one sensor (RSEN1 ) to measure one or more of the following: generated dust concentration (DR), amount of removed material, rate of removing material, change of rate of removing material, change of thickness of a workpiece (OBJ1 ), rate of change of thickness of a workpiece (OBJ1 ), gloss value, rate of change of gloss value.

13. The method according to any of the claims 1 to 12, wherein the usage parameters (GM I .GENI ) are selected from the following list: pressing force (FP), abrading pressure, torque, operating current (IC), rotation speed (NRPM) of the abrasive article (ARTi), oscillation frequency of the abrasive article (ARTi),amplitude of oscillation of the abrasive article (ARTi), relative speed of the abrasive article (ARTi) with respect to a workpiece (OBJ1 ), acoustic sound level caused by the abrading, rate of change of weight of abrasive article (ARTi).

14. The method according to any of the claims 1 to 13, wherein the first abrasive article (ARTi) is attached to an abrading apparatus (MAC1 ), which comprises at least one sensor (RSEN1 ) arranged to measure one or more of the following: pressing force (FP), abrading pressure, torque, operating current (IC), rotation speed (NRPM) of the abrasive article (ARTi), oscillation frequency of the abrasive article (ARTi), amplitude of oscillation of the abrasive article (ARTi), relative speed of the abrasive article (ARTi) with respect to a workpiece (OBJ1 ), acoustic sound level caused by the abrading, rate of change of weight of abrasive article (ARTi).

15. The method according to any of the claims 1 to 14, wherein the model (MODEL3I ,GENI , MODEL4I ,GEN I ) comprises a regression function, a neural network and / or a decision tree.

16. The method according to any of the claims 1 to 15, comprising:- determining a reference torque value (ICREF) indicative of the torque of the first abrasive article (ARTi),- abrading with the first abrasive article (ARTi),- measuring a torque parameter (ICMEAS) indicative of the torque of the first abrasive article (ARTi), and- starting operation in a fail-safe mode if the difference between the measured torque parameter (ICMEAS) and the reference torque value (ICREF) is greater than a predetermined limit.

17. The method according to any of the claims 1 to 16, comprising:- determining a reference concentration value (DRREF) indicative of a dust concentration generated by abrading,- abrading with the first abrasive article (ARTi),- measuring a concentration parameter (DRMEAS) indicative of a dust concentration generated by the abrading, and- starting operation in a fail-safe mode if the difference between the measured concentration parameter (DRMEAS) and the reference concentration value (DRREF) is greater than a predetermined limit.

18. An abrading system (GSYS1 ), comprising:- an abrading apparatus (MAC1 ) to abrade a workpiece (OBJ1 ) with an abrasive article (ARTi),- a reader (READ1 ) to read a machine-readable code (QRi) of the abrasive article (ARTi),- a user interface (LIIF1 ) for selecting a target result (RMI ,T),- a database (DMB),- a control system (CSYS1 ) to retrieve a performance model (MODEL3I ,GENI , MODEL4I ,GENI ) from the database (DBM) based on the code (QRi), wherein the control system (CSYS1 ) is configured to:- determine one or more usage parameters (GM I .GENI ) from the target result (RMI .T) by using the performance model (MODEL3I ,GENI , MODEL4I ,GENI ), and- control abrading with the first abrasive article (ARTi) according to the determined usage parameters (GM I .GENI ).

19. The abrading system (GSYS1 ) of claim 18, wherein the control system (CSYS1 ) is configured to adjust a pressing force (FP) of the abrasive article (ARTi) and / or to adjust a rotation speed (NRPM) of a motor (MOTOR1 ) of the abrading apparatus (MAC1 ) based on an electric current (IC) of a motor (MOTOR1 ) of the abrading apparatus (MAC1 ).

20. The abrading system (GSYS1 ) of claim 18 or 19, wherein the control system (CSYS1 ) is configured to adjust a pressing force (FP) of the abrasive article (ARTi) and / or to adjust a rotation speed (NRPM) of a motor (MOTOR1 ) of the abrading apparatus (MAC1 ), so as to keep an electric current (IC) of a motor (MOTOR1 ) of the abrading apparatus (MAC1 ) substantially constant.

21. The abrading system (GSYS1 ) according to any of the claims 18 to 20, comprising a dust concentration sensor (RSEN1 ) to measure a concentration of dust material particles (D1 ) abraded from the workpiece (OBJ1 ), wherein the control system (CSYS1 ) is configured to adjust a pressing force (FP) of the abrasive article (ARTi) and / or to adjust a rotation speed (NRPM) of a motor(M0T0R1) of the abrading apparatus (MAC1 ) based on a signal (DRMEAS) of the dust concentration sensor (RSEN1 ).

22. The abrading system (GSYS1) according to any of the claims 18 to 20, comprising a dust concentration sensor (RSEN1 ) to measure a concentration of dust material particles (D1 ) abraded from the workpiece (OBJ1 ), wherein the control system (CSYS1 ) is configured to adjust a pressing force (FP) of the abrasive article (ARTi) and / or to adjust a rotation speed (NRPM) of a motor (MOTOR1) of the abrading apparatus (MAC1), so as to keep a signal (DRMEAS) of the dust concentration sensor (RSEN1 ) substantially constant.

Images