Model reconstruction method, electronic device and medium for large equipment

Abstract

The present invention proposes a model reconstruction method, electronic device and medium for large equipment. The model reconstruction method selects a limited number of sample working conditions based on the working condition stress range, simulates the sample working conditions and trains the simulation results based on a machine learning algorithm to obtain predicted stress results of all working condition stress ranges, achieving a second-level solution and greatly reducing the simulation calculation time. The model reconstruction method compares the predicted stress results with the actual stress results to obtain the stress error. If the stress error exceeds the error critical value, the machine learning algorithm is replaced, and the training samples are trained again until the stress error does not exceed the error critical value. The predicted stress results are output as predicted simulation results, and fast solution improves the efficiency of result verification. The model reconstruction method obtains node coordinates and node stress values ​​based on the predicted simulation results, constructs a scatter plot based on the node coordinates and assigns colors to obtain a three-dimensional cloud map model, and the three-dimensional cloud map model has a fast visualization speed.

Description

Technical Field

[0001] The present invention relates to the field of rapid simulation technology, and in particular to a model reconstruction method, electronic equipment and medium for large equipment. Background Art

[0002] With the development of modern technology, industrial production equipment has gradually become intelligent, and emerging concepts such as digital twin technology have gradually been applied to various industries. The "virtual-real synchronization" technology consisting of real-time simulation of physical equipment operation and virtual equipment has been widely mentioned. At present, the simulation of underground engineering equipment is mostly carried out through finite element simulation software to check the strength, obtain the stress condition of the equipment, and then display it through post-processing software. If this method is used for virtual-real synchronization display, there are the following problems:

[0003] First, the working conditions of underground engineering equipment are constantly changing during the actual working process. For example, when the drilling rig is drilling, the impact force on the boom is different, and the force condition of the boom changes in real time. The finite element simulation software sets the boundary conditions of the model and then simulates it. For different working conditions (for example, the boom of the drilling rig is subjected to different impact forces), different boundary conditions need to be set, and then the finite element differential equations are iteratively solved again. The model of underground engineering equipment is complex, the number of grids is in the millions, and the model solution time is long. When the working conditions such as the impact force are constantly changing, the parameters need to be readjusted and then the solution is re-performed, which leads to low efficiency and time-consuming of the entire process.

[0004] Secondly, for the parameter range of different boundary conditions (working conditions), the results of the arm force and impact force under different working conditions within the range are calculated by existing software. However, due to the need to geometrically display the surface and volume information, it is impossible to meet the requirements of real-time display of the force cloud map, deformation cloud map and other results of the three-dimensional model.

[0005] In addition, the model reduction technology proposed to solve the real-time problem is to process a large number of data samples through an algorithm and then solve them. For the finite element model, its format is mostly based on the finite element model data format (such as odb, etc.), which is difficult to meet the basic sample format of algorithm training.

[0006] Finally, even if the required data (such as stress, deformation and other result data) is extracted and the model reduction algorithm is trained, most of them can only be displayed again in the form of a one-dimensional line graph. It is impossible to intuitively obtain a visualized three-dimensional model and see its posture under different working conditions.

[0007] The above factors all affect the synchronous mapping of the actual physical prototype status and the virtual model prototype status in the digital twin technology, which makes it difficult to meet the needs of the effective application of existing digital twin technology in underground engineering equipment.

[0008] In summary, there is an urgent need for a model reconstruction method for large equipment to solve the problems existing in the prior art. Summary of the invention

[0009] The present invention aims to provide a model reconstruction method, electronic device and medium for large equipment. The specific technical scheme is as follows:

[0010] A model reconstruction method for large equipment, the steps are as follows:

[0011] Step S100: obtaining sample simulation results, specifically, establishing a model of large equipment and determining the stress range of working conditions, selecting typical sample working conditions based on the stress range of working conditions, simulating the large equipment based on the typical sample working conditions, and obtaining a sample simulation result set;

[0012] Step S200: Obtaining the prediction simulation results, specifically, using the sample simulation result set in step S100 as a training sample, training the training sample based on the machine learning algorithm, and obtaining the predicted force results of the nodes within the force range of all working conditions; comparing the predicted force results with the actual force results to obtain the force error, and if the force error exceeds the error critical value, replacing the machine learning algorithm, training the training samples again, until the force error does not exceed the error critical value, and outputting the predicted force results and the node coordinates as the prediction simulation results;

[0013] Step S300: constructing a three-dimensional cloud map model, specifically, reading the predicted simulation results in step S200, obtaining node coordinates and node stress deformation values, the node stress deformation values ​​including node stress values ​​and node displacement data, and constructing a scatter plot based on the node coordinates; the node stress deformation values ​​and colors are matched and the nodes are colored to obtain a three-dimensional cloud map model.

[0014] Specifically, in step S100, the typical sample working condition includes the maximum value and the minimum value within the working condition stress range.

[0015] Specifically, in step S100, the typical sample working conditions also include working conditions that are equidistantly selected within the working condition stress range and account for at least 30% of the total working conditions.

[0016] Specifically, in step S200, the error critical value is 1%.

[0017] Specifically, in step S300, the specific steps of constructing the three-dimensional cloud model are as follows:

[0018] Step S301: constructing a scatter plot and adjusting the node size, specifically, reading the prediction simulation results in step S2, obtaining node coordinates and node stress and deformation values, constructing a scatter plot based on the node coordinates, and adjusting the size attributes of the nodes in the scatter plot so that adjacent nodes are in contact;

[0019] Step S302: construct a mapping between nodes and colors. Specifically, the node stress values ​​and color levels are matched and the colors corresponding to the nodes are assigned to the scatter plot in step S301 to obtain a three-dimensional cloud map model.

[0020] Specifically, in step S302, the node stress value and the color scale correspond to each other based on the gradient color, red corresponds to the maximum node stress value, and blue corresponds to the minimum node stress value;

[0021] Alternatively, the node stress value and the color level correspond based on the grayscale, the darker the color, the larger the node stress value, and the lighter the color, the smaller the node stress value.

[0022] Specifically, the large equipment is a rock drilling rig boom.

[0023] In addition, the present invention also provides an electronic device, comprising:

[0024] Memory for storing computer programs;

[0025] A processor is used to implement the above-mentioned model reconstruction method when executing the computer program.

[0026] In addition, the present invention also proposes a computer medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned model reconstruction method is implemented.

[0027] The application of the technical solution of the present invention has the following beneficial effects:

[0028] The model reconstruction method of the present invention simplifies the input conditions and only selects a limited number of operating conditions (at least including the maximum operating condition and the minimum operating condition) from the operating condition range as typical sample operating conditions. The model of the limited number of typical sample operating conditions is then trained through a machine learning algorithm to obtain simulation results of all operating condition ranges. The existing finite element software solver (differential equation iteration) requires at least tens of minutes to solve each operating condition, while the model reconstruction method of the present invention can achieve a solution in seconds, which greatly reduces the simulation calculation time for large equipment with complex operating conditions.

[0029] The model reconstruction method of the present invention performs simulation based on a machine learning algorithm, and only needs to set the algorithm coefficients of the machine learning algorithm once, which is simple and convenient and is conducive to verifying the correctness of the simulation results.

[0030] The model reconstruction method of the present invention reconstructs the node coordinates to obtain a scattered point model, and then achieves visual visualization by adjusting the size of the scattered points. Since it does not contain information such as lines and surfaces, the model reconstruction speed is faster, which provides favorable conditions for the real-time visualization display in the virtual-real synchronization of digital twins.

[0031] The model reconstruction method of the present invention can quickly obtain the stress conditions of equipment under different working conditions for different large-scale equipment, which is beneficial to the rapid design of designers and the rapid response of construction personnel, solves the problems of long solution time of finite element software and the need to repeatedly set boundary conditions for solution, and greatly improves work efficiency.

[0032] In addition to the above-described purposes, features and advantages, the present invention has other purposes, features and advantages. The present invention will be further described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0034] Figure 1 It is a flow chart of the steps of the model reconstruction method;

[0035] Figure 2 yes Figure 1 Flow chart of step S300;

[0036] Figure 3 It is a schematic diagram of (part of) a scatter plot;

[0037] Figure 4 It is a schematic diagram of the three-dimensional cloud model (grayscale contrast). DETAILED DESCRIPTION

[0038] The embodiments of the present invention are described in detail below with reference to the accompanying drawings, but the present invention can be implemented in many different ways as defined and covered by the claims.

[0039] Embodiment 1:

[0040] This embodiment discloses a model reconstruction method for large equipment. The preferred large equipment in this embodiment is a rock drilling trolley boom. The steps are as follows:

[0041] Step S100: obtaining sample simulation results, specifically, establishing a model of large equipment and determining the stress range of working conditions, selecting typical sample working conditions based on the stress range of working conditions, simulating the large equipment based on the typical sample working conditions, and obtaining a sample simulation result set;

[0042] It should be noted that for the geological environment in which large equipment in underground engineering usually serves, the extreme working conditions have been designed in the structural design process, that is, the extreme working condition input values ​​that the equipment components are subjected to, and then the working condition stress range can be obtained.

[0043] Furthermore, the typical sample working conditions include the maximum and minimum values ​​within the working condition stress range. In order to improve the accuracy of subsequent calculation results, this embodiment also selects working conditions that are equidistant within the working condition stress range and account for at least 30% of the total working conditions as typical sample working conditions.

[0044] It should be noted that the sample simulation results include node coordinates and node stress-deformation values. The node stress-deformation values ​​include node stress values ​​and node displacement data. If you only care about the stress value of the drilling rig boom, you only need the node coordinates and node stress values. If you only care about the deformation value of the drilling rig boom, you only need the node coordinates and node displacement data.

[0045] Step S200: Obtaining the prediction simulation results, specifically, using the sample simulation result set in step S100 as a training sample, training the training sample based on the machine learning algorithm, and obtaining the predicted force results of the nodes within the force range of all working conditions; comparing the predicted force results with the actual force results to obtain the force error, and if the force error exceeds the error critical value, replacing the machine learning algorithm, training the training samples again, until the force error does not exceed the error critical value, and outputting the predicted force results and the node coordinates as the prediction simulation results;

[0046] Furthermore, the machine learning algorithm includes but is not limited to a neural network algorithm, a deep learning algorithm, a supervised learning algorithm, and an unsupervised learning algorithm;

[0047] This embodiment adopts a supervised learning algorithm, and the specific training process is as follows:

[0048] The working conditions corresponding to the above training samples, i.e., sample working conditions, are used as input. For example, a set of sample working conditions is as follows: the force magnitude of the stress node is 2000N, and the corresponding stress values ​​of each node (node ​​number i, 1, 2, 3…127…275) are F i (F1=301MPa, F2=203.75MPa…F 127 =501MPa…F 275=63MPa). The samples in other groups are similar, so the node stress values ​​of these sample working conditions are normalized first to facilitate the subsequent machine learning algorithm training. Then, the processed samples are machine learned through the algorithm in the python machine learning tool Scikit-learn. After the algorithm training, the predicted stress results of each node under different stress conditions are obtained, that is, the predicted stress results of each node. The stress error is calculated according to the predicted stress results. When the stress error does not exceed 1%, the algorithm model can be considered valid, otherwise the algorithm needs to be modified. When the supervised learning Delta learning rule is adopted in this embodiment, when the final error exceeds 1%, the connection weights between neurons and the learning speed constant are adjusted to calculate again until the predicted stress value calculated by the algorithm model reaches the error range; if the actual output of the neuron is larger than the expected output, the weights of all inputs as positive connections are reduced; if the actual output of the neuron is smaller than the expected output, the weights of all inputs as positive connections are increased.

[0049] It should be noted that, for large equipment other than the drilling rig, if the algorithm cannot be achieved, it means that other algorithms need to be replaced for the large equipment.

[0050] Furthermore, the algorithms in this embodiment can be many, such as regression algorithm, or a combination of integrated algorithms, clustering algorithms, etc. can be used according to different samples of different research objects and the number of different samples; in addition, the node stress values ​​can be normalized and standardized, which is conducive to machine learning algorithm training.

[0051] Step S300: constructing a three-dimensional cloud map model, specifically, reading the prediction simulation results in step S200, obtaining node coordinates and node stress and deformation values, and constructing a scatter plot based on the node coordinates; the node stress and deformation values ​​are matched with colors and the nodes are colored to obtain a three-dimensional cloud map model;

[0052] like Figure 2 As shown in the figure, the specific steps of constructing a 3D cloud model are as follows:

[0053] Step S301: construct a scatter plot and adjust the node size. Specifically, use the Occ module or vtk module of Python to read the prediction simulation results in step S200, obtain the node coordinates and node stress and deformation values, and construct a scatter plot based on the node coordinates. Figure 3 The scatter plot shown in the figure; because there are gaps between the nodes on the scatter plot, it cannot visually present the overall appearance of the equipment. Therefore, by adjusting the size of the node in the vtk module, the node can be enlarged until the size of the adjacent node touches each other to achieve visual coherence;

[0054] Step S302: Construct a mapping between nodes and colors. Specifically, map the node stress values ​​to colors one by one and assign them to each node to obtain a three-dimensional cloud map model. Since the model is composed of scattered points in the memory, there are no lines, surfaces, or bodies during reconstruction, but only scattered points, so the processing speed and efficiency are higher.

[0055] like Figure 4 As shown, the preferred node stress values ​​and color levels in this embodiment correspond based on grayscale, dark colors correspond to the maximum node stress value, and light colors correspond to the minimum node stress value.

[0056] It should be noted that the node stress value and the color scale can also correspond based on a gradient color, red corresponds to the maximum node stress value, and blue corresponds to the minimum node stress value. As the node stress value changes from large to small, you can choose from light to dark, from cold to warm tones in the color scale, so as to achieve a cloud map effect.

[0057] In addition, this embodiment also discloses an electronic device, including:

[0058] Memory for storing computer programs;

[0059] A processor is used to implement the above-mentioned model reconstruction method when executing the computer program.

[0060] In addition, this embodiment further discloses a computer medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned model reconstruction method is implemented.

[0061] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0062] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0063] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A model reconstruction method for large equipment, characterized in that: Here are the steps: Step S100: obtaining sample simulation results, specifically, establishing a model of large equipment and determining the stress range of working conditions, selecting typical sample working conditions based on the stress range of working conditions, simulating the large equipment based on the typical sample working conditions, and obtaining a sample simulation result set; Step S200: Obtaining the prediction simulation results, specifically, using the sample simulation result set in step S100 as a training sample, training the training sample based on the machine learning algorithm, and obtaining the predicted force results of the nodes within the force range of all working conditions; comparing the predicted force results with the actual force results to obtain the force error, and if the force error exceeds the error critical value, replacing the machine learning algorithm, training the training samples again, until the force error does not exceed the error critical value, and outputting the predicted force results and the node coordinates as the prediction simulation results; Step S300: constructing a three-dimensional cloud map model, specifically, reading the predicted simulation results in step S200, obtaining node coordinates and node stress deformation values, the node stress deformation values ​​including node stress values ​​and node displacement data, and constructing a scatter plot based on the node coordinates; the node stress values ​​and colors are matched and the nodes are colored to obtain a three-dimensional cloud map model.

2. The model reconstruction method according to claim 1, characterized in that: In step S100 , the typical sample working condition includes the maximum value and the minimum value within the working condition stress range.

3. The model reconstruction method according to claim 2, characterized in that: In step S100, the typical sample working conditions also include working conditions that are equidistantly selected within the working condition stress range and account for at least 30% of the total working conditions.

4. The model reconstruction method according to claim 1, characterized in that: In step S200, the error threshold is 1%.

5. The model reconstruction method according to claim 1, characterized in that: In step S300, the specific steps of constructing a three-dimensional cloud model are as follows: Step S301: constructing a scatter plot and adjusting the node size, specifically, reading the prediction simulation results in step S2, obtaining node coordinates and node stress and deformation values, constructing a scatter plot based on the node coordinates, and adjusting the size attributes of the nodes in the scatter plot so that adjacent nodes are in contact; Step S302: construct a mapping between nodes and colors. Specifically, the node stress values ​​and color levels are matched and the colors corresponding to the nodes are assigned to the scatter plot in step S301 to obtain a three-dimensional cloud map model.

6. The model reconstruction method according to claim 5, characterized in that: In step S302, the node stress values ​​and color levels correspond to each other based on gradient colors, with red corresponding to the maximum node stress value and blue corresponding to the minimum node stress value; Alternatively, the node stress value and the color level correspond based on the grayscale, the darker the color, the larger the node stress value, and the lighter the color, the smaller the node stress value.

7. The model reconstruction method according to any one of claims 1 to 6, characterized in that: The large equipment is a drilling rig boom.

8. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the model reconstruction method according to any one of claims 1 to 7 when executing the computer program.

9. A computer medium, characterized in that The computer medium stores a computer program, and when the computer program is executed by a processor, the model reconstruction method according to any one of claims 1 to 7 is implemented.

Images