Simulation device, simulation method, and simulation program

Abstract

This acquisition unit (22) inputs, to each of a plurality of processing systems (21) that is simulated by a simulation target, a system model(31) representing the specifications of the simulation target, and an input waveform (32), which is time-series data to be given to the simulation target as input, and thereby acquires an output waveform (33), which is time-series data, from each of the plurality of processing systems (21). This classification unit (23) calculates the degrees of waveform similarity between the output waveforms (33) acquired respectively from the plurality of processing systems (21), and classifies the plurality of processing systems (21) into one or more groups on the basis of the degrees of waveform similarity.

Description

Simulation device, simulation method, and simulation program

[0001] The present disclosure relates to a technique for performing a simulation using an immature processing system.

[0002] In recent years, the industry has been promoting the spread of model-based systems engineering in order to ensure the quality of the entire system, such as products and services, while also improving development efficiency. Furthermore, model-based development, which mainly focuses on software simulation, has become established for the individual devices or equipment that make up the system.

[0003] It is being considered to apply simulation technology for devices or equipment to ensure the quality of the entire system and improve development efficiency. For example, it is being considered to model a system using the Modelica language and perform collaborative simulation using a processing system compliant with the standard technologies FMI and SSP (see Non-Patent Documents 1 and 2). FMI stands for Functional Mock-up Interface. SSP stands for System Structure and Parameterization. The processing system is also called a simulation tool, simulation engine, numerical solver, etc.

[0004] "FMI Utilization Guide Ver. 1.0.1" by the Society of Automotive Engineers of Japan, published November 2018. Open Simulation Platform, https: / / open-simulation-platform.github.io

[0005] However, the processing systems used in collaborative simulation are not yet fully mature. Depending on the processing system, even if the system can be modeled, the calculation results may differ or the calculation may fail. In addition, there are numerous processing systems available, making it difficult to evaluate the cost-effectiveness, which makes it difficult to select a processing system. The purpose of this disclosure is to make it possible to realize a configuration that can obtain reasonable calculation results using immature processing systems.

[0006] The simulation device according to the present disclosure includes an acquisition unit that acquires output waveforms, which are time-series data output from each of a plurality of processing systems that simulate the simulated object, by inputting a system model representing the specifications of the object to be simulated and input waveforms, which are time-series data to be given as input to the object to be simulated, into each of the plurality of processing systems; and a classification unit that calculates the waveform similarity of the output waveforms acquired from each of the plurality of processing systems by the acquisition unit, and classifies the plurality of processing systems into one or more groups based on the waveform similarity.

[0007] In this disclosure, the waveform similarity of the output waveforms is calculated, and multiple processing systems are classified into one or more groups based on the waveform similarity. By using the classification results to select an appropriate processing system and output waveform, it is possible to obtain reasonable calculation results.

[0008] 1 is a functional configuration diagram of a simulation apparatus 10 according to a first embodiment. 2 is a hardware configuration diagram of the simulation apparatus 10 according to the first embodiment. 3 is a flowchart showing the processing flow of the simulation apparatus 10 according to the first embodiment. 4 is an explanatory diagram of an input waveform 32 and an output waveform 33 according to the first embodiment. 5 is an explanatory diagram of a similarity calculation process according to the first embodiment. 6 is an explanatory diagram of a classification process, a group selection process, and a processing system selection process according to the first embodiment. 7 is a functional configuration diagram when the simulation apparatus 10 according to the first embodiment is applied to a HILS. 8 is a flowchart showing the processing flow of the simulation apparatus 10 according to a second embodiment. 9 is a flowchart showing the processing flow of the simulation apparatus 10 according to a third modification. 10 is a flowchart showing the processing flow of the simulation apparatus 10 according to a fourth modification. 11 is a flowchart showing the processing flow of the simulation apparatus 10 according to a fifth modification. 12 is a flowchart showing the processing flow of the simulation apparatus 10 according to a sixth modification.

[0009] First Embodiment *** Description of Configuration *** The functional configuration of a simulation device 10 according to the first embodiment will be described with reference to Fig. 1. The simulation device 10 includes a plurality of processing systems 21, an acquisition unit 22, a classification unit 23, and a selection unit 24. In Fig. 1, processing systems 21A to 21Z are shown as the plurality of processing systems 21, but the number of processing systems 21 may be any number as long as there is more than one.

[0010] Each processing system 21 is also referred to as a simulation tool, simulation engine, numerical solver, or the like. Each processing system 21 receives a system model 31 and an input waveform 32 as input, and performs a collaborative simulation by simulating the behavior of a target to be simulated. Each processing system 21 outputs an output waveform 33 as a result of the collaborative simulation. The acquisition unit 22 acquires the output waveform 33 output by each processing system 21. The classification unit 23 classifies each processing system 21 into one or more groups based on the output waveform 33. The selection unit 24 selects one of the one or more groups as a specific group. The selection unit 24 then selects a representative processing system 34 from the processing systems 21 classified into the specific group, and extracts the output waveform 33 acquired from the representative processing system 34 as a representative waveform 35.

[0011] The hardware configuration of the simulation device 10 according to the first embodiment will be described with reference to Fig. 2. The simulation device 10 is a computer. The simulation device 10 includes the following hardware components: a processor 11, a memory 12, a storage 13, and a communication interface 14. The processor 11 is connected to other hardware components via signal lines and controls the other hardware components.

[0012] The processor 11 is an IC that performs processing. IC stands for Integrated Circuit. Specific examples of the processor 11 include a CPU, a DSP, and a GPU. CPU stands for Central Processing Unit. DSP stands for Digital Signal Processor. GPU stands for Graphics Processing Unit.

[0013] The memory 12 is a storage device that temporarily stores data. Specific examples of the memory 12 include SRAM and DRAM. SRAM stands for Static Random Access Memory. DRAM stands for Dynamic Random Access Memory.

[0014] The storage 13 is a storage device that stores data. A specific example of the storage 13 is an HDD. HDD is an abbreviation for Hard Disk Drive. The storage 13 may also be a portable recording medium such as an SD (registered trademark) memory card, CompactFlash (registered trademark), NAND flash, a flexible disk, an optical disk, a compact disk, a Blu-ray (registered trademark) disk, or a DVD. SD is an abbreviation for Secure Digital. DVD is an abbreviation for Digital Versatile Disk.

[0015] The communication interface 14 is an interface for communicating with external devices. Specific examples of the communication interface 14 include Ethernet (registered trademark), USB, and HDMI (registered trademark) ports. USB stands for Universal Serial Bus. HDMI stands for High-Definition Multimedia Interface.

[0016] The storage 13 stores programs that realize the functions of the functional components of the simulation device 10. These programs are read into the memory 12 by the processor 11 and executed by the processor 11. In this way, the functions of the functional components of the simulation device 10 are realized.

[0017] 1 shows only one processor 11. However, there may be a plurality of processors 11, and the plurality of processors 11 may cooperate to execute programs that realize the respective functions.

[0018] 3 to 7, the operation of the simulation device 10 according to the first embodiment will be described. The operation procedure of the simulation device 10 according to the first embodiment corresponds to the simulation method according to the first embodiment. Furthermore, the program that realizes the operation of the simulation device 10 according to the first embodiment corresponds to the simulation program according to the first embodiment.

[0019] 3, the flow of processing of the simulation apparatus 10 according to the first embodiment will be described. (Step S11: Input Processing) The acquisition unit 22 acquires a system model 31 and an input waveform 32 and inputs them to each of the multiple processing systems 21. Then, each processing system 21 uses the system model 31 and the input waveform 32 to simulate the operation of the object to be simulated and executes a collaborative simulation. Then, each processing system 21 outputs an output waveform 33 as a result of the collaborative simulation.

[0020] The system model 31 represents the specifications of the object to be simulated. The system model 31 is a specification description of the controller and controlled object that are the objects to be simulated. The controller is a device or equipment that applies control. The controller is also called a controller. The controlled object is a physical phenomenon or an artificial object that is controlled. The controlled object is also called a plant. The specification description is written using the Modelica language or the like.

[0021] The input waveform 32 is time-series data given as an input to the system model 31 from the outside. The output waveform 33 is time-series data resulting from analysis of the system model 31 by each processing system 21. As shown in Fig. 4, waveform data is obtained by plotting and interpolating the values ​​for each time indicated by the time-series data. Here, the input waveform 32 and output waveform 33 are waveform data obtained from the time-series data.

[0022] (Step S12: Acquisition process) The acquisition unit 22 acquires the output waveforms 33 output from each processing system 21. That is, the acquisition unit 22 acquires the same number of output waveforms 33 as the number of processing systems 21. The output waveforms 33 are associated with the processing systems 21 that are the output sources. Here, if the processing of a processing system 21 is successful, waveform data is obtained as the output waveform 33. However, if the calculation in a processing system 21 fails, time-series data is not obtained as the output waveform 33, and information indicating the calculation failure is obtained.

[0023] (Step S13: Exclusion Processing) The classification unit 23 excludes from the processing target the output waveforms 33 that indicate a calculation failure among the output waveforms 33 acquired in step S12. As a result, only the output waveforms 33 output from the processing system 21 that have been successfully processed remain among the output waveforms 33 acquired in step S12.

[0024] (Step S14: Similarity Calculation Process) The classification unit 23 calculates the waveform similarity between each output waveform 33 that was not excluded in step S13. The waveform similarity is the degree to which the output waveforms 33 are similar in features such as values ​​and shapes. The classification unit 23 calculates the waveform similarity using, for example, the Euclidean distance or Manhattan distance between each output waveform 33. As shown in FIG. 5 , the waveform similarity between each remaining output waveform 33 is calculated. Here, the closer the similarity value is to 1, the more similar the waveforms are.

[0025] (Step S15: Classification Process) The classification unit 23 classifies the multiple processing systems 21 into one or more groups based on the waveform similarity calculated in step S14. Here, the classification unit 23 considers that output waveforms 33 having a waveform similarity higher than a threshold value match as time-series data. The classification unit 23 then classifies the multiple processing systems 21 into one or more groups so that processing systems 21 that have output waveforms 33 having a waveform similarity higher than the threshold value are in the same group. If the waveform similarity is calculated as shown in FIG. 5 and the threshold value is 0.8, the multiple processing systems 21 are classified into one or more groups together with the output waveforms 33 output by the multiple processing systems 21, as shown in FIGS. 5 and 6.

[0026] (Step S16: Group Selection Process) The selection unit 24 selects one of the one or more groups as a specific group based on the processing systems 21 classified into each of the one or more groups. For example, the selection unit 24 selects one of the one or more groups as a specific group based on the number of processing systems 21 classified into each of the one or more groups. As a specific example, the selection unit 24 selects the group with the largest number of classified processing systems 21 as the specific group. When selecting the group with the largest number of classified processing systems 21 as the specific group, in the example of FIG. 6 , group α is selected as the specific group.

[0027] The selection unit 24 may select a plurality of groups as the specific group. For example, the selection unit 24 may select all groups in which the number of classified processing systems 21 is equal to or greater than a reference number as the specific group.

[0028] (Step S17: Processing System Selection Process) The selection unit 24 selects a representative processing system 34 from the processing systems 21 classified into the specific group selected in step S16. For example, the selection unit 24 selects, from among the processing systems 21 classified into a specific group, the processing system 21 having the highest average value of waveform similarity with the other processing systems 21 classified into that specific group as the representative processing system 34. In the example of FIG. 6 , if group α is selected as the specific group and the processing system 21 having the highest average value of waveform similarity is selected as the representative processing system 34, the processing system 21A is selected as the representative processing system 34. At this time, if multiple groups are selected as specific groups in step S16, the selection unit 24 selects a representative processing system 34 from each specific group. Note that the selection unit 24 may select multiple processing systems 21 from one specific group as the representative processing system 34.

[0029] The selection unit 24 extracts the output waveform 33 output from the representative processing system 34 as the representative waveform 35. In the example of FIG. 6, when the processing system 21A is selected as the representative processing system 34, the output waveform 33a is extracted as the representative waveform 35.

[0030] The functional configuration of the simulation device 10 shown in FIG. 1 uses specification descriptions rather than actual controllers and controlled objects, and is generally called MILS. MILS stands for Model In the Loop Simulation. Variations of MILS include SILS, SPILS, PILS, HILS, and RCP. SILS stands for Software In the Loop Simulation. SPILS stands for Simulation-based Processor In the Loop Simulation. PILS stands for Processor In the Loop Simulation. HILS is an abbreviation for Hardware In the Loop Simulation. RCP is an abbreviation for Rapid Control Prototyping. The simulation device 10 according to the first embodiment can be applied to various forms of simulation in the same manner, such as SILS, SPILS, PILS, HILS, and RCP, in addition to MILS.

[0031] For example, when the simulation device 10 according to the first embodiment is applied to a HILS, the configuration is as shown in Fig. 7. In other words, the controller is not modeled, and the system model 31 includes a specification description of only the controlled object. An input waveform 32 is input to the controller, and input data corresponding to the input waveform 32 is input to each processing system 21 by the controller.

[0032] ***Effects of First Embodiment*** As described above, the simulation device 10 according to the first embodiment calculates the waveform similarity of the output waveforms 33 and classifies the multiple processing systems 21 into one or more groups based on the waveform similarity. By using the classification results to select an appropriate processing system 21 and output waveform 33, it is possible to obtain valid calculation results. The processing systems 21 used in collaborative simulation are not technically completely mature, and the calculation results may differ or the calculation may fail. Even when collaborative simulation is performed using such an immature processing system 21, it is possible to obtain valid calculation results.

[0033] The simulation device 10 according to the first embodiment selects specific groups based on the number of processing systems 21 classified into each group, and selects a representative processing system 34 from the processing systems 21 classified into the specific group. The simulation device 10 then extracts the output waveform 33 output by the representative processing system 34 as a representative waveform 35. By using the representative processing system 34 and the representative waveform 35, it is possible to obtain appropriate calculation results.

[0034] ***Other Configurations*** <Modification 1> In the first embodiment, each functional component is realized by software. However, in Modification 1, each functional component may be realized by hardware. The differences between Modification 1 and the first embodiment will be described below.

[0035] When each functional component is realized by hardware, the simulation device 10 includes an electronic circuit instead of the processor 11, the memory 12, and the storage 13. The electronic circuit is a dedicated circuit that realizes the functions of each functional component, the memory 12, and the storage 13.

[0036] Possible electronic circuits include a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, a logic IC, a GA, an ASIC, and an FPGA. GA stands for Gate Array. ASIC stands for Application Specific Integrated Circuit. FPGA stands for Field-Programmable Gate Array. Each functional component may be realized by a single electronic circuit, or each functional component may be distributed across multiple electronic circuits.

[0037] <Modification 2> As a modification 2, some of the functional components may be realized by hardware, and other functional components may be realized by software.

[0038] The processor 11, the memory 12, the storage 13, and the electronic circuitry are collectively referred to as a processing circuit. In other words, the functions of the respective functional components are realized by the processing circuit.

[0039] Furthermore, the term "unit" in the above description may be read as a "circuit," "step," "procedure," "process," or "processing circuit."

[0040] Second Embodiment In the second embodiment, a method for calculating waveform similarity will be described. In the second embodiment, differences from the first embodiment will be described, and a description of the same points will be omitted.

[0041] In the second embodiment, the classification unit 23 uses DTW, which stands for Dynamic Time Warping, as a method for calculating waveform similarity.

[0042] ***Description of Operation*** The flow of processing by the simulation device 10 according to the second embodiment will be described with reference to Fig. 8. The processing in step S14 differs from that in Fig. 3 .

[0043] (Step S14: Similarity Calculation Process) The classification unit 23 calculates the waveform similarity between each output waveform 33 that was not excluded in step S13, using DTW.

[0044] ***Effects of Second Embodiment*** When calculating waveform similarity using DTW, compared to when calculating waveform similarity using Euclidean distance or Manhattan distance, waveform similarity can be determined appropriately even if the lengths or periods of the two output waveforms 33 as time-series data are different. By calculating waveform similarity appropriately, it becomes possible to appropriately select the representative processing system 34.

[0045] ***Other Configurations*** <Modification 3> Derivative DTW may be used as a method for calculating waveform similarity.

[0046] The flow of processing by the simulation device 10 according to the third modification will be described with reference to Fig. 9. The processing in step S14 differs from that in Fig. 8.

[0047] (Step S14: Similarity Calculation Process) The classification unit 23 calculates the waveform similarity between each output waveform 33 that was not excluded in step S13 using Derivative DTW. Derivative DTW is an improved version of DTW. Compared to DTW, Derivative DTW has the advantage that it can calculate the waveform similarity by taking into account upward and downward trends.

[0048] By calculating the waveform similarity using Derivative DTW, the waveform similarity can be obtained appropriately. By calculating the waveform similarity appropriately, it becomes possible to appropriately select the representative processing system 34.

[0049] <Modification 4> K-shape may be used as a method for calculating waveform similarity.

[0050] The flow of processing by the simulation device 10 according to the fourth modification will be described with reference to Fig. 10. The processing in step S14 differs from that in Fig. 8.

[0051] (Step S14: Similarity Calculation Process) The classification unit 23 uses k-shape to calculate the waveform similarity between each output waveform 33 that was not excluded in step S13. K-shape is characterized by determining similarity using a distance measure that focuses on the phase and amplitude, which are waveform characteristics.

[0052] By calculating the waveform similarity using k-shape, the waveform similarity can be obtained appropriately. By calculating the waveform similarity appropriately, it becomes possible to appropriately select the representative processing system 34.

[0053] <Modification 5> PCA, which stands for Principal Component Analysis, may be used as a method for calculating waveform similarity.

[0054] The flow of processing by the simulation device 10 according to the fifth modification will be described with reference to Fig. 11. The processing in step S14 differs from that in Fig. 8.

[0055] (Step S14: Similarity Calculation Process) The classification unit 23 uses PCA to calculate the waveform similarity between each output waveform 33 that was not excluded in step S13. PCA has the characteristic of compressing data by constructing a variable that can be most efficiently explained among a large number of variables.

[0056] By calculating the waveform similarity using PCA, it is possible to obtain an appropriate waveform similarity. By calculating the waveform similarity appropriately, it becomes possible to appropriately select the representative processing system 34.

[0057] <Modification 6> The waveform similarity may be calculated using t-SNE, which stands for t-distributed Stochastic Neighbor Embedding.

[0058] The flow of processing by the simulation device 10 according to the sixth modification will be described with reference to Fig. 12. The processing in step S14 differs from that in Fig. 8.

[0059] (Step S14: Similarity Calculation Process) The classification unit 23 uses t-SNE to calculate the waveform similarity between each output waveform 33 that was not excluded in step S13. t-SNE is characterized in that it compresses high-dimensional data into low-dimensional data so that the proximity of the original data is as equal as possible, and then determines the number of nearby points.

[0060] By calculating the waveform similarity using t-SNE, the waveform similarity can be obtained appropriately. By calculating the waveform similarity appropriately, it becomes possible to appropriately select the representative processing system 34.

[0061] Aspects of the present disclosure are summarized below as appendices. (Appendix 1) A simulation device comprising: an acquisition unit that inputs a system model representing specifications of a target to be simulated and input waveforms that are time-series data to be provided as input to the target to each of a plurality of processing systems that simulate the target to acquire output waveforms that are time-series data output from the plurality of processing systems; and a classification unit that calculates waveform similarities of the output waveforms acquired from each of the plurality of processing systems by the acquisition unit and classifies the plurality of processing systems into one or more groups based on the waveform similarities. (Appendix 2) The simulation device according to Appendices 1 further comprises: a selection unit that selects one of the one or more groups as a specific group based on the processing systems classified into each of the one or more groups by the classification unit, and extracts a representative waveform from the output waveforms acquired from the processing systems classified into the specific group. (Supplementary Note 3) The simulation device according to Supplementary Note 2, wherein the selection unit selects one of the processing systems classified into the specific group as a representative processing system, and extracts the output waveform obtained from the representative processing system as the representative waveform. (Supplementary Note 4) The simulation device according to Supplementary Note 2 or 3, wherein the selection unit selects one of the one or more groups as a specific group based on the number of processing systems classified into each of the one or more groups. (Supplementary Note 5) The simulation device according to any one of Supplementary Notes 1 to 4, wherein the classification unit calculates the waveform similarity by any one of Dynamic Time Warping (DTW), Derivative Dynamic Time Warping (DTW), k-shape, Principal Component Analysis (PCA), and t-distributed Stochastic Neighbor Embedding (t-SNE).(Supplementary Note 6) A simulation method in which a computer inputs a system model representing the specifications of a target to be simulated and an input waveform that is time-series data to be given as input to the target to each of a plurality of processing systems that simulate the target to acquire an output waveform that is time-series data output from each of the plurality of processing systems, the computer calculates a waveform similarity of the output waveforms acquired from each of the plurality of processing systems, and classifies the plurality of processing systems into one or more groups based on the waveform similarity. (Supplementary Note 7) A simulation program that causes a computer to function as a simulation device that performs an acquisition process in which a system model representing the specifications of a target to be simulated and an input waveform that is time-series data to be given as input to the target to each of a plurality of processing systems that simulate the target to acquire an output waveform that is time-series data output from each of the plurality of processing systems, and a classification process in which the waveform similarity is calculated for the output waveforms acquired from each of the plurality of processing systems by the acquisition process, and classifies the plurality of processing systems into one or more groups based on the waveform similarity.

[0062] The embodiments and modifications of the present disclosure have been described above. Some of these embodiments and modifications may be combined and implemented. Furthermore, one or more of them may be implemented partially. Note that the present disclosure is not limited to the above embodiments and modifications, and various modifications are possible as needed.

[0063] 10 Simulation device, 11 Processor, 12 Memory, 13 Storage, 14 Communication interface, 21 Processing system, 22 Acquisition unit, 23 Classification unit, 24 Selection unit, 31 System model, 32 Input waveform, 33 Output waveform, 34 Representative processing system, 35 Representative waveform.

Claims

1. A simulation device comprising: an acquisition unit that inputs a system model representing the specifications of a simulated object and input waveforms, which are time-series data to be given as input to the simulated object, into each of a plurality of processing systems that simulate the simulated object, and acquires output waveforms, which are time-series data output from each of the plurality of processing systems; and a classification unit that calculates the waveform similarity of the output waveforms acquired from each of the plurality of processing systems by the acquisition unit, and classifies the plurality of processing systems into one or more groups based on the waveform similarity.

2. The simulation device according to claim 1, further comprising a selection unit that selects one of the one or more groups as a specific group based on the processing systems classified into each of the one or more groups by the classification unit, and extracts a representative waveform from the output waveforms obtained from the processing systems classified into the specific group.

3. The simulation device according to claim 2, wherein the selection unit selects one of the processing systems classified into the specific group as a representative processing system, and extracts the output waveform obtained from the representative processing system as the representative waveform.

4. The simulation device according to claim 2 or 3, wherein the selection unit selects one of the one or more groups as a specific group based on the number of processing systems classified into each of the one or more groups.

5. The simulation device according to any one of claims 1 to 4, wherein the classification unit calculates the waveform similarity using any one of DTW (Dynamic Time Warping), Derivative DTW (Dynamic Time Warping), k-shape, PCA (Principal Component Analysis), and t-SNE (t-distributed Stochastic Neighbor Embedding).

6. A simulation method in which a computer inputs a system model representing the specifications of a simulated object and input waveforms, which are time-series data to be given as input to the simulated object, into each of multiple processing systems that simulate the simulated object, thereby obtaining output waveforms, which are time-series data output from each of the multiple processing systems, and the computer calculates the waveform similarity of the output waveforms obtained from each of the multiple processing systems, and classifies the multiple processing systems into one or more groups based on the waveform similarity.

7. A simulation program that causes a computer to function as a simulation device that performs an acquisition process in which a system model representing the specifications of a simulated object and an input waveform, which is time-series data to be given as input to the simulated object, are input to each of multiple processing systems that simulate the simulated object, thereby acquiring output waveforms, which are time-series data output from each of the multiple processing systems; and a classification process in which the waveform similarity of the output waveforms acquired from each of the multiple processing systems by the acquisition process is calculated, and the multiple processing systems are classified into one or more groups based on the waveform similarity.

Images