Design method for hose fittings

Abstract

By automatically designing hose fittings that meet target performance using AI technology, it eliminates the need to rely on the designer's knowledge, experience, or intuition, resulting in a hose fitting design method with less development effort, time, and cost than conventional methods. [Solution] The computer performs the following steps: step S502 inputs design parameters and corresponding performance values ​​for the design object into an untrained AI design model; step S503 generates a trained AI design model that has learned the design parameters and corresponding performance values ​​for the design parameters; step S504 inputs the required performance values ​​for the design object into the trained AI design model; and step S505 has the trained AI design model calculate the design parameters for the design object, thereby automatically designing the design object using AI technology.

Description

Technical Field

[0001] The present invention relates to a design method for a hose joint that connects a hose body fixed by clamping a clamping fitting that has a core fitting and a clamping fitting and covers the hose body into which the core fitting is inserted to another component.

Background Art

[0002] In the fields of fashion design and clothing design, in recent years, there has been a need for high-resolution design generation technology, and there is also a growing need for design generation technology that reflects rapidly changing fashion trends. For this reason, a design generation method based on learned conditions performed by a design generation device based on learned conditions has been proposed, which includes steps of acquiring an image from information including an image and text of clothing or the like, learning the features of the acquired image, extracting text on the information and matching the extracted text with the features of the learned image, a step in which a user inputs user text for generating a design image, a step of identifying conditions corresponding to the user text, and a step of generating a design image based on the identified conditions (see Patent Document 1).

[0003] Furthermore, in the field of technology to which tires on mobile vehicles belong, there is a need to provide a tire shape creation method and a tire shape creation apparatus that can stably create a tire shape determined by measurable dimensional parameter values ​​while achieving target values ​​for tire performance. For this purpose, a first tire shape creation system that creates a tire shape using multiple dimensional parameters related to the tire shape and a second tire shape creation system that creates a tire shape using multiple non-dimensional parameters are used. Next, a tire model having the original tire shape created by the second tire shape creation system is created and the tire performance is simulated. Then, the values ​​of the dimensional parameters related to the reproduced tire shape, which reproduces the original tire shape using the first tire shape creation system, are extracted. Multiple sets of values ​​of the simulation calculation results and dimensional parameter values ​​obtained in this way are used to train a prediction model. A tire shape determination method, tire shape determination apparatus, and program are known that determine the optimal tire shape that achieves the target value using a prediction model based on the target value of the tire performance (see Patent Document 2). [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Special Publication No. 2023-550534 [Patent Document 2] Japanese Patent Publication No. 2022-63581 [Overview of the project] [Problems that the invention aims to solve]

[0005] Hose fittings, also known as hose crimping fittings, consist of a core fitting and a clamping fitting. The hose can be secured by crimping the clamping fitting, which covers the hose body into which the core fitting is inserted, causing plastic deformation. Since high-pressure fluids such as water and oil flow through the hose, if the contact pressure between the fitting and the hose is low, it can lead to hose leakage. Furthermore, a difference in stress occurs in the hose when fluid is flowing and when it is not. If this stress difference is large, repeated loading will occur over time, leading to failure or deformation. Considering these points, it is necessary to design the hose to have sufficient contact pressure while keeping the stress difference below a certain value.

[0006] Traditionally, the design of hose fittings, which connect the hose body to other components, relied on the designer's knowledge, experience, or intuition. For example, in the design of such hose fittings, skilled designers would calculate and evaluate each parameter using numerical analysis with the finite element method (FEM), and then repeatedly modify the shape based on the results, relying on the designer's knowledge, experience, or intuition to find a shape that met the required performance values.

[0007] However, if the design method for hose fittings relies on the designer's knowledge, experience, or intuition, which is a conventional design method, the design quality will depend on the skill level of the designer, potentially leading to significant variations in design quality. For example, in typical design, the design is done manually, referencing existing shape databases and the designer's experience or intuition, which is time-consuming and results in numerous corrections and rework if performance values ​​do not match. Furthermore, the design skills for hose fittings become individualized to the designer, potentially hindering the transfer of design skills to other designers.

[0008] The present invention aims to realize a hose fitting design method that automatically designs hose fittings that meet target performance using AI technology, thereby eliminating the need to rely on the designer's knowledge, experience, or intuition, and resulting in less development effort, development time, and development cost than conventional methods. [Means for solving the problem]

[0009] This invention has been made in view of the above problems, and relates to a method for designing hose fittings, Computers The steps include inputting design parameters for the design target and performance values ​​corresponding to those design parameters into an untrained AI design model, The steps include generating a trained AI design model that learns the design parameters and performance values ​​corresponding to the design parameters, The steps include inputting the required performance values ​​for the design target into the aforementioned trained AI design model, The method for designing a hose fitting is characterized in that the trained AI design model performs the step of calculating the design parameters of the object to be designed.

[0010] According to the above configuration, the computer inputs the design parameters and corresponding performance values ​​of the design object into an untrained AI design model, generates a trained AI design model that has learned the design parameters and corresponding performance values, inputs the required performance values ​​for the design object into the trained AI design model, and the trained AI design model calculates the design parameters of the design object. As a result, hose fittings can be designed using the trained AI design model, eliminating the need to rely on the designer's knowledge, experience, or intuition, and enabling a hose fitting design method with less development effort, development time, and development cost than conventional methods. [Effects of the Invention]

[0011] According to the present invention, a trained AI design model is generated by learning design parameters and performance values ​​corresponding to those design parameters. By inputting the required performance values ​​needed for the design target into the trained AI design model, the generated trained AI design model calculates the design parameters for the design target. As a result, a hose fitting can be designed using the AI ​​design model, eliminating the need to rely on the designer's knowledge, experience, or intuition, and enabling a hose fitting design method with less development effort, development time, and development cost than conventional methods. [Brief explanation of the drawing]

[0012] [Figure 1]This is a cross-sectional view of a hose fitting, which is the subject of the design of the present invention, cut along the axis of the hose. [Figure 2] This is a cross-sectional view of a fastening fitting included in the hose fitting, which is the subject of the design of the present invention, cut along the axis of the hose. [Figure 3] This is a cross-sectional view of a core fitting included in the hose fitting, which is the subject of the design of the present invention, cut along the axis of the hose. [Figure 4] The names and descriptions of the performance indicators for the hose fitting in one embodiment of the present invention are as follows. [Figure 5] This is a flowchart showing how an AI design model calculates design parameters in one embodiment of the present invention. [Modes for carrying out the invention]

[0013] The hose fitting 101, which is the subject of the design in the present invention, will be described with reference to Figure 1. Figure 1 is a cross-sectional view of the hose fitting, which is the subject of the design in the present invention, cut along the axis of the hose. The hose fitting 101 designed in this invention is a fitting used to connect a hydraulic hose used in hydraulic equipment to other components. The hose body 102, through which a high-pressure fluid such as water or oil flows, has a structure in which multiple rubber layers and metal reinforcing layers are stacked in sequence. In contrast, the hose fitting 101 consists of two components: a core fitting 103 and a fastening fitting 104. As shown in Figure 1, for example, the hose fitting 101 is fixed by inserting the core fitting 103, which has multiple grooves 106, into the interior of the hose body 102 from the end of the hose body 102, and the fastening fitting 104, which has multiple claws 105, covers the hose body 102.

[0014] In a state where the clamping fitting 104 covers the hose body 102, the clamping fitting 104 is clamped in the axial direction of the hose body 102 and the core fitting 103 by a tool, a machine, etc. from, for example, 8 directions. As a result, the clamping fitting 104 is plastically deformed, and the hose body 102 is fixed to the hose joint 101 by being sandwiched between a plurality of claws 105 of the clamping fitting 104 and a plurality of grooves 106 of the core fitting 103. In addition, in order to fix or assemble the hose joint 101 to other components or the like, the hose joint 101 may further include a nut 107 and a bolt 108, or may include a member other than the nut 107 or the bolt 108.

[0015] FIG. 2 is a cross-sectional view obtained by cutting the upper half of the clamping fitting included in the hose joint which is the design object of the present invention along the hose axis 202, and shows in more detail an example of design parameters (design variables) which are learning data of the AI design model in the clamping fitting 201 as one embodiment of the present invention. These design parameters are selected from the main dimensions of the fitting and are arbitrarily determined by the designer.

[0016] FIG. 3 is a cross-sectional view of the upper half obtained by cutting the core fitting included in the hose joint which is the design object of the present invention along the hose axis 302, and shows in more detail an example of design parameters (design variables) which are learning data of the AI design model in the core fitting 301 as one embodiment of the present invention. These design parameters are selected from the main dimensions of the fitting and are arbitrarily determined by the designer. Note that the hose axis 302 corresponds to the hose axis 202 in FIG. 2.

[0017] Next, the names and explanations of each performance index will be described based on FIG. 4. FIG. 4 is an example of each performance index of the hose joint in one embodiment of the present invention. It is desirable to select those having a strong correlation with the target performance obtained by calculation for this performance index. As described above, the hose joint, also called a hydraulic hose clamp fitting, has a core fitting and a clamp fitting, and by clamping the clamp fitting that covers the hose body into which the core fitting is inserted and plastically deforming it, the hose body can be fixed.

[0018] For example, since high-pressure fluids such as water and oil flow through the hose, if the contact pressure between the fitting and the hose is small, it will lead to hose leakage. Therefore, it is desirable that the performance index sum_contact_pressure [N / m] be above a certain value.

[0019] Also, there is a difference in the stress applied to the hose when the fluid is flowing and when the fluid is not flowing. If this stress difference is large, repeated loads will occur during continuous use, leading to breakage or deformation. Considering these points, it is necessary to design such that there is sufficient contact pressure and the stress difference is below a certain value.

[0020] The performance index 1st_layer_ave_stress [Mpa] is the average of the stress differences in the first layer of the reinforcing layer (wire layer) of the hose body. As described above, it is desirable that it be below a certain value. The performance index 2nd_layer_ave_stress [Mpa] is the average of the stress differences in the second layer of the reinforcing layer (wire layer) of the hose body. As described above, it is desirable that it be below a certain value. The performance index 3rd_layer_ave_stress [Mpa] is the average of the stress differences in the third layer of the reinforcing layer (wire layer) of the hose body. As described above, it is desirable that it be below a certain value.

[0021] 4th_layer_ave_stress [Mpa] is the average of the stress differences in the fourth layer of the reinforcing layer (wire layer) of the hose body. As described above, it is desirable that it be below a certain value.

[0022] The performance index all_layer_ave_stress [MPa] is the average stress difference between the 1st to 4th reinforcing layers (wire layers) of the hose body, and as mentioned above, it is desirable that it be below a certain value.

[0023] Figure 5 shows an example flowchart of how an AI design model calculates design parameters in one embodiment of the present invention. In general, the design method for hose fittings is performed by an information processing device such as a computer, which has a central processing unit (CPU) and a main memory (memory) connected to the central processing unit.

[0024] The process of generating an AI design model for the design method of hose fittings starts from the beginning of S501 (S501). Specifically, in S501, depending on the selected AI design model, initial values ​​are set for each variable, as well as for hyperparameters such as the learning rate, which represents how much the weight parameters are changed at once; the batch size, which is the number of data points per group when the dataset to be trained is divided into multiple groups; and the number of epochs, which is the number of iterations that have been used to train the data (S501).

[0025] In the following step 502, the design parameters and corresponding performance values ​​are input into the untrained AI design model (S502). Specifically, the design parameters (design variables) that serve as training data for the AI ​​design model in the fastening fitting 201 exemplified in Figure 2, the design parameters (design variables) that serve as training data for the AI ​​design model in the core fitting 301 exemplified in Figure 3, and the corresponding performance values ​​for each performance index of the hose fitting exemplified in Figure 4 are input into the AI ​​model before training (S502).

[0026] Then, in step 503, an AI design model is generated that has been trained using the design parameters and the performance values ​​corresponding to those design parameters (S503). Specifically, based on the design parameters (design variables) that serve as training data for the AI ​​design model of the fastening fitting 201 exemplified in Figure 2 and the design parameters (design variables) that serve as training data for the AI ​​design model of the core fitting 301 exemplified in Figure 3, which are input in S502, and the corresponding performance values ​​for each performance index of the hose fitting exemplified in Figure 4, training is performed according to the learning rate, batch size, and number of epochs set in step S501, thereby generating a trained AI design model that has learned the correlation between the design parameters and the performance values ​​corresponding to those design parameters (S503).

[0027] Next, in step S504, the required performance values ​​for the hose fitting to be designed are input to the trained AI design model generated in step S503 (S504). Specifically, for each performance index of the hose fitting exemplified in Figure 4, input the required performance value for the hose fitting under design (S504).

[0028] Then, in step S505, the trained AI design model generated in step S503 calculates and outputs the design parameters of the hose fitting to be designed (S505). Specifically, the trained AI design model generated in step S503 calculates and outputs the design parameters of the hose fitting to be designed, which are inference results corresponding to the required performance values ​​for the hose fitting to be designed that were input in step S504 (S505).

[0029] In step S505 described above, the design parameters of the hose fitting to be designed, which are the inference results, are calculated and output. Finally, in step S506, the AI ​​design model in Figure 5 completes the operation of calculating the design parameters (S506).

[0030] Although embodiments of the present invention have been described, the present invention is not limited to the embodiments described above, and various design modifications are possible without departing from the spirit of the invention. It goes without saying that the present invention can be implemented in a variety of forms within the scope of the spirit of the invention. Contribution to the United Nations-led Sustainable Development Goals (SDGs)

[0031] With the SDGs being advocated for the realization of a sustainable society, this invention is considered to have the potential to become a technology that contributes to creating livable cities and other such initiatives. [Explanation of Symbols]

[0032] 101...Hose fitting, 102...Hose body, 103...Core fitting, 104...Clamping fitting, 105...Claw, 106...Groove, 107...Nut, 108...Bolt

Claims

[Claim 1] Computers The steps include inputting design parameters for the design target and performance values ​​corresponding to those design parameters into the AI ​​design model before training, The steps include generating a trained AI design model that learns the design parameters and performance values ​​corresponding to the design parameters, The steps include inputting the required performance values ​​for the design target into the trained AI design model, A method for designing a hose fitting, comprising the steps of: the trained AI design model calculating the design parameters of the object to be designed; and the other step of the trained AI design model.

Images