Computer-aided method and device for generating a digital representation of a technical structure and corresponding computer program product

Abstract

The invention relates to a computer-aided method (1) and a device (C1) for generating a digital representation (17) of a technical structure, in particular an electric motor, and a corresponding computer program product (1, C2). In the method (1), a domain-specific model based on digital geometric data of a component (2, 18, 19, 20, 21, 22) is provided. Thus, a model reduction is performed, wherein the domain-specific model is converted into modal coordinates and the spectral behavior of the component (2, 18, 19, 20, 21, 22) is determined by means of a corresponding modal analysis. Based on this, a corresponding state space representation is generated as a reduced-order spectral model. In order to simulate the technical structure as a whole, the spectral models are coupled to each other to form a digital representation (17) for simulation, which describes the behavior of the technical structure across domains.

Description

Technical Field

[0001] The invention relates to a computer-aided method and device for generating a digital representation of a technical structure and to a corresponding computer program product. Background Art

[0002] Computer-aided methods, such as precise component modeling based on finite elements and other methods summarized in CAE (English: Computer-aided Engineering), have been used in modern product development. Analysis based on this can provide precise insights into the characteristics of the corresponding modeled components and thus accelerate development. However, the disadvantage is that the corresponding data is often not accessible and understandable, and, for example, finite element experts (FE: finite element) are retained. In addition, the computing hardware that is actually available today and in the foreseeable future cannot fully model and simulate complex systems within a reasonable time, because this involves too much computing effort and complexity.

[0003] Therefore, up to now, people have often limited themselves to the individual modeling of individual components and / or replaced complex systems or structures with completely simplified representations, which can only provide very rough basic statements and results. Thus, taking the electric motor as an example, the complex structural dynamics of the motor housing itself can be simplified to two masses and two springs as a simplified starting model for the rotor dynamics, or the vibration characteristics of the rotor can be roughly approximated as a simplified basis for electrical calculations, in which only the rotor mass moment of inertia is taken into account.

[0004] Therefore, no method to date has been able to obtain in a practicable manner detailed and reliable representations and knowledge about complex structures or systems consisting of several components that interact or come into contact with each other. This is precisely what is of interest for further improved and digitalized product development in the future. Summary of the invention

[0005] The object of the invention is to provide an improved possibility for digital processing of technical structures. According to the invention, this object is achieved by the subject matter of the independent claims. Advantageous embodiments and improvements of the invention are described in the dependent claims, the description and the drawings.

[0006] The method according to the invention serves to generate a digital representation of a predetermined technical structure, i.e. a digital model or digital twin. A technical structure in the sense of the invention may be, for example, an apparatus, a machine, an instrument, a system, a facility or the like, in particular consisting of a plurality of components or parts. An example of such a technical structure may be a high-voltage electric motor. Ultimately, however, the method according to the invention can be successfully used for almost any digital, i.e. computer-aided, representable or modeled structure.

[0007] In the method steps of the method according to the invention, domain-specific models of these components are provided or obtained based on digital three-dimensional geometric data of predetermined components of the technical structure. The three-dimensional geometric data indicate the construction or structure of the corresponding component. For this purpose, Cartesian coordinates can be used in particular, but cylindrical coordinates, spherical coordinates or the like can also be used. These coordinates, in which the three-dimensional geometric data or the domain-specific model is provided or exists, are referred to as primitive coordinates here. In other words, the geometry of the corresponding component can therefore be described by primitive coordinates, especially three-dimensional geometric data in Cartesian coordinates. The three-dimensional geometric data can exist, for example, in the form of a CAD data set (CAD: Computer-Aided Design), as a finite element model or the like. In the example of an electric motor (the present invention is not limited thereto), the components can be, for example, the housing, rotor, stator, cooling device, active electrical components and more of the electric motor.

[0008] Here, a domain-specific model is a model that models or describes the properties or behavior of a corresponding component at least primarily in a physical-technical domain. For example, these domains may be (classical) mechanics, electricity or electrodynamics, heat or thermodynamics and more similar, i.e., sub-domains or specialized fields of physics or technology, respectively. Compared to a complete model that attempts to fully model or describe the corresponding component in all domains or aspects, the corresponding domain-specific model has a significantly lower complexity. Therefore, domain-specific models can be advantageously generated and managed with correspondingly less effort.

[0009] Providing or retrieving domain-specific models may, for example, mean or include their computer-assisted or semi-automatic generation and / or their loading or retrieval from a corresponding model library of a provided data store or component.

[0010] In a further method step of the method according to the invention, a model reduction is performed based on the domain-specific model. Thus, the order or complexity required for calculating or numerically processing the model and thus the required computational effort is reduced. For this purpose, the domain-specific model is at least partially converted or transformed from its original coordinates, in particular Cartesian coordinates, into modal coordinates. In this case, these modal coordinates are a set of minimal coordinates, so that particularly efficient and labor-saving calculations or simulations can advantageously be achieved.

[0011] The spectral behavior of the component is determined using these modal coordinates with the aid of a corresponding modal analysis. In other words, a numerical characterization of the dynamics of the component is achieved here by means of a modal analysis, wherein these systems and / or components, which are understood to be vibrating, are assigned decay times or time constants as the basis for the modal analysis. The dynamics or spectral behavior of the component can be characterized or described, for example, by modal parameters, natural frequencies, characteristic forms or natural vibration types, modal masses, if appropriate modal damping and / or the like, which are determined within the scope of a corresponding modal analysis.

[0012] Based on this, a state space representation (ZRD) is then generated for the component as a reduced-order spectral model of the component. In other words, the component or its behavior or properties are described or characterized by the spectral model in the corresponding state space representation. Compared to the original domain-specific model or the corresponding finite element model of the individual components, the model reduction can already be derived here, that is, only the modes of the component as a whole, that is, the corresponding vibration states, are calculated or modeled, and no longer the individual movements or properties of all parts or regions of the corresponding component in all three Cartesian space directions. The spectral model is therefore able to describe the dynamic or vibration properties of the component in a simpler or more compact way than the domain-specific model or the complete finite element model of the component.

[0013] In another method step of the method according to the invention, in order to simulate the technical structure, the spectral models of the components are generally coupled to each other or connected to each other to form a digital representation, which describes the behavior of the technical structure across domains. In other words, the digital representation of the technical structure can therefore be generated (similar to assembling the corresponding real technical structure from the individual components) by coupling the spectral models for the individual components. Because the spectral model is based on domain-specific models, the model can in particular describe components in different domains, which leads to the characterization, description or modeling of the entire technical structure in the corresponding different domains. Here, a component can be described by a spectral model, or the characteristics or behavior of a component in different domains can be described by multiple spectral models of the component.

[0014] By coupling the individual spectral models to each other, it is advantageously possible to model or simulate the interaction between components, especially across different domains, and to obtain corresponding knowledge of the performance or characteristics of the entire technical structure, which cannot be obtained from the domain-specific or finite element models of the individual components that have been considered separately so far. Due to the model reduction, the computational work for such modeling and simulation of the entire technical structure is advantageously still feasible even with the computing hardware widely available today. In addition, the method according to the invention provides further options for accelerating and reducing the required computational workload, which will be explained in more detail below. Due to the reduced required computational workload, the present invention can still provide more detailed, more accurate and more reliable results and insights into the performance of the entire technical structure, rather than obtaining some basic idealized elements through a very rough approximation of the components described at the beginning. Therefore, the present invention makes it possible, for example, to identify potential problems or weaknesses in the design of the technical structure in a purely digital and computer-aided manner with improved accuracy and reliability in the scope of the development or design process, and to better adjust the design of the components to the loads that actually occur. Therefore, in the end, this can improve the reliability of the real technical structure, and at the same time, if necessary, can also save material.

[0015] The method according to the invention can be performed, in particular automatically or semi-automatically, by means of a correspondingly arranged data processing device, for example, a device for data processing according to the invention. For example, a conventional computer, a conventional workstation, a mainframe computer or the like can be used as such a data processing device. Therefore, the data processing device can in particular include a processor, a microchip, an integrated circuit, a hardware circuit or the like for executing a computer program or program code encoding or representing the method. In addition, the data processing device can in particular have a volatile and / or non-volatile data memory connected thereto, one or more interfaces for receiving and outputting data and / or more of the like.

[0016] The data processing device can have one or more sub-devices for carrying out the individual method steps of the method according to the invention. This can be, for example, an acquisition or provision device for acquiring or providing domain-specific models, a model reduction device for carrying out model reduction and the processes described in connection therewith, and a coupling device for generating digital representations of technical structures by coupling spectral models to one another. Similarly, a simulation device can be provided for simulating technical structures based on the digital representations generated in this way. These sub-devices can be or include corresponding hardware circuits or hardware modules, but also corresponding program parts or program modules of the named computer program, i.e., for example, an operating program for the data processing device.

[0017] In an advantageous embodiment of the invention, the mass matrix M and the stiffness matrix K of the modal decoupling are respectively determined in the corresponding modal analysis for the component. In general, the corresponding damping matrix D can also be determined here. Based on the equation of motion for the corresponding component, a new arrangement is then achieved by means of a state space representation, i.e. in the form of a state space representation. In this case, the differential equation of motion (short for equation of motion) can be expressed for a component with one degree of freedom as follows:

[0018]

[0019] Where s is the position vector and F is the external force. For the special case of modal analysis, F = 0 can be applied or set. The corresponding state space representation can be expressed as:

[0020] And y = C·x + D·u

[0021] Where t is time, x is the position vector, u is the input vector, y is the output vector, A is the system matrix, B is the input matrix, C is the output matrix and D is the transmission matrix. The matrices for the state space representation are given here in normal representation:

[0022] And C = [φ 0].

[0023] The transmission matrix D can be a zero matrix here, i.e. it is omitted if the technical structure or the corresponding component is a mechanical system, since mechanical systems cannot jump in position and speed. In many technical structures or for many components of technical structures, such as structural components of electric motors, it can be assumed that the damping is so weak that it can be ignored, so that the damping matrix D of the equation of motion can also be omitted. By combining and rearranging the parameters, the special case of negligible damping is then obtained:

[0024] And y = [φ 0]·q = C·q

[0025] The following are provided: state vector q and eigenform matrix φ. Therefore, the modal analysis can be performed here as an undamped modal analysis, which advantageously results in simplified calculations and a correspondingly reduced workload.

[0026] With the method proposed here, the method according to the invention, namely the digitization and modeling or simulation of technical structures, can be carried out particularly efficiently and with a practicable and manageable computing effort.

[0027] In another advantageous design of the present invention, the corresponding modal analysis is performed only for a predetermined lower frequency range and / or only for a predetermined number of the highest energy modes. For example, the modal analysis can be limited to a frequency range between 0 Hz and 2 kHz, or between 0 Hz and 1 kHz, or between 0 Hz and 500 Hz. Additionally or alternatively, the modal analysis can be limited to relatively high energy modes. For this purpose, the modes can be classified, for example, according to their energy content or their contribution to the total energy of the vibration or the dynamics of the corresponding component. Among them, only the n modes with the highest energy can be considered or used for further calculations, wherein n can be a predetermined number or a predetermined value such as a percentage or ratio. Similarly, for example, an assumption or estimate can be performed or predetermined according to the type, basic structure or a predetermined reference or comparison value, in which frequency range the highest energy mode of the corresponding component may be, and the modal analysis is limited to this. By limiting the modal analysis proposed here, the required computational workload can be further reduced. However, according to experience, the specific limitations provided here can still be used to obtain particularly relevant statements and understandings of the performance and characteristics of the component and the entire technical structure.

[0028] In another advantageous embodiment of the invention, for the model reduction, after the modal analysis, the transformation from Cartesian coordinates to modal coordinates and / or the reverse transformation from modal coordinates to Cartesian coordinates is performed only for a predetermined selection of discrete observation points, wherein the external influences of the technical structure are determined as a whole during the simulation at these discrete observation points. For example, depending on the complexity of the technical structure, up to 1000 or up to 100 observation points can be predetermined or used. These observation points can in particular be nodes, connection points or attachment points of components, storage or support points or the like. It has been shown that by such a restriction to certain observation points that are important for the behavior of the technical structure and its properties, data that is usually needed or required today can be obtained, for example for designing components and / or other components in a loading environment or in contact with the technical structure. In this case, the computational effort or the amount of data to be processed can be reduced by several orders of magnitude compared to the modeling and simulation of the complete (e.g., based on finite elements) technical structure. Due to the structure of the modeling method according to the invention proposed here, the accuracy of the results of the properties of the technical structure at the observation points can still be within the range of accuracy that can also be achieved using a complete finite element analysis, for example, a maximum deviation of 5% or a maximum deviation of 10%. Therefore, the restriction to predetermined observation points can provide a particularly effective and efficient contribution to model order reduction.

[0029] At a predetermined observation point, a bidirectional connection can finally be established between the description of the internal (i.e. model-internal) modes of a component or technical structure and the external Cartesian description. In the present case, this connection is generated with the aid of or based on the corresponding characteristic forms. These are part of the modal analysis, just like the corresponding natural frequencies, and can therefore be used anyway, and do not generate any further or additional computational effort. Therefore, not only can the amount of data to be processed in the method according to the invention itself be reduced accordingly, but also, for example, the data derived from it or obtained on this basis, which can be used by other applications or programs, can be reduced accordingly. The design of the invention proposed here is based on the recognition that for many steps, such as in the development, manufacture, installation, evaluation and testing of technical structures, only a limited number of core information is required, rather than a complete global knowledge of the detailed characteristics of the corresponding technical structure at all points or in all domains. For example, in the case of an electric motor, its characteristics at the interface with the loader, at the power connection and at the support or foundation may be particularly important. Compared to previous methods, the method according to the invention makes it possible to obtain detailed and reliable information about the properties of these or corresponding observation points with improved accuracy or at practically manageable costs, even for complex technical structures consisting of multiple components coupled to one another.

[0030] In an advantageous improvement of the invention, observation points in the three-dimensional geometric data or the corresponding finite element model of the component are automatically determined according to the category of the corresponding component. The category can be predetermined, for example, in a table or database or as a parameter or in metadata for the corresponding geometric data, the corresponding finite element model or the corresponding component. The category can indicate the kind or type of the component. For example, the component can be classified as or a housing, a radiator, a bearing, an axle or shaft, a rotor, an active electrical component, a foundation and / or more similar. For example, it should be possible to predetermine which point or points will be used as observation points for each category. As described, these can be, for example, bearings or support points, nodes, connection points or attachment points, but can also be extreme points, such as points or areas of minimum diameter or minimum material thickness, the lowest point, the highest point or the center or center of gravity or the like. These types of observation points can then be automatically determined for the corresponding components in the geometric data or finite element model. In this way, a more extensive automation of the method according to the invention can be achieved, thereby further minimizing the effort that must be paid to obtain useful results or knowledge.

[0031] Likewise, some or all of the observation points can be manually predetermined or determined or adjusted. In any case, once the observation points have been determined automatically or manually, further measures or steps can be automatically performed based on them. The scanning rate or sampling rate or frequency for the modal analysis or simulation, the number of characteristic forms or modes to be determined, the frequency range to be considered or analyzed, the coordinate system to be used, the number and type of degrees of freedom to be considered and / or more of the like can then be automatically determined or selected, for example, depending on the category of the corresponding component and / or the number and / or type of observation points determined. For this purpose, corresponding regulations and / or allocation tables can be predetermined.

[0032] In another advantageous design of the present invention, when generating a digital representation of the technical structure, an idealized coupling mechanism is used to connect at least some spectral models of the components to each other and / or to other predetermined elements. Such an idealized coupling element can be, for example, an ideal spring, a damper, a mass or a mass vibrator or the like. Such an idealized coupling element can represent or model specific basic characteristics, respectively, in a mathematical or mechanical idealized manner, that is, in a manner abstracted from real components. Therefore, the idealized coupling mechanism and the corresponding connection of the components or other elements can be modeled and calculated in a particularly simple manner. However, since the components themselves connected to each other in this way are modeled with higher accuracy, the loads occurring at the ends or connection points of the idealized coupling mechanism can accordingly accurately reflect the performance of the corresponding actual technical structure, so that despite the use of an idealized coupling mechanism, sufficient accuracy can ultimately be obtained for practical purposes.

[0033] The other elements mentioned can, for example, represent bearing stiffness, friction effects, in particular air friction of moving components and / or more similar things. Therefore, other elements can especially represent effects or influences that cannot be easily modeled or represented in the form of three-dimensional geometric data or as independent finite element models. In particular, nonlinearities can be modeled or simulated by additional elements, which can be presented at the simulation level of a corresponding more comprehensive system composed of components and other elements. Such nonlinearities can occur, for example, in sliding bearings, in the air gap of high-voltage motors and / or at other points. Other elements can also be idealized or abstracted, or predetermined by a single model (for example based on causal or physical modeling).

[0034] In another advantageous embodiment of the present invention, at least one of the spectrum models of the components is also connected to at least one predetermined digital environment component, which represents the environment of the planned technical structure in actual use. Thus, a digital representation of the facility including the technical structure is generated for the simulation. For example, the environment component can be or represent a technical structure that forms a facility or is part of a facility together with the original technical structure. Examples of such environment components can be or represent a foundation supporting the technical structure, a power connection or power supply device, via which power or energy is provided to the technical structure, a loader operated or driven by the technical structure, a support for the technical structure, a machine room in which the technical structure is arranged, and / or more similar things. The environment component can therefore represent or map the corresponding actual use and / or working environment for the technical structure. This is particularly advantageous because the performance of the technical structure can be affected by its environment and / or knowledge can be obtained in this way, based on which the arrangement or design of the environment can be optimized. Taking into account the corresponding environment, it is advantageous to achieve individual adjustments or evaluations for different application scenarios of the technical structure. This was not feasible before, because on the one hand a complete finite element analysis would result in an unmanageable workload, and on the other hand, the greatly simplified representation of the technical structure mentioned at the beginning could not provide or achieve sufficiently accurate data and interaction only through some idealized elements.

[0035] In an advantageous refinement of the invention, at least one environmental component is initially represented by at least one idealized element. In the iterative refinement of the digital representation of the facility, the idealized element is then replaced by a finite element model (FE model) of the environmental component or a reduced-order spectral model derived therefrom. In other words, the digital representation of the facility can be gradually adapted and improved or designed in more detail, for example once a corresponding detailed model of at least one environmental component becomes available during the corresponding development or design process. A particular advantage of the invention is that this is easily achieved due to the modular structure of the digital representation. Since the environmental component is initially represented by an idealized element, the initial computational effort can be minimized, so that first results can be obtained particularly quickly and at an early stage. Thus, the efficiency and parallelism of the development and design process for a technical structure or facility can be improved. Idealized elements here refer to descriptions based on idealized assumptions that ignore complex effects occurring in reality, for example, known from the simplest models and theories of classical mechanics, electrodynamics and thermodynamics.

[0036] Another aspect of the invention is a computer program product comprising command or control instructions, which, when executed by a computer, in particular the device for data processing according to the invention or the data processing device mentioned in conjunction with the method according to the invention, causes the computer to perform at least one variant of the method according to the invention, in particular automatically or semi-automatically. The computer program product according to the invention can be a computer program. Accordingly, the method according to the invention can be fully or partially computer-implemented or computer-implementable, i.e. encoded or represented by such a computer program or a corresponding program code. Likewise, the computer program product according to the invention can be a computer-readable data carrier on which the corresponding computer program is stored.

[0037] Another aspect of the present invention is the already mentioned device for data processing. The device has a mechanism for executing at least one variant of the method according to the invention, in particular automatically or semi-automatically. These mechanisms can in particular be the mechanisms described in conjunction with the mentioned data processing equipment. Therefore, the device according to the present invention can in particular have a corresponding processor, a data memory and at least one input and / or output interface for executing the method or a corresponding computer program or program code. The device according to the present invention can in particular include a computer program product according to the present invention. Since the device according to the present invention can be particularly configured to execute the method according to the present invention and therefore the method according to the present invention can be executed by means of the device according to the present invention, the device according to the present invention can accordingly have some or all of the characteristics and / or features described in conjunction with the method according to the present invention.

[0038] The present invention also includes improvements of various aspects of the present invention, i.e., the method according to the present invention, the computer program product according to the present invention, and the device according to the present invention, which have features described only in conjunction with one or some aspects of the present invention. In order to avoid unnecessary redundancy, corresponding improvements of the present invention or its various aspects are no longer described separately for all these aspects.

[0039] The invention also includes combinations of features of the described embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The embodiments of the present invention are described below. As shown in the figure:

[0041] Figure 1 A schematic diagram of an apparatus for performing a method for providing a digital representation of a technical structure is shown, wherein the method is represented by an exemplary flow chart;

[0042] Figure 2 A schematic diagram for illustrating a first description of an assembly is shown;

[0043] Figure 3A schematic diagram for illustrating a second description of the components is shown; and

[0044] Figure 4 A schematic diagram is shown for explaining the numerical representation. DETAILED DESCRIPTION

[0045] The embodiments explained below are preferred embodiments of the present invention. In the embodiments, the components of the described embodiments represent the various features of the present invention respectively, and these features are considered independently of each other and further improve the present invention independently of each other. Therefore, the present disclosure should also include combinations other than the illustrated features of the embodiments. In addition, the described embodiments can also be supplemented by other features of the present invention that have been described.

[0046] For various reasons, for example to support product development and testing, it is desirable and advantageous to have a digital image or a digital representation of the technical structure available. In the following, it is assumed purely by way of example that a high-voltage motor (HV motor) consists of multiple components and parts and therefore represents a complex technical structure. For the design of the R&D part, there are only two digital alternatives, namely computer-aided consideration of such complex technical structures, in which multiple components, which each mainly act on different physical-technical domains, are coupled to one another. This involves firstly complex multi-physics simulations and secondly the consideration of the coupling of the entire technical structure is neglected. Due to the resources required, the former is currently not feasible, so that in practice a correct simulation of the entire system is usually not required. The same applies, for example, to the preventive analysis of the entire drive system.

[0047] It should also be noted that complex technical structures can often be supplied in different configurations depending on requirements or customer wishes. This means that even if analyses or simulations are performed for extreme values, such as the heaviest, lightest, longest, shortest, etc. configuration or variant, these often do not yield reliable information or an optimized design for a customer-specific variant or configuration that lies in the middle of the corresponding parameter field and may have unrecognizable unfavorable combinations or interactions between individual components.

[0048] Instead of a complete (i.e. physically at least substantially correct) consideration of the entire system (i.e. the entire technical structure), the simulation and the corresponding design of the individual components of the corresponding technical structure can currently be performed individually and domain-specifically. This means, therefore, that the simulation or computer-aided checking or analysis is limited to a single component and a single domain, respectively, and that the individual components as well as the entire technical structure are designed only in a single domain without a representation of the overall system. Furthermore, in today's established processes, even the corresponding simulation data of the individual components are usually only made available to experts, for example in specialized programs for finite element modeling (FE modeling) and are therefore not available to other departments (e.g. engineering or sales) or to the corresponding customers or users or end consumers. However, it can be seen from the requests for digitized images of transient simulations of complete technical structures that there is a corresponding need.

[0049] For this purpose, these are often reduced to very rough information, rather than providing correspondingly accurate simulation data or models. For example, the complex structural dynamics of the motor housing have been simplified to two masses and two springs in order to be able to be combined with the rotor dynamics. For example, for the electrical calculation, the vibration characteristics of the rotor can be very roughly approximated by only considering the rotor mass moments of inertia. In addition, the individual components are not fully simulated here, but are represented by equivalent circuit diagrams or simple geometrical basic shapes or abstract mass points.

[0050] In order to solve these difficulties, problems and deficiencies, a modeling method is proposed here, by which core information can be extracted from the detailed analysis of each domain and combined with each other across domains to achieve a universal overall system consideration with higher accuracy. Therefore, this advantageously has the possibility of simulating specific variants or configurations of the corresponding technical structure across domains. Here, the corresponding specific variant or configuration is composed of a digital combination of individual domain-specific models of the components or main components of the corresponding technical structure.

[0051] Advantageously, environmental knowledge of the respective technical structure can also be taken into account in this case in order to ultimately simulate how the respective specific configuration of the technical structure behaves in a specific intended use or operating environment.

[0052] The individual components are modeled domain-specifically, in particular based on the corresponding digital three-dimensional geometric data provided. If necessary, multiple domain-specific models of different domains can be generated or provided for a single component, each of which can be based on the same three-dimensional geometry of the component. In this case, geometric dimensions or corresponding three-dimensional geometric data according to the design or according to the specifications can be used, so that the components are not simplified or abstracted according to their geometry or according to their structure, so that an improvement in the accuracy of the result can ultimately be achieved. Environmental components or environmental elements, i.e. elements or parts of an environment or those representing an environment, can be provided or represented in the same way or, if necessary, by simplified or idealized elements.

[0053] With the technology presented here, the relevant spectral behavior of the corresponding technical structure can be extracted from the various domains, so that previously unmanageable complete simulation calculation models (which, for example, can include 10 million equations just for the structural dynamics of the motor housing) can be effectively reduced to relatively few, for example, 10 to 100 or at most 1000 equations can be packed. By the type of extraction and processing, it is even possible to advantageously enable the corresponding data or models to be passed on to third parties, for example to external companies or customers, without disclosing knowledge related to development or production, i.e. complying with confidentiality requirements. Thus, a third party can receive a digital image of the corresponding technical structure in a specifically requested configuration for the transient calculation of the relevant parameters, but no explicit details about the inner workings or internal structure of the technical structure can be inferred from it.

[0054] In this regard, Figure 1 The schematic diagram of a data processing device C1 (e.g. a computer) for executing a method for providing a digital representation of a technical structure is shown. The data processing device C1 has a data memory C2, a processor C3 connected to the data memory, and an interface C4 connected to the processor for receiving and outputting data. A computer program for executing the method is stored in the data memory C2. The corresponding program module or method sequence is represented by a flow chart with method steps S1 to S10. Figure 1 In the following, the process will be described with reference to other figures. Figure 1 Explain the method in more detail.

[0055] The method is mainly used to generate a digital twin of the corresponding technical structure (in this case, a high-voltage motor) and provide a corresponding workflow. In method step S1, the digital three-dimensional geometric data of the individual components of the high-voltage motor are first exported from the development environment, here from the PLM system (PLM: Product Lifecycle Management). Here, a neutral, i.e., manufacturer-independent, standardized and not associated with a specific application file format is used, such as STEP (Standard for the Exchange of Product Model Data). Here, STEP advantageously achieves a particularly small amount of data and can provide particularly good conversion results based on experience.

[0056] In method step S2, the geometric data derived in this way, i.e. the corresponding data set, for example three-dimensional CAD data, are imported into a program for finite element modeling or finite element simulation, such as "Ansys Mechanical". In this case, the data set is processed in a CAE-compliant (CAE: Computer-Aided Engineering) manner. For this purpose, elements that are not part of the basic structure of the corresponding component and are therefore irrelevant or less relevant to the performance of the corresponding component can be removed in order to reduce the amount of data. These can be, for example, elements of the component itself, such as holes for fixing other elements, lubricant channels, cable brackets, etc., or other elements contained in the imported data set and fixed to the actual component. This CAE-compliant processing can already be regarded as a first part of the model reduction, thereby further reducing or limiting the amount of data to be processed and the corresponding amount of calculation or simulation, so that the method described here can be carried out with conventionally available computing hardware, such as a common work computer or workstation, and without the need to use a mainframe or supercomputer.

[0057] In a finite element program, the geometry of a component is described in primitive coordinates. Typically, these can be Cartesian coordinates, whereby there are three systems of equations for the height, depth and width of the corresponding component.

[0058] In method step S3, the correspondingly obtained analysis basis, ultimately a set of geometric or finite element data or a finite element model, is transformed from original coordinates, in particular Cartesian coordinates, into minimum modal coordinates. Preferably, this is performed only for a predetermined set of discrete points, which constitutes a relatively small part of all points of the finite element model. This represents a further part of the model reduction, by which subsequent calculations or simulations are simplified.

[0059] In method step S4, a corresponding modal analysis is performed on the individual components according to the domain-specific model in order to determine the corresponding spectral behavior of the components. The modal analysis is limited to a specified set of points, i.e., specified observation points 27 (see Figure 4), at which the behavior of the corresponding components is particularly relevant in practice. For example, these observation points can be nodes, connection points or coupling points to the environment or to other components, sites, extreme points of the components, midpoints or points marked in some other way. In addition, it is preferred to consider only relatively low-frequency but relatively high-energy modes, for example frequencies of up to 500 Hz. These restrictions on the basic data of the basic points can reduce the amount of data to be processed, for example by about five orders of magnitude, i.e. by about 99.999%. Therefore, the results finally obtained in the selected observation points 27 can still be roughly as accurate as a complete finite element analysis. The limitations of the modal analysis described describe an important part of the model reduction.

[0060] In the present case, an undamped modal analysis is performed, in which the corresponding differential equations of motion are included or used for the basic form of the component. In the modal analysis, the corresponding modally decoupled mass matrix M and stiffness matrix K are calculated for the component. These are converted into a state space representation (ZRD) and rearranged in method step S5. The simplification for the degrees of freedom can be expressed as:

[0061] And y = C·q, where:

[0062] And C = [φ 0]

[0063] The diagonal matrix of position vector x, external force F, time variable t, state vector q, system matrix A, input matrix B, output matrix C, output vector y, natural frequency ω and characteristic form matrix φ

[0064] Here, the transmission matrix sometimes used in the state space representation can be a zero matrix and can therefore be omitted, since mechanical systems cannot jump in position and velocity and the technical structures considered here can also always be mechanical systems. Likewise, a negligibly small damping is assumed here by way of example, so that the damping matrix that would otherwise appear here is also zero.

[0065] Figure 2 A schematic diagram for illustrating a first description of a component 2 corresponding to a differential equation of motion is shown. Here, the component 2 is schematically represented by a mass 3 and a support 4 and their connection by means of a spring 5 and a damper. The behavior of the component 2 is characterized by a position vector 7, which is also schematically represented here, corresponding to the variable x. In the present case, the damping matrix can be omitted in the differential equation of motion for the component 2, since, for example, the structural components of a high-voltage motor are only weakly damped anyway. In this way, the limitations of finite element solvers can also be circumvented, since damped modal analysis is currently not feasible.

[0066] Figure 3A schematic diagram for illustrating a second description of the assembly 2 is shown, corresponding to the normal representation of the state space representation. Here, a time-dependent input vector 8 corresponding to the variable u is combined with an input matrix 9 corresponding to the variable B. The result thus produced is added to a first combination element 10 with a system matrix 11 corresponding to the variable A. The result obtained is combined with an attenuation element 12 so that a time constant or a proportionality to 1 / second, i.e. a frequency, is introduced. The result obtained is combined with an output matrix 13 corresponding to the variable C. The corresponding result is combined additively in a second combination element 14 with a transmission matrix 15 corresponding to the variable D, as long as it is not omitted. Finally, the result is a time-dependent output vector 16 corresponding to the variable y.

[0067] By combining the differential equations of motion and ZRD, it corresponds to Figure 2 and Figure 3 A combination of the descriptions of and a rearrangement of the variables, ignoring damping, yields:

[0068] as well as

[0069] y=[φ 0]·q=C·q.

[0070] This can be understood as the mathematical basis of the method described here.

[0071] Modal analysis and model reduction are performed for all components of the corresponding technical structure, here for all components of the high-voltage motor. Here, the corresponding natural forms are used for the bidirectional connection of the internal modal description and the external Cartesian description. Like the corresponding natural frequencies, they are an essential part of the modal analysis and therefore no further calculation work is required. The transformation of Cartesian coordinates and modal coordinates (modes) takes place only at the observation point, so that ultimately only the external influences of the corresponding component or the corresponding technical structure or system are calculated here. The calculation is carried out via the characteristic form matrix φ. The modes represent k states of the corresponding system, which are summarized in the state vector q(t). If present, these can be rigid modes, bending modes, torsion modes and / or many similar ones.

[0072] The system excitation corresponding to the input vector u(t) is decomposed here into the excitations of k states q(t) in the following way:

[0073] u i (t) = ∑ i，k θ i，k ,q k (t),

[0074] where the index i represents the corresponding excitation (input, position and degree of freedom), and θ i，kThe element of the feature form matrix φ is specified as the weight element of the corresponding state. The response y of each defined output j j (t) - is the combination of all outputs j in the output vector y - which is weighted by the state q(t) and the feature form factor θ j，k Composition:

[0075] y j (t)=Σ j，k θ j，k ·q k (f).

[0076] In this way, the output y can be calculated for each defined observation point 27 and degree of freedom j (t). The state space representation is fully described via the A, B and C matrices. In this way, a reduced order model can be created overall.

[0077] By means of the method described here, including the transformation from Cartesian coordinates to modal coordinates and the restriction to predetermined observation points 27, a significant model reduction can be achieved and opens up a huge potential for accelerating calculations and simulations.

[0078] The results of method step S5, i.e., the corresponding state space matrices, are preferably imported into the system simulation, i.e., the corresponding software or corresponding programs, such as "Ansys Twin Builder", in method step S6 in a neutral format. Here, the reduced-order models of the individual components, i.e., the so-called component twins, are assembled into a digital representation of the entire technical structure. In this case, the component twins are connected to one another, wherein, if necessary, further elements can be added and the combination of components can be used in its planned target application, i.e., the corresponding application or working environment. For this purpose, for example, the environment, boundaries and initial conditions can be defined and modeled. If necessary, further elements that can be added can, for example, represent more bearing stiffness, represent air friction and / or the like. Ultimately, almost any other elements can be inserted as long as they can be described mathematically. This applies in particular to loads or couplings or the like with different characteristic curves. Other physical effects or domains can also be added using this method. Ultimately, the result obtained here is a digital representation, i.e., an overall model or a spectral image or an image of the entire technical structure, including its surroundings if necessary, which is also referred to as a digital twin.

[0079] In this regard, Figure 4A schematic diagram is shown for illustrating such a digital representation in the form of a digital twin 17 of a high-voltage motor, including some environmental components, i.e. a digital twin of the corresponding overall facility. The main components of the high-voltage motor are here a housing 18, a cooling element 19 connected thereto, a stator 20, a rotor 21 and an active electrical component 22. The stator 20 and the rotor 21 are installed in the housing 18, for example, and are therefore at least indirectly mechanically connected thereto. Other components connected to the high-voltage motor may be, for example, an electric power regulator (represented here by a power supply network or a corresponding power supply interface 23), a foundation 24 (the housing 18 should be placed and installed on this foundation), a coupling not shown in detail here, and a load 25 (the load is schematically connected to the high-voltage motor via a loader 26).

[0080] The connections between these components shown here represent the interaction or interaction or coupling between the components. These connections can be represented or modeled, for example, so that the power, load or influence occurring at the component is used as an input variable for modeling or simulating other components connected thereto. The connection can also be represented by other separate elements. For this purpose, for example, an ideal transmission mechanism, spring, damper and / or the like can be inserted into the digital twin 17. Therefore, the high-voltage motor can eventually be simulated in all modeling domains in its planned subsequent application environment, especially in a particularly time-saving manner and with higher accuracy compared to the methods actually used in the traditional way. This is achieved by a combination of different measures and methods, which have proven to be particularly effective. Modeling of the environment (here, for example, represented by the grid interface 23, the foundation 24 and the loader 26) is a new feature in particular, which represents the special advantages of the digital twin 17 or the method proposed here relative to conventional methods. Therefore, this provides the opportunity to model the environment and any other components and the corresponding coupling and its corresponding spectrum performance. Therefore, this allows the performance of the high-voltage motor and its interaction with the environment, i.e., the overall performance of the digital twin 17, to be calculated or simulated.

[0081] Here, some observation points 27 are marked as examples, and the performance or external influence of the digital twin 17 at these observation points is actually simulated in method step S8. Here, for example, not at any point of the housing 18, but for example only at its position point or at the connection to the remaining components of the digital twin 17 and the performance of the digital twin 17 is calculated accordingly for the remaining components. The user or user of the digital twin 17 is usually mainly interested in some core information related to the corresponding user-specific application. This core information includes the results of the defined interface and the defined position, such as the observation point 27. The loader 26 represents such a particularly relevant interface. For example, here the values ​​of the speed and torque can be provided according to actual requirements or measurements and calculated or simulated according to the digital twin 17. Another particularly relevant interface is the power interface, such as the power interface 23 or the active electrical component 22. For example, in the case of a high-voltage motor, a three-phase alternating current with a defined voltage and frequency (DOL, direct online) from the power supply or a three-phase alternating current from an inverter power supply with regulated voltage and frequency for any speed (VSD, variable speed drive) can be provided or simulated here. Another interface, ie another observation point 27, which may be particularly relevant to the design in the context of product development, is the base 24 or its connection to the housing 18. In interaction with the components of the high-voltage motor, this defines, for example, the speed adjustment range.

[0082] These and, if necessary, further interfaces are defined in the digital twin 17 in the form of observation points 27. Others can be modeled subsequently as required. In the same way, monitoring points can or will be defined as a supplement or part of the observation points 27. Such monitoring points can be set at any point in the digital twin 17 in order to mine data or information there, i.e. to track the performance of the digital twin 17 within the framework of the simulation. This is possible in particular also at locations that are not easily accessible at the real counterpart. This technology can also be called virtual sensors or soft sensors, since it is a virtual, digital observation or analysis. Overall, this approach is particularly advantageous and offers particularly great potential for improving previous processes, since, for example, the position of rotating and stationary components relative to each other can be tracked without complex physical measurement processes.

[0083] In order to iteratively improve the simulation of the digital twin 17, in method step S9 (as long as they are available or as soon as they are available), detailed models (e.g. finite element models) of other components, such as the environment of the high-voltage motor or the coupling between components of the high-voltage motor and / or for other physical effects, are integrated into the digital twin 17. In this case, if necessary, previously used, less detailed models can be replaced. Accordingly, some or all of method steps S6 to S10 can be run iteratively multiple times, whereby increased efficiency, effectiveness and parallelism in the development process can be advantageously achieved.

[0084] In (optional) method step S10, for example, the results of the simulation can be compared with a real counterpart of the digital twin 17 in order to be able to evaluate the digital twin 17 or the method described here and / or to determine, for example, whether the real counterpart, i.e. the corresponding real plant, is operating within given specifications. The knowledge gained can then be used, if necessary, to improve the method or the component twin, to improve production, to improve the operating strategy and / or the like.

[0085] Finally, in addition to the results obtained such as the speed-torque curve, the method presented here also allows a physically coherent modeling, i.e. a more complete representation of the digital twin 17, including the couplings between the components, whereby this enables, for example, the determination of position vibrations due to unbalances or loads such as anchor bolts and / or more similar, which was not possible in previous component-specific individual considerations or individual models. Thus, parts such as anchor bolts can be designed to better adapt to the actual loads and valuable information can be obtained, such as the transmitted forces applied to the various components, the total forces (e.g. on the foundation 24), the vibration excitation of the high-voltage motor for further structural engineering building analysis and more similar.

[0086] In the method used here, a combination of two different modeling types can be used, namely causal modeling and physical modeling (conservative modeling). In causal modeling, defined input-output performance is given. For example, physical modeling can use the node potential method (Kirchhoff's law) to create a bidirectional connection between two nodes and between two components. In this sense, a node can be a connection of a reduced-order model of a component to the environment or to another component or its reduced-order model. In particular, simplified or idealized other elements, such as springs, masses or damping elements can also be integrated via physical modeling. Therefore, the intuitive applicability of the method described here can be advantageously enabled to be used by a wide range of professionals, such as engineers from different disciplines.

[0087] The central process of the method presented here can be divided into two parts. On the one hand, in the context of finite element analysis or finite element modeling, these are the generation of component twins, and on the other hand, in system simulation, the simulated use of the entire twin, i.e. the complete digital twin 17. For example, the "Ansys Mechanical" or "Ansys Twin Builder" programs are currently available for practical use for this purpose. This therefore allows the process or method to be automated at least to a large extent, for example by control via ACT (Application Compatibility Toolkit) or via IronPython (implementation of the Python programming language), i.e. by using the potential of modern programming.

[0088] Furthermore, in the context of process automation, the observation points 27 can be automatically set or generated, if necessary, for example according to the corresponding predetermined categories of the components. This can be advantageously realized, for example, according to geometric data in a finite element program. Here, the corresponding coordinate system, the degrees of freedom to be simulated, the boundary and retention conditions, the matching settings for the finite element solver and / or more of the like can also be advantageously defined or set automatically, and the simulation or modeling can be started automatically.

[0089] According to the predetermined specifications of the corresponding technical structure, the corresponding component twin can be retrieved from the corresponding model library provided if necessary, for example based on the corresponding machine-readable product name (MLFB). Here, the corresponding MLFB can indicate or specify the corresponding variant, configuration or design of the component or the entire technical structure. For example, the shaft height or the length number of the active component 22 or the like can be predetermined for the high-voltage motor via the MLFB. Then, all required components or their provided models can be automatically or semi-automatically imported into the corresponding simulation environment and connected with their adjacent components and / or other elements inserted in between, such as springs, dampers, transmission elements, etc. For example, the electromechanical energy conversion of the high-voltage motor can also be inserted as a corresponding component twin and, for example, the correct coefficients of the corresponding T equivalent circuit diagram based on the MLFB can be provided automatically or semi-automatically.

[0090] For example, previously available software tools can be used to calculate the rotor dynamics of high-voltage motors separately, determine the permissible speed adjustment range, or derive the system response of the housing 18 to excitations such as an adjacent diesel engine or an earthquake or the like based on structural dynamics. Separate electrical simulations can determine the characteristics of high-voltage motors using previous methods, but using an idealized two-mass oscillator and a very simplified rotor model as the mechanical basis instead of the actual three-dimensional geometric data of the rotor 21 in accordance with the specification.

[0091] However, if, for example, the structural dynamics influence the rotordynamic effects, these conventional software tools fail, since the corresponding complete analysis requires computing time and computing resources that are not easily manageable in practice. As a result, the interactions between components have not been considered so far, or at least not in a realistic way. However, such interactions are likely to become even more important in the future, i.e. of great significance, since, for example, simultaneous demands for material savings, performance improvements, efficiency and utilization of complex systems will lead to an intensive involvement of coupled effects in components.

[0092] With the method presented here, the entire physics of complex technical structures can also be modeled in a consistent manner across domains with finite element accuracy, at least for the ultimately relevant observation points 27, with the computational effort being limited due to the measures of model order reduction, which allows a complete simulation run of the digital twin 17 in a time frame of, for example, minutes instead of days or weeks. In this regard, despite the model order reduction, new insights can be obtained compared to previous methods, since not only individual components and domains are considered individually, but also, for example, characteristic patterns of multiple components and / or combinations from multiple domains in combination or in interaction with one another, i.e. can be simulated. Since the domains are modeled coherently, results can be generated from interactions across the respective domain boundaries.

[0093] Overall, the examples described here show how the digital processing of complex technical structures can be improved by appropriate generation of digital twins of the technical structures.

[0094] In summary, the present invention therefore relates to a computer-aided method 1 and a device C1 for generating a digital representation 17 of a technical structure, in particular an electric motor, as well as a corresponding computer program product 1, C2. In the method 1, a domain-specific model based on digital geometric data of components 2, 18, 19, 20, 21, 22 is provided. On this basis, a model reduction is performed, wherein the domain-specific model is converted into modal coordinates so that the spectral behavior of the components 2, 18, 19, 20, 21, 22 is determined with the help of the corresponding modes. Based on this, the corresponding state space representation is generated as a reduced-order spectral model. In order to simulate the technical structure as a whole, the spectral models are coupled to each other to form a digital representation 17 for simulation, which describes the behavior of the technical structure across domains.

Claims

1. A computer-aided method (1) for generating a digital representation (17) of a predetermined technical structure, wherein: - providing a domain-specific model of a predetermined component of the technical structure based on digital three-dimensional geometric data of said component, - performing model reduction according to the domain-specific model, wherein - transforming the domain-specific model into modal coordinates, - using the modal coordinates to determine the spectral behavior of the component by means of a corresponding modal analysis, and - generating a state space representation for said component as a reduced order spectral model based on said spectral representation, and - In order to simulate the technological structure as a whole, the spectral models of the components are coupled to each other to form the digital representation (17) for the simulation of the technological structure, which digital representation describes the behavior of the technological structure across domains.

2. The method (1) according to claim 1, characterized in that: A method is provided for generating a digital representation as a digital model or digital twin of the predetermined technical structure.

3. The method (1) according to claim 1 or 2, characterized in that: The predetermined technical structure is an electric motor.

4. The method (1) according to claim 3, characterized in that: In the example of an electric motor, the assembly includes at least one of a housing, a rotor, a stator, a cooling device, and active electrical components of the electric motor.

5. The method (1) according to claim 1 or 2, characterized in that: The domain-specific models respectively model or describe the characteristics or behaviors of each corresponding component in the physical technical domain.

6. The method (1) according to claim 1 or 2, characterized in that: The domain-specific models relate to mechanics, electricity, electrodynamics, heat and / or thermodynamics, wherein the domain-specific models have a significantly lower complexity than the complete model, so that they can be generated and processed with correspondingly lower effort.

7. The method (1) according to claim 1 or 2, characterized in that: The respective modal analysis is performed only for a predetermined frequency range and / or only for a predetermined number of modes with the highest energy.

8. The method (1) according to claim 7, characterized in that The predetermined frequency range is a frequency range between 0 Hz and 2 kHz.

9. The method (1) according to claim 1 or 2, characterized in that: For the model reduction, after the modal analysis, a transformation from the original coordinates of the domain-specific model to the modal coordinates and / or a reverse transformation from the modal coordinates to other coordinates is performed only for a predetermined selection of discrete observation points (27) at which external influences on the structure are determined as a whole during the simulation.

10. The method (1) according to claim 9, characterized in that The number of discrete observation points is at most 1000.

11. The method (1) according to claim 9, characterized in that: The other coordinates are the original coordinates.

12. The method (1) according to claim 9, characterized in that Based on the predetermined category of the corresponding component, an observation point in the three-dimensional geometric data of the component is automatically determined.

13. The method (1) according to claim 1 or 2, characterized in that: When generating the digital representation (17) of the technical structure, idealized coupling mechanisms are used to connect at least some of the spectral models of the components to each other and / or to other predetermined elements.

14. The method (1) according to claim 1 or 2, characterized in that: At least one of the spectral models of the components is also connected to at least one predetermined digitized environmental component (23, 24, 25, 26), and a digital representation (17) of the facility including the technical structure for the simulation is generated via this connection, the environmental component representing the environment of the technical structure in the planned actual application.

15. The method (1) according to claim 14, characterized in that At least one of the environmental components (23, 24, 25, 26) is first represented by at least one idealized element (3, 4, 5, 6), and during an iterative improvement of the digital representation (17) of the facility, the idealized element is replaced by a finite element model of the environmental component (23, 24, 25, 26) or a reduced-order spectral model derived from the finite element model.

16. A computer program product comprising instructions which, when executed by a computer (C1), cause the computer to automatically perform the method (1) according to any one of the preceding claims.

17. A device for data processing, wherein: The device is used for carrying out, automatically carrying out, the method (1) according to any one of claims 1 to 15.

Images