Method and apparatus for simulating machining on a machine tool using a self-learning system - Patents.com

Abstract

To adapt a behavior of a digital machine model to an actual machine tool to efficiently optimize a machine work process on a machine tool.SOLUTION: The present invention is directed to a method of and a device for simulating a machine work process of workpiece on a machine tool subjected to NC control by a self-learning artificial neural network AAKI. The AAKI is provided with process parameters from both the machine work process on an actual machine tool located in a manufacturing process FA and a digital machine model mounted in a simulation section SA, for learning the behavior of the machine tool. The parameters are converted in format into input parameters E1-EN through mathematic conversion. Through learning a behavior of the machine work process, the AAKI sends an output file F1 back to a simulation software of the simulation section SA to make the simulation parameters more adaptive.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an apparatus and method for simulating a machining process of a workpiece on a machine tool by means of a self-learning artificial neural network, which is capable of receiving process and parameter data of the machining process from both a real machine tool and a digital machine model and using them to optimize the simulated and / or real machining process. [Background technology]

[0002] Due to the continuous increase in the complexity of machining processes of today's workpieces, especially in the area of ​​machine-performed or machine-assisted machining of workpieces, new types of machine tools are usually faced with a number of increasing qualitative or economic requirements. The increasingly difficult process mechanics require more powerful and / or more precise machine kinematics with improved functionality of the machine dynamics, drive trains or control systems, but in most cases still result in increased set-up times and difficult, costly and especially expensive commissioning.

[0003] The simulation of the machine tool preferably reproduces the course of the machining process of the respective workpiece on the digital machine tool model. For this purpose, various machine models, such as multi-mass models, geometric kinematics or finite element models, are used to describe the physical properties and interactions of the machine elements and the workpiece, and are combined with control software for moving the machine elements. Furthermore, the simulation of the process based on the penetration calculation between the workpiece and the tool may also be advantageously used.

[0004] EP 1 901 149 B1 shows in particular a machine simulation for defining a sequence for machining a workpiece on a machine tool, in which data structures are integrated into the simulation that make it possible to integrate data or behavior of elements recorded by sensors on a real machine tool and thus improve the description of the control of the implemented machine model.

[0005] Furthermore, WO 2012 / 168427 A1 shows a mechanical simulation of a work process on a machine tool using a virtual machine, where CNC controlled sub-processes can be distributed to different processor cores working in parallel and therefore calculated in parallel to speed up the simulation process.

[0006] However, in the simulation of machine tools within the machining process of the tools according to the prior art, there has always been the problem that precise specification of all state parameters of the machine tool, the tools used and / or the workpiece, and in particular the evolution of their physical properties over time, is not possible without considerable effort. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] European Patent No. 1901149(B1) [Patent Document 2] International Publication No. 2012 / 168427(A1) Brochure Summary of the Invention [Problem to be solved by the invention]

[0008] Therefore, one object of the present invention is to provide a method for simulating a machining process of a workpiece on a machine tool, which solves the above-mentioned problems of the prior art and in particular makes it possible to adapt the digital machine model in the simulation of the process to the conditions and characteristics of the real machine tool, and / or to improve it as efficiently, cheaply and quickly as possible.Furthermore, it is an object to optimize the simulation and the associated adaptation of the simulation parameters in such a way that it can be automated and therefore performed as far as possible independent of human error. [Means for solving the problem]

[0009] The features of the independent claims are proposed to achieve the above mentioned objects. The dependent claims relate to preferred embodiments of the invention.

[0010] The present invention details a method and an apparatus for simulating a machining process of a workpiece on a machine tool, which is configured to use the generation of simulation data by simulation of the machining process carried out in the simulation section of the method on a digital machine model and the recording of machining data of the machining process on a real machine tool carried out in an independent manufacturing section of the method, in order to collect comparison data between the simulated machining process and the real machining process and to supply said comparison data to an artificial intelligence (AI, artificial intelligence) implemented in the analysis section of the method in order to improve the effect of the simulation. Advantageously, an artificial neural network is used as the artificial intelligence, and it is particularly advantageous that the data model for storing the simulation data is configured as a time-continuous data model.

[0011] The control of the digital machine model and the real machine tool is preferably performed according to predefined NC and / or PLC data, and the analysis section is configured to learn the behavior of the machine tool, at least one tool, and / or workpiece by feeding the AI ​​machine with simulation and machining data, and outputting simulation modification parameters, e.g., for modifying and / or optimizing the properties of the simulation, as an output data set (output of the AI ​​for process control, prediction, and optimization). Furthermore, the AI ​​modeled conditions may be used to interpret and optimize the conditions of the real machine tool.

[0012] The present invention, by (preferably automatic) improvement of the simulation, also called machining process simulation, in particular by a self-learning AI, thus generates a simulation environment that is more accurate and at the same time more cost-effective and efficient compared to the prior art, preferably working independently and therefore not requiring additional pauses or waiting times. In addition, evaluating and interpreting the resulting simulation data allows a simplified and at the same time highly accurate optimization of the machining process of the real tool, in particular since possible inefficient settings in the machine tool can be recognized by the simulation of the machining process quickly and without the aid of complex sensors and can be adjusted for the respective work steps.

[0013] Preferably, the digital machine model could already be a remake of the machine tool in the machining section that is as accurate as possible even before AI optimization, particularly preferably a digital twin (digital image of the real machine tool), so that changes to the simulated machining process in the simulation section can output a prediction about the outcome of the machining process on the machine tool as close as possible (analysis of the machining process) as long as the same changes are made.

[0014] At the same time, preferably, the AI ​​located in the analysis section may be configured such that, through the above-mentioned learning process, the AI ​​is able to recognize the differences between the machining process of the digital tool model and the machining process of the machine tool in the machining section, and can be used for automatic improvement of the simulation of the machining process by outputting simulation change parameters.

[0015] For example, in a preferred exemplary embodiment, the AI ​​may be configured to adapt the state of the simulated machining process to the machining process of the machine tool by outputting simulation change parameters, such that the smallest possible differences between the simulated machining process and the machining process on the machine tool are realized. This has the advantage, in particular, that by adapting the simulation process with high accuracy, not only can more accurate and therefore more realistic simulation predictions for the respective machining process be made, but these can also be used for a more accurate optimization of the machining process on the machine tool.

[0016] In a further preferred exemplary embodiment, the AI ​​may be further configured to use in the simulation section (e.g. by changing the rotational speeds, the travel paths, the tools used, the workpiece geometry or the process trajectory) the most efficient settings of the future machining process (or to provide real-time optimization parameters in parallel with the currently executed, i.e. ongoing, real machining process) by learning the behavior of the machine tool, the workpiece to be machined and the various tools. In other words, the AI ​​may be used to improve the simulation of the machining process, preferably not only by optimizing and / or adapting the existing simulated process, but also by predictive methods preceding the simulation process (or in parallel with the real machining process). For this purpose, it may also be preferred that the simulation data are described by the simulation changed parameters integrated in the output data set, so that on the one hand, known data structures are kept in any optimization methods and thus efficiently used, and on the other hand, all run-in times of the simulated and / or real machining processes can be reduced to a minimum by the predictions generated by the AI.

[0017] Preferably, the output data set of the analysis section, including the simulation change parameters, may be fed back to the simulation section to optimize the simulation of the machining process, e.g. to match the simulated machining process with the machining process of a machine tool and / or to predict efficient process sequences, such that preferably a closed program loop is generated and each simulation can be optimally configured.

[0018] The introduction of the output data set into the simulation section may preferably be associated with the simulation software in such a way that at least one, but preferably all, of the simulation parameters declared as modified by the analysis section are adjusted by integrating the output data set of the analysis section into the simulation section. For this purpose, the individual simulation modification parameters of the output data set may preferably be given, for example, a digital simulation marker, which is preferably read in the simulation software and instructs the simulation software to set the simulation parameter associated with the simulation marker to the value stored as the simulation modification parameter. It may also be advantageous for the artificial neural network to be configured to optimize the simulation parameters of the simulation of the machining process in such a way that the smallest possible difference between the selected process parameters of the machining data and the selected process parameters of the simulation data occurs.

[0019] The simulation change parameters of the output data set may include various, preferably all, of the simulation parameters of the simulation software that are changed, such as the geometry of the equipment or machine elements of the tool, the grinding or cutting conditions, the trajectory or physical properties of the elements of the machine or workpiece (temperature, elasticity, friction coefficient, etc.), but may also include basic changes to the complete set of commands or settings, for example the selection of the respective machine model, thereby giving the artificial intelligence the maximum number of degrees of freedom to optimize the simulation.Similarly, the output data of the analysis section may preferably have the same data format as the simulation data of the simulation section and the processing data of the manufacturing section, so that they can be optimally used by omitting parsers that may have different loading and reading speeds for each.

[0020] In a particularly preferred embodiment, a program loop comprising at least simulating the machining process, feeding the simulation data to the analysis section, introducing an output data set into the simulation section, and setting the simulation parameters based on the output data set may be executed iteratively, preferably automatically, with a number of iterations n being at least n > 1. As mentioned above, a new machining process, e.g. using a modified workpiece and / or modified tools or tool settings, may preferably be executed within the simulation section at each iteration and may be taught to and adapted by the AI, thereby generating a continuously evolving optimization method at each iteration step.

[0021] The mutually essential process steps in the iterations of the program loop may be understood as a seamless method, which may preferably start with the transmission of an output data set based on the information learned by the AI ​​to the simulation section, thereby allowing the optimization of the simulation parameters before the actual simulation based on the simulation change parameters implemented in the output data set. In a next step, a simulation of the respective machining process may then be performed, and the result data thus obtained may be supplied to the analysis section including the artificial intelligence as process parameters for teaching the AI. The latter may preferably be performed in such a way that both the process parameters of the simulation section and the process parameters generated on the machine tool in the machining section are recorded in the analysis section, linked to each other, and then sent directly to the AI ​​as input parameters used to teach the AI. Such a program loop thus allows the simulation of every further machining process to show an improved optimization capability of the system, since the AI ​​constantly receives and learns new information about the machine tool, the workpiece, and the tool.

[0022] However, the sequence and mode of actions of the mentioned process steps are not limited to the mentioned exemplary embodiment: for example, a simulation of the machining process may be preferably performed as a first step, and the process parameters thus obtained may be fed as "actual values" of the current simulation to the AI, which may interpret and in turn optimize it by forwarding the output data to the simulation section.

[0023] Likewise, it may be preferred that the optimization process is performed independently of the machining processes in the machining section or on the machine tool. In a particularly preferred exemplary embodiment, the AI ​​may only use information / knowledge about the machine tool, the workpiece and the tool acquired in a previous learning phase, and thus optimize each simulation according to the operator's / user's wishes. The AI's learning may be completed before the actual simulation or may be extended at a later point in time, and additional process parameters of the machining section may be generated, for example, by machining processes performed later or previously on the machine tool, to provide maximum flexibility of the specified optimization process.

[0024] The process parameters of the simulation and machining data may preferably differ from the simulation parameters of the simulation of the machining process in that they only contain information related to the machining process, i.e. the characteristics of the actual (real) or simulated machine tool, the tools used, and the workpiece, and, at least in the case of the process parameters of the simulation data, preferably do not contain information related to the settings of the higher-level simulation (e.g. information from the simulation model or simulation functions used). Thus, in this case, similar to a person engaged in a manual optimization process, the AI ​​has only learning parameters that can be measured on the machine tool available to teach and optimize the simulation, thus reducing to a minimum the occurrence of human errors or wrong decisions by the former.

[0025] Preferably, the simulation data may also be generated such that, at least for each process parameter of the machining data, an equivalent process parameter exists or can be generated from the simulation data or can be derived from the individual process parameters in the simulation data, thereby further maximizing the number of training parameters used for the AI ​​and thus allowing the AI ​​to be trained as diversely as possible. In a particularly preferred embodiment of the invention, the data sets of individual process parameters in the simulation and machining data may also be time-dependent, preferably such that at least one corresponding process parameter in the simulation data is present at any time of a process parameter of the machining data and can be assigned to a process parameter of the machining data.

[0026] Besides the described link of the process parameters of the simulation section to the process parameters of the machining section, the usual generation of simulation data by simulation of the machining process also has the advantage that these can not only be used as comparative data for the learning of the AI, but also preferably allow direct conclusions regarding the (e.g. ongoing) machining process on the machine tool. Using the data generated from the simulation, for example wear, temperature or movement data of individual machine or workpiece elements, the user / employee may also identify errors or incorrect settings on the machine tool and therefore take measures with the help of a simulation of the machining process (for example running in parallel) to optimize the machining process. In other words, the mentioned optimization process by the AI ​​may not only be used for analysis and / or for a more accurate simulation of the machining process, but may also help to optimize the machine tool of the machining section so that the machining process can be performed more efficiently and with higher quality.

[0027] In a particularly preferred embodiment, these measures may be performed fully automatically, preferably by feedback between the simulation section and the machining section, and thus may be considered as a further optimization process of the AI. Process parameters present in both the machining section and the simulation section (e.g. PLC and / or NC data, tool selection or speed) may for example be transmitted from the simulation section to the machine tool in a time-dependent manner and used for improved control of the machine tool in the machining process. In a further exemplary embodiment, these process parameters may be transmitted directly with the help of output files of the AI. In other words, the AI ​​may preferably be configured in such a way that it can control both the machining process on the machine tool and on the digital machine model, and thus optimize them simultaneously by implementing output files in the machining section and in the simulation section. Such a method may thus independently improve the machining process based on previous work processes, thereby not only reducing the time and costs of the optimization process to a minimum, depending on the level of training of the AI, but also guarantee a digital machine model from which a large amount of information important for the machining process can be obtained.

[0028] To further improve the method, the process parameters of the simulation and machining sections may be further stored in a data memory provided for this purpose, preferably after their generation on the respective manufacturing or simulation section, or even before being fed to the analysis section, and the corresponding simulation data are preferably added to a simulation database implemented in an iterative loop, the process parameters of the manufacturing section being added to a manufacturing database that is arranged separately or independently of the simulation sequence and that may be declared accessible for the respective operator. This has the particular advantage that the operator can also see the results of previous AI decisions at a later point in time, for example after further learning and / or optimization iterations, evaluate these results if necessary and use these results to adopt improvements or modifications of the particular machining process. Furthermore, the intermediate storage of the individual process parameters in the respective databases allows the AI ​​to be reset to a previous learning stage in the event of erroneous or insufficiently resulting parameters.

[0029] In order to learn and optimize the machining process by the AI, some process parameters of the simulation section and the manufacturing section, preferably from the simulation and manufacturing databases, may also be introduced into the analysis section, where they are converted into individual input parameters adapted to the AI ​​by comparison and fed to the latter. The input parameters are preferably not limited to the process parameters themselves or to a direct comparison of those process parameters between the machining process carried out in the machining section and the simulated machining process, but may also represent, for example, a combination of several process parameters, a mechanical, economic or qualitative evaluation based on the process parameters or related functions, so that the AI ​​can be freely taught according to the operator's wishes and used to optimize the simulation of the machining process.

[0030] Moreover, to generate these input parameters, the analysis section may preferably include at least one data link section and a data interpretation section, which includes a number of processing segments, but is preferably connected upstream or in parallel with the input to the AI. The data link section may preferably be configured such that it receives the simulation and machining data stored in the manufacturing and simulation databases, searches for comparable process parameters accordingly, first links these process parameters as data bundles, and transfers them to the data interpretation section, and linking the corresponding simulation and machining data may preferably be performed by continuous data mapping. (In particular, a continuous time mapping method may be used, whereby the mapping is advantageously performed, for example, by association of the performed operation / machining steps, NC lines, and / or axis positions.) The data interpretation section may in turn, in a particularly preferred embodiment, analyze the linked simulation and machining data in this respect, compare them with other data bundles, and generate from them any number of input parameters for the AI, which in turn transfers these input parameters to the artificial intelligence. With the help of such processing segments, which preferably work independently of one another, optimized and particularly fast methods for repeated supply of input parameters to the AI ​​can be implemented, which methods can be freely adjusted for the respective operator, for example by the code implemented for controlling the processing segments, preferably through an additional input interface.

[0031] Furthermore, the artificial intelligence which finally obtains the input parameters may preferably be connected directly to the output of the processing segment and configured as a cluster method, a support vector machine or, particularly preferably, an artificial neural network. The latter may also preferably have a typical network structure, for example a single or multi-layer forward propagation or recurrent network, which may also be modified in the analysis section so that the best improvement method may be selected depending on the complexity of the machining process to be optimized. In order to optimize the simulated machining process, any learning algorithm of the artificial neural network with the desired activation function may furthermore be implemented, said algorithm being preferably configured in such a way that at least the simulation modification parameters or simulation parameters generated with the help of the artificial neural network finally result in the desired optimization of the simulation of the machining process, for example by adapting the simulation to predictions of the conditions on the machine tool or of the targeted settings for the simulation of a new process.

[0032] As mentioned above, the use of AI for optimizing the simulated machining process has the advantage that it can be adapted in various ways to the respective problem or to the respective preferences of the operator and can be operated during the optimization process without human intervention, in particular minimizing time and labor costs. However, there are also other objectives, namely that in certain preferred embodiments, the learning of the AI ​​and the optimization of the simulated process by the AI ​​can also be performed in parallel and / or independently of the respective real machining process on the machine tool, which can lead to further time savings. Likewise, the training process of the AI ​​can also be configured in such a way that it is performed in a segment preceding the simulation and treatment process, for example by making available multiple training data sets before machining begins. In particular, in this way, it is also possible that an already trained network can be used for the simulation of multiple processes, since the training data, of course, do not necessarily have to depend on the structure of the respective machine tool or the respective machining process and can therefore provide an optimal basis for a more specific learning phase. Thus, the artificial neural network can be trained with the training data sets in a training process preceding the simulation and machining process.

[0033] In a further preferred exemplary embodiment, the AI ​​may also store the generated output data sets in an extensible technology database, where the introduced simulation change parameters can both be seen and returned to the AI. Thus, on the one hand, the operator may be able to understand the simulation data generated in each iteration based on the simulation change parameters and continue to use them if appropriate, and on the other hand, preferably, the latter may also be used to further improve the above-mentioned learning and optimization process, for example by adding individual simulation change parameters back to the recursive network system. Furthermore, it may be preferred that at least the technology database, the simulation database and the manufacturing database are located in an external system, for example in the cloud or an external network, so that external employees besides the operator or manufacturer can efficiently access the stored data and use them, for example, for further machining or change processes, preferably for multiple machine tools.

[0034] In order to ensure generally the same requirements for the machining process in the simulation and manufacturing sections, the two method sections may also include further characteristics. For example, it may be preferred that the digital machine model implemented in the simulation section generates the overall geometry of the machine tool, the tools used, and the workpiece to be machined, but at least the parts required for the corresponding machining process according to their real equivalents. Furthermore, in a highly preferred exemplary embodiment, the NC data and / or PLC data used to control the individual machine elements of the machine tool may also be used in the same form to move the simulated elements of the digital machine model, so that an exact digital copy of the real machine tool is available in the simulation section.

[0035] Such a machine model, already described as a digital twin, has the particular advantage that any properties or results of the simulation of the machining process can be directly compared with the data of the real machine tool. Furthermore, the simulation of the machining process on such a model may be preferred to output physical parameters of the machine tool or tool and / or of the workpiece to be machined, which are difficult to determine in reality or can only be determined with great effort, such as temperature, elasticity, friction coefficient, etc., so that all problems in the machining process can be identified quickly and efficiently. For this reason in particular, the digital machine model may therefore be preferred to output time-dependent data series of at least these physical parameters, which may be defined, for example, by additional time and position markers that are linked to the data series depending on the time of the respective machining process and / or the respective work step, and which may be output as process parameters.

[0036] In order to simultaneously further improve the equivalence of the digital machine model and the real machine tool, an analysis of the machine tool elements, workpieces and / or tools used in each case carried out during the product analysis section can also be carried out on the real machine tool, preferably during and / or after the machining process on the real machine tool, the analysis being carried out for example by sensors implemented on the machine tool or by an externally mounted analysis unit. Here again, the collected data is preferably defined as process parameters and stored in the abovementioned manufacturing database.

[0037] Furthermore, it may be preferred that the above-mentioned NC and / or PLC and geometry data used to define the digital machine model and the real machine tool are adapted to the process growth in the input data pre-processing section before the machining process. For example, the work steps of the machining process defined by at least the NC and / or PLC data on the machine tool and on the digital machine model may be first generated on an upstream CAD / CAM system and transferred from the CAD / CAM system to the mentioned input data pre-processing section as motion files, which should be understood as a general data package having information about the geometry of the machine tool and the workpiece, the sequence of movements of the individual elements and / or an identification structure such as a UUID. Depending on the data format of the motion files, the input data pre-processing section may also include one or more parsers that convert the former into a suitable input format before further processing, and the input data pre-processing section itself can process motion files that are formatted differently and / or are generated externally.

[0038] The motion files created in the CAD / CAM system may then be converted to NC data format in an input data pre-processing section and transferred to a manufacturing section or to a simulation section together with corresponding manufacturing and status data of the simulated machine tool, tool and / or machined workpiece, which are similarly reformatted. This has the particular advantage that the motion files may be reformatted independently of the respective simulation cycle and therefore do not unnecessarily disturb or slow down the simulation cycle.

[0039] Furthermore, as a result of reformatting the motion files into an NC format to identify the respective work steps, it may be preferable for additional step markers to be inserted into the NC data so that the real machine tool and the digital machine tool model, e.g. with the help of a control model, are able to interpret the step markers in the machining process and therefore understand during the machining process which work step the (digital) machine tool is currently in.

[0040] In a further preferred embodiment, the work steps already stored in the NC data can be further output from the input data pre-processing section as a formatted structure file, for example as an XML or STEP file for interpretation in other simulation systems, so that the above-mentioned formatting process can also be used when changing the respective simulation engine and can therefore be viewed independently of the simulation process.Similarly, on the other hand, the machine tool and the digital machine model can also be preferably configured in such a way that the information of said work steps and / or processes can be obtained, for example by a further parser module, from other data formats, transferred and interpreted for carrying out the respective work steps involved, thereby preferably enabling communication with other simulation networks or even parallel processing of different simulations.

[0041] In short, based on the above-mentioned characteristics, a method for simulating machining on a machine tool using a self-learning artificial intelligence can be implemented, in which a simulation for the analysis and improvement of a machining process used on a machine tool can be optimally adapted to the characteristics of the respective real machining process using a learning process of the AI ​​by comparison data based on both the data of the simulation and the data of the machine tool, and can therefore be used to predict the machining process to be further developed. Here, the optimization process of the simulation preferably mainly includes the execution of a simulation of the machining process on a digital machine model implemented in a simulation section, the transfer of the simulation results as process parameters to an artificial intelligence arranged in an analysis section, the learning of the artificial intelligence based on a comparison of the process parameters of the simulation section with the process parameters of the machine tool, and the feedback of the modified simulation parameters to the simulation section by an output file generated by the AI. The necessary optimization steps can be performed independently or in parallel with the respective machining process on a real machine tool, and can be performed fully automatically after teaching the AI, which can also be performed before the actual production. Furthermore, the method has the advantage that the same NC data, and in the case of simulation, geometry data, is used to control both the real machine tool and the digital machine model, which allows the generation of an exact digital copy of the real machining process, supported by additional analysis of the machine tool, the workpiece and the tools used.

[0042] Therefore, it may be preferable for a corresponding apparatus to be implemented which realizes the above mentioned features and which can be used for simulating the machining process of a workpiece on a machine tool by means of a self-learning artificial intelligence.

[0043] Therefore, the apparatus may preferably include at least one machine tool for machining the respective workpiece using the abovementioned NC data, a simulation device, e.g. a mainframe computer or a server, which may be controlled independently of the machine tool, for simulating the machining process on the digital machine model, and an analysis unit connected to the machine tool and the simulation device, configured by means of an implemented artificial intelligence to adapt the simulation parameters in the simulation device, to learn the behavior of the machine tool, the workpieces used and / or the tools, and to output the results using an output data set. Furthermore, the machine tool may be configured to transfer machining data generated during or after the machining process to the analysis unit equivalently to the simulation data made in the simulation device, whereby the abovementioned program or iterative loop between the analysis unit and the simulation device may result.

[0044] Similarly and equivalently, for separate optimization of the simulation of the process, at least the analysis unit and the simulation device may preferably work independently of the machining process of the machine tool, such that multiple optimization or simulation processes guided by AI may be iteratively executed by the above-mentioned program loops, but these processes may still be able to be extended at any time with the process parameters of the machine tool, for example by accessing a corresponding database with further information.

[0045] In order to transfer information between the individual apparatus elements, the machine tool, the simulation device and the analysis device may further preferably be configured in such a way that they can transmit data independently of each other, in particular that the abovementioned simulation and machining data as well as the output files generated by the analysis device as well as hardware and / or program data can be transferred to each other, which may preferably be stored in corresponding data memories or central storage servers so that the simulation and optimization processes can be optimally performed. In a particularly preferred exemplary embodiment, this transmission may be performed, for example, by means of an intranet and / or the Internet. [Brief description of the drawings]

[0046] [Figure 1] FIG. 2 shows a network representation of a first exemplary embodiment of the detailed method. [Diagram 2] FIG. 2 illustrates a network representation of an exemplary embodiment of the analysis section of the detailed method. [Figure 3A] FIG. 2 illustrates a network representation of an exemplary embodiment of the artificial neural network of the method. [Figure 3B] FIG. 3B illustrates a network representation of the learning process of the artificial neurons of the artificial neural network of FIG. 3A. [Figure 4] 1 is a flowchart showing the steps of a learning process for an artificial intelligence (machine learning device). [Diagram 5] 13 is a flowchart of an operation procedure of a simulation section. [Figure 6] FIG. 2 illustrates a network representation of a second exemplary embodiment of the method described. [Figure 7] FIG. 1 illustrates an exemplary list of parameters obtained from a virtual and real machine tool. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0047] In the following, exemplary embodiments of the present invention are described in detail based on exemplary figures. The features of the exemplary embodiments can be combined in whole or in part, and the present invention is not limited to the described exemplary embodiments.

[0048] 1 shows a first simplified embodiment of the method according to the invention for simulating the machining process of a workpiece on an NC-controlled machine tool by a self-learning artificial intelligence AAKI (in particular an artificial neural network) in a network representation. Firstly, the geometric structure of the machine tool, which is reproduced in a digital machine model including the workpiece and the tools used, as well as the work steps required for the machining process, are defined by the CAD / CAM system and transferred in an operation file D1 to the input data pre-processing section PRE. Here, the operation file D1 may already exist as an NC file or the work steps involved may have already been converted into an NC format in the CAD / CAM system, but any other file format can be read and interpreted in the input data pre-processing section PRE, inter alia, by a parser implemented in the input data pre-processing section PRE.

[0049] Within the input data pre-processing section PRE, the operation file D1 is then divided according to its functional components, i.e. for example the geometric data of the machine tools, tools and workpieces used as well as the data of the respective work steps of the machine tool related elements, which are analysed and, if necessary, existing work steps are reformatted into the NC file G1. Furthermore, for an improved identification of each sub-process, each registered work step in the NC file receives a specified UUID which can be queried by the respective machine tool or digital machine model during the machining process and thus can be used to identify the currently ongoing machining process.

[0050] In a next step, the files of information required are then transferred to the simulation section SA or to the manufacturing section FA provided for this purpose, thereby simultaneously dividing the process into a real part (lower section of the network) and a part belonging to the simulation (upper section of the network). In order to control the real machine tools used in the manufacturing section FA, at least the NC data formatted in the input data pre-processing section PRE are exported from the input data pre-processing section PRE to the manufacturing section FA in a machine file G1 together with other program data used for the control of the machine, such as identifiers or backup files, which are transferred to the machine tools after repeated checks by the machine operator or by automated checking algorithms (not shown). On the other hand, with the help of the input data pre-processing section PRE, the simulation section SA including the simulation software is provided with a separate simulation information file G2 which, in addition to the NC data of the machine file G1, further contains at least the geometrical construction data of the machine tool required for the generation of the digital machine model and further preliminary information such as the physical properties of the individual machine elements, the workpiece or the tools used, so that the simulation section has at least sufficient information or parameters to start the corresponding simulation.

[0051] In addition, the individual data processed in the input data pre-processing section PRE can be reformatted, if necessary, into another data format, such as XML or STEP, for example by an implemented compiler, and these can be stored in a separate data bundle G3 for use by external simulation software or hardware, thereby enabling in particular a direct comparison of different simulation engines or structures, notably parallel simulation on multiple simulation sections SA, and thus even a speed-up of the simulation process using multiple processor cores working simultaneously.

[0052] Furthermore, after the real section of the method, the real machining process in the manufacturing section FA can now be performed by providing the NC file G1 to the machine tool. However, the course of this process is not dependent, in time or otherwise, on the above-mentioned section of the method related to the simulation of the machining process, but rather serves to generate reference or process parameters that are used to teach the artificial neural network AAKI. Similarly, the number of completed machining processes in the manufacturing section FA is not fixedly defined, but can be manually specified at any time by the operator and / or can be increased at a later time for a more precise verification of the individual parameters. The latter can be obtained not only by a number of sensors attached to the machine tool or external, but also by manual input of a qualitative analysis process, for example by evaluation of the finished workpiece by an expert during or after the machining process in the so-called machine, process and production analysis section PA, and the process parameters thus finally obtained and bundled are first stored in a database DB1 directed to the real manufacturing and then transferred as machining data R1 to the analysis section AA provided with the artificial neural network AAKI.

[0053] In the part corresponding to the manufacturing section of the method, when the simulation information file G2 is received in the section of the method related to the simulation of the machining process, the initiation of the simulation of the machining process and therefore also of the associated optimization and / or learning process can be undertaken. Firstly, in the simulation section SA, using the implemented simulation software, with the help of the simulation information file G2, a corresponding digital machine model including the tool and the workpiece is generated, which is as similar as possible to the real machine tool of the manufacturing section and which can be controlled by the machining process with the help of the NC data also received. In parallel to the real method section, all process parameters used are also determined using an analysis module integrated in the simulation section SA, which are stored by the simulation section SA in a separate simulation database DB2 in a format that is preferably the same as that of the machining data and transmitted to the analysis section AA as simulation data R2. Here again, the number and type of process parameters selected can be manually selected depending on the application and purpose of the optimization process and / or adapted to the process parameters obtained in the real method section, so that the optimization process and / or the learning process can be used as efficiently as possible depending on the machine tool used or the machining process to be adjusted.

[0054] Continuing with the optimization and / or learning process, in the next step, the simulation data R2 and the machining data R1 (or the machining data of the machine tool) are processed in the analysis section AA and sent as input parameters to the artificial intelligence AAKI for learning, which in turn is configured in such a way that it learns the behavior of the machining process on the manufacturing section FA by comparing the simulation data R2 with the machining data R1 (or the machining data of the machine tool) and outputs, if necessary, some simulation change parameters (analysis of the machining process) in the form of an output file F1, which, after its generation, is output in an iterative loop declared by an output 1 A1, stored in a technology database DB3 provided for this purpose and fed back to the simulation section SA in order to optimize the simulation of the digital machine model.

[0055] As mentioned above, the learning of the artificial intelligence AAKI and therefore also the generation process of the simulation data R2 can be carried out independently of the machining process on the machine tool and therefore can also be carried out before or after (or during) the manufacture of the workpiece on the machine tool. It is also possible to train the artificial intelligence AAKI by externally introduced input parameters, shown in Fig. 1 by the process "Learning AI" AL, so that maximum flexibility of the optimization process is realized. Furthermore, the output file F1 created by the artificial intelligence AAKI can contain not only the individual parameters modified by the simulation software, which can be adopted for the simulation by additional markers added to the data, but also the entire settings (simulation model, simulation time, frame rate, etc.) or the functions to be implemented (e.g. interaction functions in the model), as long as these have already been declared in the artificial intelligence AAKI.

[0056] The implementation of the simulation change parameters of the output file F1 in the simulation section SA can start various optimization processes depending on the previous steps and the devised process sequence, i.e. if, for example, a machining process has already been simulated for teaching in the artificial intelligence AAKI, this process can be adjusted in the simulation section SA by the introduction of the simulation change parameters and can be examined for further analysis (for example for comparison with the machining process on a machine tool). However, it is also possible to generate predictions regarding the optimization of a machining process that has not yet been implemented, in particular by the artificial intelligence AAKI already taught or trained using external data, and thus to generate a simulation of a completely new machining process that is as efficient as possible, by transferring the output file F1 to the simulation section SA.

[0057] In a subsequent process, it is also possible to draw further conclusions from the simulation optimization process. For example, if the simulation is successfully optimized, the decision results and system information (system change parameters, weightings, AI parameters, etc.) generated by the artificial intelligence AAKI can be obtained from the final output (output 2 A2) and used to gain insight into and improve the machining process of the actual (i.e. real) machine tool. The same happens with the simulation parameters and information generated within the simulation section, which in this case is shown by the process of "performance output" A3 in Figure 1.

[0058] Furthermore, by optimizing the simulation of the machining process, a direct improvement of the machining process on the machine tool and of the learning process of the artificial intelligence AAKI may be realized. For example, by directly connecting the simulation section SA to the manufacturing section FA or the machine tool by data transmission of the corresponding process data, the optimization of the simulated machining process may be easily implemented in the machining process of the machine tool. In other words, the system shown here allows the simultaneous adaptation of the real machining process to the improved simulated process sequences, which in turn allow the optimization of the machine tool to be significantly accelerated and therefore more cost-effective and efficient. Furthermore, it is possible to feed the optimized simulation data back to the simulation database DB2, so that the optimized simulation data can be used by the artificial intelligence AAKI for training and further improvement of the optimization process. Each optimization process thus has the option of adapting the artificial intelligence AAKI to the conditions on the real machine tool with even greater precision and thus improving the judgments and results of the artificial intelligence AAKI iteratively, i.e., continuously after completing an optimization or iteration loop.

[0059] Fig. 2 also shows a more detailed network description of the analysis section AA of the exemplary embodiment already shown by Fig. 1, by which the implementation of the learning parameters of the artificial neural network AAKI can be explained again in more detail. Firstly, the process parameters contained in the simulation data R2 and the machining data R1 are introduced into the data link section AADV implemented in the analysis section AA, in which they are firstly analyzed by their associated markers and corresponding comparable process parameters emerging from the simulation data R2 and the machining data R1 are linked to each other. The latter can be implemented in several ways, for example by copying and saving the separate process parameters in an intermediate memory location provided for this purpose, or for example by marking them with their own ID number, and only concerns the connection of any kind of process parameters leading to the generation of the learning parameters, also called input parameters of the artificial neural network AAKI. Furthermore, attempted comparison processes such as digital mapping may be used to link more complex structures (e.g. time-resolved data sets) so that an optimal comparison between simulated process parameters and process parameters acquired on the machine tool is possible.

[0060] The process parameter pairs and the individual process parameters thus linked are then passed in a next step from the data link section AADV to a data interpretation section AADI of the analysis section AA, where they are again identified and converted into the desired input parameters E1-EN for the subsequent artificial neural network AAKI and finally introduced into the artificial neural network AAKI. Any combination or mathematical processing of the process parameters of the simulation data R2 and the machining data R1 and / or input parameters such as the NC data G1 generated by the input data pre-processing section PRE, the combined geometric data G2 or the motion file D1 generated in the CAD / CAM system can be understood as a conversion or generation of the input parameters E1-EN.

[0061] Furthermore, Figures 3A and 3B show a more accurate representation of the learning process of the artificial intelligence AAKI, which in this exemplary embodiment is represented as an artificial neural network in the form of a further network representation, where Figure 3A shows the general inputs and outputs in the network and Figure 3B shows a detailed representation of the decision-making process carried out in the intermediate layer in the artificial neuron. Here, the depicted structure of the shown network should be understood as merely an exemplary representation.

[0062] As mentioned above, first a set of potential input parameters E1 to EN is generated by the accumulation of process parameters generated by the manufacturing section FA and the simulation section SA as well as other information generated in the previous method, which in turn can be further combined by the data interpretation section AADI and finally introduced into the first learning layer of the network as a newly combined element KOM or as one of the parameters mentioned above as input. Thus, according to the rules of the self-learning system, in the relevant neurons of each network layer, they are then given a weighting W1 to WN for each input parameter, which is assigned a random value before the start of the learning phase and can be gradually changed by a "trial and error" algorithm with each individual learning iteration up to the desired decision weighting W1 to WN. In the usual sense, the interaction of the input parameters E1 to EN with the respective weighting W1 to WN can be performed by a simple product Ei x Wi of the respective i-th parameter with the i-th weighting, although in other exemplary embodiments other or more complex functions can be selected.

[0063] In the next step, the input-weight combination is then combined into the network input by a transfer function Fi. For example, following the rules of artificial neural networks, Σ i E i W i A simple summation of the form: could be used to generate the network inputs, although this approach could be modified depending on the problem and application of each optimization attempt.

[0064] And now the network input determines whether the artificial neuron will activate ai upon receiving all the weighted and summarized inputs, and thus be allowed to pass the information to the next layer, by inputting a predefined activation function Ai. In general, a comparison is made as to whether the activation function Ai at the time of the network input exceeds a certain threshold that must also be learned (thus activating the neuron ai), or whether the resulting value is insufficient and the neuron remains inactive. As in the previous case, the activation function may be chosen freely, but at least in this exemplary embodiment, a sigmoid function is preferred due to its shape, which is continuous at all points and differentiable.

[0065] By teaching the artificial neural network, the selected inputs EN can thus be transferred from one network layer to the next using the weights WN obtained with the help of the learning process, and thus finally transferred to the outputs 1 and 2 in the last layer as preferred simulation parameters. Depending on the network structure, they can also be returned to the original set of inputs, for example in a recurrent network system, or to a specific neuron or layer, so that a feedback system generated from the output file grows. The learning of the network or the above-mentioned weights W1 to WN, and in special cases also thresholds, can be carried out before the input parameters E1 to EN are actually provided, for example with already defined training data, and can therefore be seen independently of the actual optimization process.

[0066] FIG. 4 again illustrates the structure of the learning process using a flow chart showing the learning procedure, where "start" is understood as the normal operating concept for performing a training process of the AI, and "end" is understood as successfully teaching the AI ​​the desired characteristics. Each learning process starts with a normal decision SA01 of the system (or a person) wanting to perform a learning process of the artificial intelligence AAKI, and various factors may be included as the basis of these decisions. The learning process may be used to adapt the simulation of an existing machining process to the desired situation, for example the operating procedure of the machine tool, other than to make predictions about new tools to be inserted, operating modes or newly restored machine tools, thus increasing the accuracy and efficiency of the simulation. In case of a successful decision, the machine information SA02 and machining conditions SA03 required for the training process are collected and first prepared for introduction into the analysis section AA integrated with the artificial intelligence AAKI. All physical information about the machine tool that can be used to successfully describe the machining process to be learned, i.e. for example the shape, size, material properties or physical characteristics of the tools, workpieces and / or individual machine tool elements, may be considered as machine information. Similarly, all those machine settings or conditions, such as tool trajectories, rotational speeds or machining speeds, that must be applied to the individual (simulated) elements of the machine tool to perform the above mentioned machining processes, may be considered as machining conditions. Said data may be generated from externally stored data sets (e.g. server or cloud). However, they may be supplemented or overwritten by parameters resulting from a parallel or previously executed simulation of the machining process SA04 or a machining process on the machine tool SA05, so that current information on the machining of the tool can be continuously fed to the learning process of the artificial intelligence at any time.

[0067] Furthermore, a direct comparison of the simulation results with the real machining conditions allows a further analysis and decision process SA06 to be carried out in the next step. For example, after generating the input parameters defined for the artificial intelligence AAKI in the above-mentioned manner, a decision can be made by the input or other parameters themselves other than by comparing the simulation data R2 with the machining data R1 as to whether learning of the artificial intelligence AAKI is even necessary or, for example, whether the current simulation settings already meet the desired conditions. For this purpose, for example, the above-mentioned parameters can be compared with particularly defined limit values, and if they are below the limit values, they can be used to automatically continue the learning process, thereby realizing an individual trigger for each simulation or learning process. As a result of a positive decision ("yes"), the input parameters can also be introduced into the analysis section AA or the artificial intelligence AAKI by the above-mentioned steps (SA07) and the learning process SA08 can be started.

[0068] 5 further illustrates the structure of the operating procedure and the continuing process of the simulation of a machining process based on another flow chart, where "start" should be regarded as the normal operating concept for starting the simulation of a machining process, and "end" should be regarded as the successful completion of the same simulation.

[0069] Similar to the diagram structure of the learning process, each simulation process first starts with a decision SB01 to execute the learning process, which can be made either manually, for example by an employee, or automatically, for example by an integrated program code or an artificial intelligence AAKI. Similarly, if the input is positive ("Yes"), the (digital) machine information SB02 and machining conditions SB03 required for the simulation are obtained from available servers, cloud services or other types of databases and prepared for implementation in the corresponding simulation section Sat. The former can include both the above-mentioned information data about the physical or kinematic conditions of the machining process and the general software settings related to the simulation (for example, the engine used, the simulation model, the parameters set in the software).

[0070] In a next optional step, the introduced machine information and machining conditions or the simulation parameters obtained therefrom may also be adapted by introducing the simulation modification parameters generated by the analysis section AA and thus optimized using the judgment of the trained artificial intelligence AAKI (SB04). Depending on the machining process, the simulation modification parameters or the output file F1 containing these parameters may be output directly from the analysis section AA or obtained from an existing database (for example the technology database DB3), provided that the machine characteristics related to them correspond to the machine information and machining conditions of the simulation of the machining process now to be simulated.

[0071] If the simulation parameters used for the simulation then correspond to the desired specifications, the simulation of the machining process is started and then evaluated (SB05), which means that the actual simulation process may be considered to be finished. However, in a continuing manner, after comparing certain conditions (e.g. does the efficiency or result of the simulation correspond to certain requirements?) (SB06), it is possible to use the results or knowledge obtained from the simulation of the machining process for the machining process on the machine tool. For example, after ensuring the quality of the simulation results (SB06-"Yes"), the latter may first be output separately (SB07), transmitted and / or reused for further use, for example to analyze possible inefficiencies in the machine tool. Furthermore, besides the simulation results, it is also possible (and may be) to transfer directly to the machine tool (SB08), for example the simulation parameters used in the simulation of the machining process classified as efficient, thus realizing the efficiency and process improvements obtained by the optimized simulation also on the machine of the real tool.

[0072] Fig. 6 also shows a detailed view of the network representation of an exemplary embodiment similar to the network representation shown in Fig. 1, in particular further elements and interactions between the various method sections are shown. In this figure, for example, communication paths of individual equally exemplary elements of the input data pre-processing section PRE are shown, linked to the components of the simulation section SA and the manufacturing section FA. As an example of this, elements of the program structure, for example the geometric data of the digital machine model or the NC data already defined as G-code in this case, are supplied to both the manufacturing section FA recorded for manufacturing the real workpiece and the simulation section SA placed in the simulation engine lo. Other elements not yet described are the G-code interpreter of the simulation section SA, which first reads the G-code of the NC data generated in the input data pre-processing section PRE and introduces a kinematic solver for implementing the movement data of the individual machine elements in the model used for the simulation, and the virtual NC, which generates a virtual copy of the NC data in parallel with the interpretation of the G-code of the respective NC data, thus making them receivable for the above-mentioned kinematic solver. Furthermore, reference should be made to the individual parameters stored in the above-mentioned databases DB1, DB2, DB3, but in particular to the parameters stored in the simulation database DB2, which parameters contain, in addition to the process parameters of the simulation section SA already described, also further information related to the simulation, such as TCP, thus providing a wide variety of parameters used to teach the artificial intelligence AAKI.

[0073] Also, FIG. 7 shows an exemplary comparison of various analysis parameters obtained using a real machine tool and a digital machine model, where the indicated x represents the possibility of obtaining the respective parameter. Said data may be used, for example, as training data for the artificial intelligence AAKI. It is clear from this example that many other factors, such as the depth or width of engagement of the respective tool on the workpiece, in particular compared to a real machine tool, can be easily obtained and employed for the analysis and improvement of the respective machining process. In this respect, the implementation of machining processes in a digital machine model offers a cost-effective and efficient way to improve the process flow, especially during the complex, difficult to implement and / or costly introduction phase of a new machine tool.

[0074] These features, components, and specific details can be interchanged and / or combined to produce further embodiments depending on the intended application. All modifications within the knowledge of one skilled in the art are implicitly disclosed in this description.

Claims

1. 1. A computer-implemented method for simulating a machining process of a workpiece on a machine tool in response to NC data and / or PLC data, wherein a digital machine model of the machine tool is used to simulate the machining process, - executing, in a simulation section (SA), based on said NC data and / or said PLC data, a digital machining process by simulating said machining process on said digital machine model and storing simulation data; - recording machining data (R1) of said machining process on said machine tool, said machining process being executed according to said NC data and / or said PLC data, said machining data (R1) comprising a plurality of process parameters; - supplying said simulation data of said digital machining process and said machining data of said machining process on said machine tool to an analysis section (AA) and linking said simulation data and said machining data; adjusting the machining process on the machine tool based on the simulation data; Including, linking said simulation data to said machining data by carrying out a continuous data mapping based on the association of the positions of the operations to be performed, NC lines and / or axes, and associating sensor data of the machine tool with the simulation data obtained by the simulation in the simulation section (SA); said analysis section (AA) comprising a machine learning device for analyzing said machining process based on the linked data, said analysis section (AA) outputting the results of the analysis; A computer-implemented method, wherein the analysis section (AA) learns the behavior of the machine tool based on the simulation data of the digital machining process and the machining data of the machining process, and outputs the behavior as an output data set (F1).

2. 2. The computer-implemented method of claim 1, wherein the output data set (F1) of the analysis section (AA) is fed back to the simulation section to adapt the digital machining process.

3. 3. The computer implemented method of claim 1 or 2, wherein the simulation of the machining process, the supply of the simulation data (R2) to the analysis section (AA), the feedback of the output data set (F1) to the simulation section (SA) and the modification of simulation parameters in the simulation section (SA) based on the output data set (F1) are formed into a program loop for continuously adapting the simulation of the machining process.

4. At least one simulation parameter is modified in the simulation section (SA) based on the output data set (F1) of the analysis section (AA), The computer-implemented method according to any of claims 1 to 3, wherein the simulation data (R2) generated from a simulation or at least one process parameter of the simulation data (R2) is stored in a simulation database (DB2) before being fed back to the analysis section (AA).

5. 5. A computer implemented method according to any of claims 1 to 4, wherein the digital machining process is performed in parallel with or before the machining process on the machine tool, and instructions for optimizing the machining process on the machine tool are output.

6. 6. The computer-implemented method according to any of claims 1 to 5, wherein the machine learning device is an artificial neural network (AAKI) configured to optimize simulation parameters of the simulation of the machining process such that differences between selected process parameters of the machining data (R1) and the simulation data (R2) are as small as possible.

7. 7. A computer implemented method according to any of claims 1 to 6, wherein the training of the machine learning device and the optimisation of the simulation of the machining process are performed by the machine learning device in parallel with and / or independently of the machining process on the machine tool.

8. The output data set (F1) output by the machine learning device is stored in an extensible database (DB3); 8. The computer-implemented method of claim 1, wherein the machine learning device accesses the output dataset (F1) stored in the database (DB3) for feedback of the learning process.

9. A computer-implemented method according to any one of claims 1 to 8, wherein in order to link work steps of the machining process and work steps of a simulation of the machining process, the same NC data are used for the machining process on the machine tool and for the simulation of the machining process on the digital machine model.

10. A computer-implemented method according to any of claims 1 to 9, wherein physical parameters of the machine tool, the tool and the workpiece to be machined are output by a simulation of the machining process on the digital machine model, and the physical parameters of the machine tool, the tool and the workpiece to be machined are defined as a function of the time of the machining process and / or of the respective work steps.

11. the NC data for identifying each work step is provided with additional markers; A computer implemented method according to any of claims 1 to 10, wherein the machine tool and digital machine model can be interpreted using the markers in the NC data, thereby making it possible to track at which work step and / or in which position the machine tool and / or the digital machine model are at a determinable point in time.

12. The work steps of the machining process are further output as construction data (G3), A computer implemented method according to any of the preceding claims, wherein the machine tool and the digital machine model are operable to extract and execute work step and / or process information by a parser.

13. A computer-implemented method according to any of the preceding claims, wherein the structure data (G3) is output as an XML file or as a STEP file.

14. 1. An apparatus for controlling a machining process of a workpiece by a machine tool in response to NC data and / or PLC data, comprising: - said machine tool for machining said workpiece using specified NC and / or PLC data; and - a processor, the processor a simulation device which can be controlled independently of the machine tool for simulating the machining process on a digital machine model based on the specified NC data and / or PLC data; said processor connected to a memory storing instructions for acting as an analysis unit connected to said machine tool and to said simulation device for adapting simulation parameters in said simulation device, The machine tool is configured to transmit machining data (R1) of the machining process on the machine tool to the analysis unit, and the simulation device is configured to transmit simulation data (R2) of the machining process simulated on the digital machine model to the analysis unit, the analysis unit (AA) is configured to link the transmitted machining data (R1) and simulation data (R2) and, based on the linked data, to cause a machine learning device arranged in the analysis unit to learn the behavior of the machine tool and output a result of an analysis of the machining process; The device is characterized in that it links the simulation data to the machining data by performing continuous data mapping based on the association of the positions of the operations performed, NC lines and / or axes, and associates sensor data of the machine tool with simulation data obtained by simulation in a simulation section (SA).

15. the machine tool, the simulation device and the analysis unit are configured to transmit data, in particular parameter and / or performance data and / or hardware and / or program data, to one another, The apparatus of claim 14 , wherein the data is transmitted over an intranet and / or over the Internet.

16. the simulation device and the analysis unit are independent of the machining process on the machine tool; 16. The apparatus according to claim 14 or 15, wherein the analysis unit is configured to continuously match the machining process on the digital machine model with the machining process of the machine tool by transmitting an output dataset (F1) of the machine learning device to the simulation device.

Images