Reducing friction in machine tools

Abstract

The invention relates to a computer-implemented method for reducing friction within a machine tool (MT), comprising the method steps: a) reading (S1) a plurality of surrogate models (SM) for approximating a given friction compensation within a machine tool, wherein each surrogate model (SM1,..., SMm) is configured such that it assigns a friction compensation result value to a given set of friction compensation parameters for reducing friction within a machine tool, and wherein a weighting factor (w1,..., wm) is assigned to each surrogate model (SM1,..., SMm), b) reading (S2) a set of friction compensation parameters (CP), c) determining (S3) a friction compensation result value (CPR1,..., CPrm) for each surrogate model (SM1,..., SMm) using said set of compensation parameters (CP), d) determining (S4) a weighted average friction compensation value (CPRav) of said friction compensation result values using the respective weighting factors (w1,..., wm) of the respective surrogate models (SM1,..., SMm), e) deriving (S5) a quality indicator (Q) for said set of friction compensation parameters (CP) based on said weighted average friction compensation value (CPRav), f) outputting (S6) said set of friction compensation parameters (CPopt) if said quality indicator (Q) fulfils a given quality criterion (QC), or repeating (S7) steps b) to e) until said quality indicator fulfils said given quality criterion, g) applying (S8) the outputted set of friction compensation parameters (CPopt) to said machine tool for reducing friction within said machine tool.

Description

Technical Field

[0001] This invention relates to a computer-implemented method and apparatus for reducing friction within machine tools. Background Technology

[0002] Computer numerical control (CNC) machines are machine tools that can automatically produce workpieces with high precision, even for complex shapes. These machine tools enable high-precision manufacturing in industry. However, friction within the machine tool—that is, friction between mechanical components—can introduce deviations between the machine tool's controlled position and its actual execution position. This deviation caused by friction affects the target tolerances of the produced parts.

[0003] Therefore, it is necessary to reduce friction between the mechanical parts of machine tools. Friction compensation controllers with adjusted parameters correct for errors caused by reverse forces introduced during manufacturing. Due to variations in force and the friction generated by different machine parts, controllers must be individually configured for each machine. Furthermore, these parameters must be recalibrated during the machine's service life.

[0004] Traditionally, these parameters are manually adjusted by skilled technicians, requiring production to be interrupted and machines to be shut down. Furthermore, irregular adjustments among technicians and inconsistent quality can lead to tolerance losses and consequently reduce the quality of finished parts. Summary of the Invention

[0005] Therefore, the purpose of this invention is to improve friction compensation within machine tools.

[0006] The objective of this invention is achieved through the features of the independent claims. The dependent claims contain further improvements to the invention.

[0007] According to a first aspect, the present invention provides a computer-implemented method for reducing friction within a machine tool, comprising the following steps:

[0008] a) Read multiple alternative models for approximating friction compensation within a given machine tool, wherein each alternative model is configured to assign friction compensation result values ​​to a given set of friction compensation parameters to reduce friction within the machine tool, and wherein a weighting factor is assigned to each alternative model, the weighting factor representing the goodness of fit of the alternative model to the machine tool.

[0009] b) Read the friction compensation parameter set.

[0010] c) Use the aforementioned set of compensation parameters to determine the friction compensation result value for each alternative model.

[0011] d) Determine the weighted average friction compensation value using the corresponding weighting factors of the appropriate alternative model.

[0012] e) Determine the quality indicator of the friction compensation parameter set based on the weighted average friction compensation value.

[0013] f) If the quality indicator meets the given quality criterion, output the friction compensation parameter set; otherwise, repeat steps b) to e) until the quality indicator meets the given quality criterion.

[0014] g) Apply the output friction compensation parameter set to the machine tool to reduce friction within the machine tool.

[0015] Unless otherwise specified, the terms “operation,” “execution,” “computer-implemented,” “calculate,” “determine,” “generate,” “configure,” “reconfigure,” etc., are preferably associated with actions and / or processes and / or steps that alter and / or generate data, wherein the data may specifically be physical data and may be executed by a computer or processor. The term “computer” can be interpreted broadly and can refer to a personal computer, server, mobile computing device, or processor such as a central processing unit (CPU) or microprocessor.

[0016] The machine tool can be, for example, a computer numerical control (CNC) machine tool. An alternative model is preferably a computerized model configured to approximate or fit the friction compensation behavior within the machine tool. Within the machine tool, friction compensation (i.e., applying opposing or balancing forces to reduce friction, for example) depends on the set of friction compensation parameters applied to control the machine tool.

[0017] The alternative model can be, for example, a fitting model, a regression model, or an artificial neural network. Preferably, the alternative model is suitable for representing physical behavior, i.e., friction between machine parts of a machine tool due to internal forces. The friction compensation parameter set can be understood as the input values ​​of the alternative model. Furthermore, the friction compensation parameter set is input to configure the machine tool to reduce internal friction. The weighting factor preferably represents the likelihood that the corresponding alternative model correctly reproduces the friction compensation response within the machine tool. In other words, the weighting factor represents the goodness of fit of the corresponding alternative model to the friction compensation response of the machine tool.

[0018] The proposed method has the advantage of automatically finding an optimized set of friction compensation parameters for setting up a machine tool, thereby reducing internal friction within the machine tool. Furthermore, due to its computational speed, the method can be applied in-situ (i.e., in parallel with machine tool operation).

[0019] The friction compensation parameter set is used to configure the machine tool so that, for example, a counterforce is applied in a way that reduces friction between machine parts. Less calibration work is required to calibrate the machine tool and / or better calibration results can be obtained. Furthermore, this invention is particularly applicable to unknown machines.

[0020] This invention can use multiple alternative models of friction forces similar to those of machine tools to determine an optimized set of parameters for friction compensation within the machine tool.

[0021] In a preferred embodiment of the computer-implemented method according to the first aspect of the invention, the actual friction compensation result value of the machine tool can be measured based on the applied friction compensation parameter set, each weighting factor of the corresponding alternative model can be modified according to the difference between each friction compensation result value obtained from the corresponding alternative model and the measured actual friction compensation result value, and steps b) to g) above can be repeated.

[0022] By adjusting the weighting factors of the alternative model based on friction compensation results measured at the actual machine tool, the approximation of friction compensation for that machine tool can be optimized. Preferably, the adjustment of the weighting factors of the alternative model and the application of friction compensation parameters are performed iteratively until a given stopping criterion is met. This stopping criterion could be, for example, a specific friction compensation quality, an expired adjustment period, or the discovery that the selected parameter set is optimally suited to the actual machine tool.

[0023] In another embodiment of the computer-implemented method according to the first aspect of the invention, multiple alternative models can be generated by means of a regression method based on a given set of multiple datasets, wherein each dataset includes a set of friction compensation parameters and corresponding friction compensation result values ​​for the reduced friction obtained in the respective machine tool.

[0024] Preferably, this dataset used to generate alternative models of the machine tools is stored in a database. Based on the available data, alternative models can be trained individually for each dataset using regression methods. The learning objective of the alternative models is to estimate the friction compensation results for a given set of parameters. Therefore, the provided datasets can be used as training data for training these alternative models. Possible regression techniques include, for example, linear and multinomial models, regression trees, artificial neural networks, or Gaussian processes.

[0025] In one embodiment of the computer-implemented method according to the first aspect of the invention, a dataset can be generated based on friction measurements at a real machine tool.

[0026] In one embodiment of the computer-implemented method according to the first aspect of the invention, a dataset can be generated based on a dedicated computer-aided simulation of a machine tool.

[0027] The dataset can be stored in a database. It provides an alternative model for friction compensation within machine tools.

[0028] In one embodiment of the computer-implemented method according to the first aspect of the invention, the friction compensation parameter set can be generated by means of a goodness-of-fit function, wherein the goodness-of-fit function depends on the alternative model and the corresponding weighting factor.

[0029] The friction compensation parameter set can be determined using a computerized search of the parameter space based on a goodness-of-fit function. The goodness-of-fit function preferably uses an alternative model and corresponding weighting factors to calculate a scalar value, which serves as an indicator of the quality of the parameter set used.

[0030] In one embodiment of the computer-implemented method according to the first aspect of the invention, multiple alternative models can be selected based on machine-specific identification data of the machine tool.

[0031] Preferably, the machine-specific identification data of the machine tool includes manufacturing information data and / or machine type data. The optimization process can be further improved by using prior knowledge of the machine tool. Preferably, before starting the on-site optimization process for the selected machine tool, multiple alternative models for approximating the friction compensation response of the machine tool are pre-selected. Therefore, additional alternative models that are not very suitable for approximating the machine tool can be excluded.

[0032] According to a second aspect, the present invention relates to a device for reducing friction within a machine tool, comprising:

[0033] a) An input unit configured to read multiple alternative models for approximating friction compensation within a given machine tool, wherein each alternative model is configured to assign friction compensation result values ​​to a given set of friction compensation parameters for reducing friction within the machine tool, and wherein a weighting factor is assigned to each alternative model, the weighting factor representing the goodness of fit of the alternative model to the machine tool.

[0034] b) The analysis unit is configured as follows:

[0035] -Read in the friction compensation parameter set,

[0036] - Use the aforementioned set of compensation parameters to determine the friction compensation result value for each alternative model.

[0037] - The weighted average friction compensation value is determined using the corresponding weighting factors of the appropriate alternative model, and

[0038] -A quality indicator for determining the friction compensation parameter set based on the weighted average friction compensation value.

[0039] c) An output unit configured to output a set of friction compensation parameters if the quality indicator meets a given quality criterion, or otherwise repeat the steps performed by the analysis unit.

[0040] d) An application unit configured to apply the output set of friction compensation parameters to the machine tool to reduce friction within the machine tool.

[0041] The device is preferably connected to or is part of a machine tool. The device and / or at least one unit thereof may also include at least one processor or computer to perform the method steps according to the invention. The corresponding unit may be implemented in hardware and / or software. If the unit is implemented in hardware, it may be embodied as a device, such as a computer or processor or part of a system. If the unit is implemented in software, it may be embodied as a computer program product, function, routine, program code, or executable object. The output unit preferably provides a data structure including optimized compensation parameters. This data structure may, for example, be transmitted to the machine tool's control unit for corresponding machine tool settings.

[0042] According to one embodiment of the apparatus, the application unit can also be configured to receive measured true friction compensation result values ​​of the machine tool based on the applied friction compensation parameter set, and the analysis unit is configured to modify each weighting factor of the corresponding alternative model according to the difference between each friction compensation result value and the measured true friction compensation result value and obtained from the corresponding alternative model, and repeat steps b) to e) of the computer-implemented method according to the first aspect of the invention.

[0043] According to one embodiment, the apparatus may include a generator configured to generate multiple alternative models for friction compensation based on a given set of multiple datasets using a regression method, wherein each dataset includes friction compensation parameters for setting the machine tool and the corresponding reduction in friction compensation obtained in the respective machine tool.

[0044] According to another embodiment, the device may be connected to a database, wherein the database is configured to store datasets and / or alternative models.

[0045] The invention also includes a computer program product that can be directly loaded into the internal memory of a digital computer, including software code portions for performing the steps of the method when the product is run on the computer.

[0046] Computer program products, such as computer program devices, can be embodied in memory cards, USB sticks, CD-ROMs, DVDs, or files that can be downloaded from servers on a network. Attached Figure Description

[0047] The invention will be explained in more detail with reference to the accompanying drawings.

[0048] Figure 1A flowchart illustrating method steps involved in one embodiment of a method for reducing friction within a machine tool is shown;

[0049] Figure 2 A schematic representation of an embodiment of a method for reducing friction within a machine tool; and

[0050] Figure 3 A schematic representation of one embodiment of a device for reducing friction within a machine tool is shown.

[0051] Equivalent components in different figures are indicated by the same reference numerals. Detailed Implementation

[0052] Figure 1 A flowchart illustrating the method steps involved in a computer-implemented method for reducing friction within a machine tool, preferably a CNC machine tool, is shown. The method preferably provides an optimized set of parameters that can be applied to the machine tool to achieve optimal friction compensation. Furthermore, the method allows for the determination of optimized alternative models to approximate friction compensation within the machine tool. The machine tool can be, for example, a CNC machine tool used for milling, laser cutting, stamping, or other industrial applications.

[0053] The first step S0 of the method involves generating multiple alternative models for different machine tools based on training data using a regression method. The training data includes datasets. Each dataset includes a set of friction compensation parameters, also called a parameter set, for setting the machine tool, and corresponding friction compensation result values. Friction compensation result values ​​are generated by applying the parameter set to the machine tool and measuring the resulting frictional force. Therefore, the friction compensation result values ​​can be understood as an indicator of the resulting frictional force within the machine tool; that is, they can be determined, for example, based on sensor measurements of friction between machine parts.

[0054] The dataset can be generated based on friction measurements at real machine tools or different machine tools and / or based on dedicated computer-aided simulations of at least one machine tool. Corresponding alternative models for the machine tools are generated based on at least one dataset. To generate the alternative models, computer regression methods can be used, such as linear or multinomial models, regression trees, artificial neural networks, or Gaussian processes. Preferably, alternative models are generated for multiple different machine tools. The generated alternative models are preferably stored in a database.

[0055] In the next step S1, multiple alternative models for approximating friction compensation within the machine tool are read in. Preferably, a sample of alternative models is selected from available alternative models stored in a database based on machine-specific identification data of the machine tool, such as machine type. A weighting factor is assigned to each alternative model, wherein the weighting factor preferably represents the goodness of fit of the corresponding alternative model to the friction compensation response of the real machine tool. At the start of the optimization process, the weighting factors of each alternative model can be particularly evenly distributed, for example, all set to 1.

[0056] In the next step S2, the friction compensation parameter set is read in. Preferably, the friction compensation parameter set is generated based on a given weighting criterion, which will be explained below. Typically, the friction compensation parameter set preferably includes at least one parameter controlling the machine tool, which is also input for the corresponding alternative model. The friction compensation parameter set can be proposed based on the evaluation of the alternative model, as described below. The initially proposed parameter set can be, for example, an initial estimate.

[0057] In the next step S3, based on the input friction compensation parameter set, a friction compensation result value is determined for each input alternative model. In other words, based on this parameter set, each alternative model is evaluated to provide a friction compensation result value.

[0058] In the next step S4, the weighted average value is determined based on the compensation result value and the weighting factor of the corresponding alternative model.

[0059] In the next step S5, a quality indicator for the set of friction compensation parameters used is determined based on the weighted average compensation result value. The quality indicator represents the quality of the proposed set of parameters used to reduce friction when applied to a machine tool. The quality indicator may, for example, have a value corresponding to the weighted average compensation result value and / or multiplied by a given factor or similar value.

[0060] If the quality indicator meets a given quality standard, such as exceeding a given threshold, a friction compensation parameter set is output (step S6), and the friction compensation parameter set is applied to the machine tool (step S8) to configure the machine tool in a way that reduces internal friction. The friction compensation parameter set can, for example, be transmitted to a machine control unit used to control the machine tool in order to reduce friction between the machine tool's mechanical components.

[0061] If the quality indicator does not meet the given quality criteria, in step S7, a second friction compensation parameter set different from the first input parameter set is selected and input. Using this second parameter set, a second friction compensation result value for the alternative model is determined. Preferably, the weighting factor of the alternative model is not modified. The weighted average of the obtained second friction compensation result values ​​is determined to derive the quality indicator of the second parameter set. If the quality indicator of the second parameter set meets the given quality criteria, the second friction compensation parameter set is output. If it does not meet the quality criteria, the search for a suitable parameter set is repeated. Therefore, a suitable parameter set is searched based on this iterative process. In particular, this parameter search can be implemented using a goodness-of-fit function that uses the alternative model and its corresponding weighting factor.

[0062] Step S8 involves applying a set of friction compensation parameters that meet quality standards to the machine tool, and step S9 involves measuring the actual friction compensation result value based on the applied parameter set. In the next step S10, the weighting factor of the alternative model can be modified based on the difference between each friction compensation result value output by each alternative model and the measured actual friction compensation result. For example, a small difference between a measured friction compensation result and a modeled friction compensation result can be converted into a higher weight for the corresponding alternative model. Based on the modified weighting factor of the alternative model, steps S2 to S8 can be repeated, preferably up to S10, to further improve the modeling and parameter determination of the machine tool.

[0063] Figure 2 A schematic representation of an embodiment of a method for reducing friction within a machine tool MT is shown. This representation includes the generation of alternative models using a generator 103. The model generator 103 preferably includes or is connected to a database DB. The database DB preferably includes training data DATA, which is used to generate alternative models suitable for approximating and reproducing frictional forces within the machine tool. The training data DATA is based on measurement result data and / or simulation data. The training data DATA includes multiple datasets, each consisting of a friction compensation parameter set CP' and a corresponding friction compensation result value CPR'. Using the training data DATA, the generator 103 can generate multiple alternative models SM1, ..., SMn using a regression method RM. The alternative models SM1, ..., SMn can be stored in the database DB.

[0064] Samples of at least one alternative model SM1, ..., SMm are selected from these multiple alternative models SM1, ..., SMn. This selection is preferably based on machine-specific identification data of the machine tool MT. Weighting factors w1, ..., wm are assigned to each alternative model SM1, ..., SMm. The weighting factors w1, ..., wm represent the goodness of fit of the corresponding alternative model approximating the frictional response of the machine tool.

[0065] The selected alternative models SM1, ..., SMm are read in by the analysis unit 102. Additionally, a friction compensation parameter set CP is read in by the analysis unit 102. Preferably, the friction compensation parameter set CP is determined using a fit function based on the selected alternative models SM1, ..., SMm and their respective weighting factors w1, ..., wm.

[0066] For each alternative model SM1, ..., SMm, the corresponding friction compensation result values ​​CPR1, ..., CPRm are determined based on the input parameter set CP. Using the corresponding weighting factors w1, ..., wm, the weighted average CPRav of these friction compensation result values ​​CPR1, ..., CPRm is calculated. A quality indicator Q is derived from this weighted average CPRav to determine the matching quality of the friction compensation parameter set CP. If the quality indicator meets the given quality criterion QC, the friction compensation parameters CP are output as the optimized friction compensation parameters CPopt and applied to the machine tool MT. Otherwise, another parameter set can be proposed and evaluated until the parameter set meets the quality criterion QC.

[0067] At the machine tool MT, the actual friction compensation result value CPR_MT can be measured. For example, a sensor can be used to measure the friction force between two machine parts. By comparing this measured friction compensation result value CPR_MT with the individual friction compensation result values ​​CPR1, ..., CPRm output by the alternative models SM1, ..., SMm, the weighting factors w1, ..., wm of these alternative models can be adjusted. In other words, the weighting factors of the corresponding alternative models are modified based on the fitting quality of the corresponding models. Preferably, the alternative models SM1, ..., SMm that predict compensation result values ​​close to the measured values ​​are given greater weight. Performing these iterative steps further improves the weighting of the alternative models and the parameter set search, resulting in reduced friction within the machine tool. Once a given stopping criterion is met, the iterative parameter search and / or model weighting can be stopped.

[0068] Figure 3 A schematic representation of one embodiment of a device 100 for reducing friction within a machine tool MT is shown. The device 100 is preferably connected to the machine tool MT using a wireless or wired connection.

[0069] The device includes an input unit 101 configured to read multiple alternative models for approximating friction compensation within a given machine tool. Each alternative model is configured to assign friction compensation result values ​​to a given set of friction compensation parameters to reduce friction within the machine tool. Weighting factors are assigned to each alternative model.

[0070] The apparatus 100 also includes an analysis unit 102 configured to read in a set of friction compensation parameters and use the set of compensation parameters to determine the friction compensation result value for each alternative model. The analysis unit 102 is further configured to use the corresponding weighting factor of the corresponding alternative model to determine a weighted average friction compensation value for the friction compensation result value, and to derive a quality indicator for the friction compensation parameter set based on the weighted average friction compensation value. The apparatus 100 also includes an output unit 103 configured to output the set of friction compensation parameters if the quality indicator meets a given quality criterion, and an application unit 104 configured to apply the output set of friction compensation parameters to a machine tool to reduce friction within the machine tool.

[0071] The application unit 104 can also be configured to receive measured friction compensation result values ​​from the machine tool based on the applied friction compensation parameter set. Measurements can be performed, for example, using sensors located at or inside the machine tool. The analysis unit 102 can be configured to modify each weighting factor of the corresponding alternative model based on the differences between each friction compensation result value and the measured true friction compensation result value, as well as those obtained from the corresponding alternative model, and repeat the parameter set selection step.

[0072] The apparatus 100 may also include a generator 105 configured to generate multiple alternative models for friction compensation based on a given set of multiple datasets using a regression method. Each dataset includes friction compensation parameters for setting the machine tool and corresponding friction compensation results for the resulting reduction in friction within the machine tool. Alternatively, the generator 105 may be installed separately and connected to the apparatus 100.

[0073] The device 100 and / or generator 105 may also be connected to a database DB, wherein the database is configured to store alternative models and / or friction compensation data to generate alternative models for approximating friction compensation in a machine tool.

[0074] Although the invention has been described in detail with reference to preferred embodiments, it should be understood that the invention is not limited to the disclosed examples, and many additional modifications and changes can be made to it by those skilled in the art without departing from the scope of the invention.

Claims

1. A computer-implemented method for reducing friction within a machine tool (MT), comprising the method steps of: a) reading (S1) a plurality of alternative models (SM) for approximating a friction compensation within a given machine tool, wherein, each surrogate model (SM1,..., SMm) is configured such that it assigns a friction compensation result value to a given set of friction compensation parameters for reducing friction within a machine tool, and wherein a weighting factor (wi,..., wm) is assigned to each surrogate model (SM1,..., SMm), said weighting factor representing a goodness-of-fit of the surrogate model to the machine tool, b) reading (S2) a set of friction compensation parameters (CP), c) determining (S3) a friction compensation result value (CPR1,..., CPRm) for each surrogate model (SM1,..., SMm) using said set of compensation parameters (CP), d) determining (S4) a weighted average friction compensation value (CPRav) of said friction compensation result values using the respective weighting factors (wi,..., wm) of the respective surrogate models (SM1,..., SMm), e) determining (S5) a quality indicator (Q) for said set of friction compensation parameters (CP) based on said weighted average friction compensation value (CPRav), f) outputting (S6) said set of friction compensation parameters (CPopt) if said quality indicator (Q) fulfills a given quality criterion (QC), or else repeating (S7) steps b) to e) until said quality indicator fulfills said given quality criterion, g) applying (S8) the output set of friction compensation parameters (CPopt) to said machine tool for reducing friction within said machine tool.

2. The computer-implemented method according to claim 1, further comprising the steps of: - measuring (S9) a real friction compensation result value (CPR_MT) of said machine tool based on the applied set of friction compensation parameters (CPopt), - modifying (S10) each weighting factor (wi,..., wm) of the respective surrogate models (SM1,..., SMm) depending on a difference between the measured real friction compensation result value and each friction compensation result value (CPR1,..., CPRm) produced by the respective surrogate model, and - repeating steps b) to g) of claim 1.

3. The computer-implemented method of claim 1 or 2, wherein, said plurality of surrogate models (SM1,..., SMn) is generated (S0) based on a given plurality of data sets (DATA) by means of a regression method (RM), wherein each data set (DATA) comprises a set of friction compensation parameters and a corresponding friction compensation result value of a produced reduced friction within a machine tool.

4. The computer-implemented method of claim 3, wherein, said data sets (DATA) are generated based on friction measurements at real machine tools.

5. The computer-implemented method of claim 3, wherein, said data sets (DATA) are generated based on dedicated computer-aided simulations of machine tools.

6. The computer-implemented method of claim 1 or 2, wherein, said set of friction compensation parameters (CP) is generated by means of a goodness-of-fit function, wherein said goodness-of-fit function depends on said surrogate models (SM1,..., SMm) and the respective weighting factors (wi,..., wm).

7. The computer-implemented method of claim 1 or 2, wherein, said plurality of surrogate models (SM) is selected based on machine-specific identification data of said machine tool.

8. An apparatus (100) for reducing friction within a machine tool (MT), comprising: a) an input unit (101) configured to read a plurality of surrogate models for approximating a friction compensation within a given machine tool, wherein each surrogate model is configured such that it assigns a friction compensation result value to a given set of friction compensation parameters for reducing friction within a machine tool, and wherein a weighting factor is assigned to each surrogate model, the weighting factor representing a goodness-of-fit of the surrogate model to the machine tool, b) an analysis unit (102) configured to: - read in a set of friction compensation parameters, - determine a friction compensation result value for each surrogate model using the set of compensation parameters, - determine a weighted average friction compensation value of the friction compensation result values using the respective weighting factor of the respective surrogate model, and - determine a quality indicator for the set of friction compensation parameters based on the weighted average friction compensation value, c) an output unit (103) configured to output the set of friction compensation parameters if the quality indicator fulfills a given quality criterion, or to repeat the steps the analysis unit (102) is configured to perform otherwise, and d) an application unit (104) configured to apply the output set of friction compensation parameters to the machine tool for reducing friction within the machine tool.

9. The apparatus (100) of claim 8, wherein The application unit (104) is configured to receive a measured real friction compensation result value of the machine tool based on the applied set of friction compensation parameters, and the analysis unit is configured to modify each weighting factor of the respective surrogate model according to a difference between each friction compensation result value and the measured real friction compensation result value and resulting from the respective surrogate model, and to repeat steps b) to g) of claim 1.

10. The apparatus (100) according to claim 8 or 9, comprising a generator (105) configured to generate the plurality of surrogate models for friction compensation by means of a regression method (RM) based on a given plurality of data sets (DATA), wherein, Each data set comprises a set of friction compensation parameters (CP') for setting a friction compensation of a machine tool and a corresponding friction compensation result (CPR') of a resulting reduced friction within the respective machine tool.

11. The apparatus of claim 8 or 9, wherein, The device is connected to a database (DB), wherein the database is configured to store data sets (DATA) and / or surrogate models (SM).

12. A computer program product directly loadable into the internal memory of a digital computer, comprising software code portions for performing the steps of any one of claims 1 to 7 when said computer program product is run on a computer.

Images