Automatic optimization method and system for cross section of rod system component
By constructing a loss function and using gradient descent to optimize the cross-sectional parameters of the truss components, the problems of inefficiency and material waste in traditional design are solved, achieving efficient cross-sectional optimization and material saving.
Patent Information
- Application Number
- CN202410951545.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-16
- Publication Date
- 2024-11-08
AI Technical Summary
Traditional bar structure cross-section design relies on experience-based adjustments, leading to inefficiency, difficulty in finding optimal parameters, and easy material waste.
Construct a loss function and perform backpropagation to obtain the gradient, initialize the cross-sectional parameters, calculate the internal forces based on the design load, and optimize the cross-sectional parameters through gradient descent until the iteration termination condition is met.
It achieves automatic optimization of the cross-sectional parameters of the rod system components, meeting strength and stability requirements while reducing material usage and improving the economy and environmental friendliness of the structure.
Smart Images

Figure CN118917184A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of intelligent structural design, and in particular to a method and system for automatically optimizing the cross section of a bar system component. Background Art
[0002] Bar components are one of the most commonly used structural systems in the field of engineering construction, such as steel truss main beams, steel frame structures, Bailey beams, etc. The traditional way to design the cross-section of bar components is generally determined based on experience. When the cross-section parameters of bar components do not meet the requirements of the specifications or regulations, designers need to manually adjust and modify them. This method is not only inefficient, but also difficult to find the optimal cross-section parameters, which can easily lead to unreasonable cross-section design and waste of materials. Summary of the invention
[0003] The embodiments of the present application provide a method and system for automatically optimizing the cross-section of a bar system component to solve the problem that the manual adjustment and modification method in the related art is not only inefficient, but also difficult to find the optimal cross-section parameters, which easily leads to unreasonable cross-section design and material waste.
[0004] In a first aspect, a method for automatically optimizing a cross section of a bar system member is provided, comprising:
[0005] Set the design goal of the verification result and use it as a construction condition to construct a loss function for the verification result;
[0006] Based on the loss function, obtaining the gradient of the loss function;
[0007] Initializing the cross-sectional parameters of the bar system component, and obtaining the internal force exerted on the bar system component based on the design load, and obtaining the calculation result based on the cross-sectional parameters and the internal force exerted;
[0008] Based on the gradient of the loss function, the learning rate, the check result and the section parameters, the updated section parameters are obtained, and the check result is obtained again until the iteration end condition is met, and the final section parameters are used as the optimization results of the bar system components.
[0009] In some embodiments, the calculation results include the cross-sectional area of the bar system component, the maximum cross-sectional stress, and the maximum width-to-thickness ratio of the plate section included in the bar system component;
[0010] The design objectives of the verification result include: a first design objective, a second design objective and a third design objective;
[0011] Among them, the first design goal is that the maximum stress of the section does not exceed the first allowable value, the second design goal is that the maximum width-to-thickness ratio of the section does not exceed the second allowable value, and the third design goal is that the cross-sectional area of the bar system component reaches a minimum value under the premise of meeting the first design goal and the second design goal.
[0012] In some embodiments, based on the gradient of the loss function, the learning rate, the check result and the cross-sectional parameters, obtaining updated cross-sectional parameters specifically includes:
[0013] Substitute the check result into the gradient of the loss function and find the product of the learning rate and the gradient of the loss function;
[0014] The product is subtracted from the cross-sectional parameter to obtain an updated cross-sectional parameter.
[0015] In some embodiments, the learning rate ranges from 0.0001 to 0.001.
[0016] In some embodiments, the learning rate is 0.0001.
[0017] In some embodiments, the iteration termination condition includes that the loss function change amplitude is less than an amplitude threshold.
[0018] In some embodiments, the amplitude threshold is 1×10 6 .
[0019] In some embodiments, the iteration end condition includes the number of iterations reaching a threshold number.
[0020] In some embodiments, the number threshold is 15 times.
[0021] In a second aspect, a bar member cross-section automatic optimization system is provided, comprising:
[0022] The first module is used to: set the design target of the inspection result and use it as a construction condition to construct a loss function about the inspection result;
[0023] The second module is used to: obtain the gradient of the loss function based on the loss function;
[0024] The third module is used to: initialize the cross-sectional parameters of the bar system component, obtain the internal force exerted on the bar system component based on the design load, and obtain the calculation result based on the cross-sectional parameters and the internal force exerted;
[0025] The fourth module is used to obtain updated cross-sectional parameters based on the gradient of the loss function, the learning rate, the check result and the cross-sectional parameters, and return to obtain the check result until the iteration end condition is met, and use the final cross-sectional parameters as the optimization result of the bar system component.
[0026] The beneficial effects of the technical solution provided by this application include:
[0027] This application constructs a loss function in advance and performs back propagation to obtain the gradient of the loss function, then initializes the cross-sectional parameters of the bar, calculates the internal forces of the bar system components based on the design load; calculates based on the cross-sectional parameters and the internal forces of the bar system components to obtain the verification results; performs gradient descent based on the gradient of the loss function, the learning rate, the verification results and the cross-sectional parameters to obtain new cross-sectional parameters; repeats the adjustment until the cross-sectional optimization reaches the iteration end condition, and finally uses the final cross-sectional parameters as the optimization results of the bar system components. This application can realize the automatic optimization of the cross-sectional parameters of the bar system components, so that the material usage of the bar system components is as small as possible while meeting the requirements of strength, stability, structure, etc., to achieve the purpose of green and low-carbon, while improving the economy of the structure. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0029] Figure 1 A flow chart of a method for automatically optimizing the cross section of a bar system member provided in an embodiment of the present application;
[0030] Figure 2 A block diagram of a bar member cross-section automatic optimization system provided in an embodiment of the present application;
[0031] Figure 3 A schematic diagram of a bidirectionally stressed I-section simply supported beam provided in an embodiment of the present application;
[0032] Figure 4 A schematic diagram of a cross-section optimization process provided in an embodiment of the present application;
[0033] Figure 5 These are the various indicators in the cross-section optimization process provided in the embodiments of this application. DETAILED DESCRIPTION
[0034] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0035] See also Figure 1 As shown, the embodiment of the present application provides a method for automatically optimizing the cross section of a bar system component, which includes the following steps:
[0036] 101: Set the design goal of the verification result and use it as a construction condition to construct a loss function about the verification result.
[0037] Among them, the loss function can be recorded as loss(R), and the vector R is the verification result.
[0038] 102: Based on the loss function, perform back propagation to obtain the gradient of the loss function.
[0039] Among them, the gradient of the loss function is recorded as vector G.
[0040] 103: Initializing the cross-sectional parameters of the bar system component, and obtaining the internal force exerted on the bar system component based on the design load, and obtaining the calculation result based on the cross-sectional parameters and the internal force exerted on the bar system component.
[0041] Here, all cross-sectional parameters of the bar system components are represented as vector θ.
[0042] θ=[θ1,θ2,...,θ n ], where θ n is the nth section parameter.
[0043] It can be understood that the initial cross-sectional parameters of the bar system components are usually known quantities.
[0044] The internal force of the bar system components is obtained based on the design load. The calculation method can directly adopt the existing scheme, which will not be repeated here.
[0045] The check result can be recorded as vector R.
[0046] R=[R1(θ), R2(θ),...,R m (θ)], where R m (θ) is the mth check result of a single bar system component.
[0047] It can be seen that the essence of the loss function loss(R) is actually a function of the cross-sectional parameter vector θ.
[0048] It can be written as follows:
[0049] loss(R)=loss(R1(θ), R2(θ),..., R m (θ))=L(θ).
[0050] Accordingly, the gradient G of the loss function can be written as follows:
[0051]
[0052] The back propagation process can be expressed as follows:
[0053]
[0054] 104: Based on the gradient of the loss function, the learning rate, the check result and the cross-sectional parameters, the updated cross-sectional parameters are obtained, and the process returns to step 103 to perform iterative optimization until the iteration end condition is met, and the final cross-sectional parameters are used as the optimization results of the bar system component.
[0055] Set the learning rate and perform gradient descent on the gradient of the loss function according to the current section parameters to obtain the new section parameters.
[0056] The learning rate is denoted as α, which may be a constant value. For example, the learning rate ranges from 0.0001 to 0.001. Preferably, the learning rate is 0.0001. The learning rate may also be a function.
[0057] The gradient descent process is shown as follows: θ = θ - α × G.
[0058] The current section parameter moves a learning rate in the opposite direction of the current gradient to obtain a new section parameter.
[0059] This application constructs a loss function in advance and performs back propagation to obtain the gradient of the loss function, then initializes the cross-sectional parameters of the bar, calculates the internal forces of the bar system components based on the design load; calculates based on the cross-sectional parameters and the internal forces of the bar system components to obtain the verification results; performs gradient descent based on the gradient of the loss function, the learning rate, the verification results and the cross-sectional parameters to obtain new cross-sectional parameters; repeats the adjustment until the cross-sectional optimization reaches the iteration end condition, and finally uses the final cross-sectional parameters as the optimization results of the bar system components. This application can realize the automatic optimization of the cross-sectional parameters of the bar system components, so that the material usage of the bar system components is as small as possible while meeting the requirements of strength, stability, structure, etc., to achieve the purpose of green and low-carbon, while improving the economy of the structure.
[0060] It is understandable that the above calculation results are calculated according to the corresponding specifications or certain specific requirements. For example, as an example, the calculation results include the cross-sectional area of the bar system component, the maximum cross-sectional stress and the maximum width-to-thickness ratio of the plate section included in the bar system component.
[0061] The design objectives of the verification result include: a first design objective, a second design objective and a third design objective;
[0062] Among them, the first design goal is that the maximum stress of the section does not exceed the first allowable value, the second design goal is that the maximum width-to-thickness ratio of the section does not exceed the second allowable value, and the third design goal is that the cross-sectional area of the bar system component reaches a minimum value under the premise of meeting the first design goal and the second design goal.
[0063] The first allowable value and the second allowable value are both manually set according to actual needs.
[0064] In the above step 104, based on the gradient of the loss function, the learning rate, the check result and the cross-sectional parameters, the updated cross-sectional parameters are obtained, which specifically includes:
[0065] 201: Substitute the check result into the gradient of the loss function and calculate the product of the learning rate and the gradient of the loss function.
[0066] 202: Subtract the product from the cross-sectional parameter to obtain an updated cross-sectional parameter, that is, θ=θ-α×G.
[0067] The iteration termination condition includes that the change amplitude of the loss function is less than an amplitude threshold.
[0068] For example, as an example, the amplitude threshold is 1×10 6 .
[0069] The iteration end condition includes that the number of iterations reaches a threshold number.
[0070] For example, as an example, the number threshold is 15 times.
[0071] See also Figure 2 As shown, based on the above method, the embodiment of the present application also provides a bar member cross-section automatic optimization system, which includes:
[0072] The first module is used to: set the design target of the inspection result and use it as a construction condition to construct a loss function about the inspection result;
[0073] The second module is used to: obtain the gradient of the loss function based on the loss function;
[0074] The third module is used to: initialize the cross-sectional parameters of the bar system component, obtain the internal force exerted on the bar system component based on the design load, and obtain the calculation result based on the cross-sectional parameters and the internal force exerted;
[0075] The fourth module is used to obtain updated cross-sectional parameters based on the gradient of the loss function, the learning rate, the check result and the cross-sectional parameters, and return to obtain the check result until the iteration end condition is met, and use the final cross-sectional parameters as the optimization result of the bar system component.
[0076] The present application is further described in detail below with a specific example.
[0077] like Figure 3As shown in the figure, a simply supported beam structure with a span of 10m is subjected to a uniform vertical load of 500kN / m (z direction in the figure) and a uniform horizontal load of 200kN / m (y direction in the figure), both of which are design loads. The beam section is set to an I-shaped section that is symmetrical in both the upper and lower and left and right directions, and the specified value of its section maximum stress is determined to be no more than 300MPa, and the specified value of the section maximum width-to-thickness ratio of the upper and lower plates and the middle vertical plate of the I-shaped section is determined to be no more than 40.
[0078] Figure 4 It is a schematic diagram of the cross-section optimization process of the bar system component. The optimization process is as follows:
[0079] (1) Set the design goal of the verification result and use it as the construction condition to construct the loss function about the verification result.
[0080] Specifically, the first design goal is to ensure that the maximum stress of the cross section does not exceed the allowable value of 300 MPa.
[0081] The second design goal is to set the maximum width-to-thickness ratio of the upper and lower flat plates and the middle vertical plate to not exceed the allowable value of 40.
[0082] On the premise of meeting the first and second design goals, the third design goal is to minimize the cross-sectional area of the I-shape.
[0083] Combining the above three design objectives, the maximum cross-sectional stress, the maximum width-to-thickness ratio of the plate and the cross-sectional area are recorded as R1, R2 and R3 respectively, and the loss function can be constructed:
[0084] loss(R)=loss(R1, R2, R3).
[0085] (2) Initializing the section parameters of the I-section, and obtaining the internal force applied to the I-section based on the design load, and obtaining the calculation results based on the section parameters and the internal force applied.
[0086] The width and thickness of the upper and lower plates of the I-shaped bar system component are set to be θ1 and θ2 respectively; the width and thickness of the middle vertical plate are set to be θ3 and θ4 respectively.
[0087] Therefore, the cross-sectional parameters of the I-section can be expressed as: θ = [θ1, θ2, θ3, θ4].
[0088] The internal force is calculated based on the vertical uniformly distributed load of 500 kN / m and the transverse uniformly distributed load of 200 kN / m. Combined with θ=[θ1, θ2, θ3, θ4], the maximum section stress R1(θ), the maximum width-to-thickness ratio R2(θ) of the upper and lower flat plates and the middle vertical plate, and the cross-sectional area R3(θ) under the current section parameters can be calculated.
[0089] Thus we get R = [R1(θ), R2(θ), R3(θ)].
[0090] Furthermore, the following results can be obtained:
[0091] loss(R)=loss(R1(θ), R2(θ), R3(θ))=L(θ1, θ2, θ3,
[0092] θ4).
[0093] Based on the loss function, back propagation is performed to obtain the gradient of the loss function.
[0094]
[0095] (3) Based on the gradient of the loss function, the learning rate, the check result and the cross-sectional parameters, the updated cross-sectional parameters are obtained, and the process returns to step 103 until the iteration end condition is met, and the final cross-sectional parameters are used as the optimization results of the bar system component.
[0096] Set the learning rate α = 0.0001.
[0097] Perform gradient descent on the current cross-sectional parameters to obtain new cross-sectional parameters.
[0098] The gradient descent process is shown as follows: θ = θ - α × G.
[0099] Until the loss function changes by less than 1×10 6 , or the number of iterations reaches the threshold.
[0100] Figure 4 1, 2, 3, and 4 are the initial cross-sections of the I-shaped bar system and the cross-sections after 5, 10, and 15 iterations of optimization, respectively. The initial cross-section parameters are that the width and thickness of the upper and lower plates of the I-shaped cross-section are 1m and 0.2m respectively; the width and thickness of the middle vertical plate are 0.96m and 0.2m respectively. Each iteration of optimization will adjust the cross-section parameters to a certain extent based on the previous one, so that the cross-section will continue to evolve towards the local optimal cross-section. Figure 5The various indicators in the section optimization process are displayed, including the changes in cross-sectional area, maximum cross-sectional width-to-thickness ratio, and maximum cross-sectional stress. It can be seen that although the stress and width-to-thickness ratio of the initial section meet the specified requirements (the maximum cross-sectional stress is less than 300MPa, and the maximum cross-sectional width-to-thickness ratio is less than 40), it is too conservative, the cross-sectional material is large, and the material properties are not fully utilized. As the optimization proceeds, the various cross-sectional parameters of the I-section are gradually adjusted, but the maximum stress and width-to-thickness ratio of the section are always controlled within the allowable range. When the section optimization is completed, the maximum cross-sectional stress is 297Mpa, which is less than the 300MPa allowed by the material; the maximum cross-sectional width-to-thickness ratio is 37, which is less than the allowable 40; and the cross-sectional area is increased from the initial 52dm 2 Down to 13dm 2 It can be seen that the cross-section optimization method can adjust the original bar cross-section so that the cross-section meets the specified requirements while reducing the cross-section material as much as possible and improving the material utilization rate to achieve the purpose of economy and environmental protection. In addition, compared with traditional methods, this optimization process does not require human intervention and can be completed in a very short time.
[0101] In the description of the present application, it should be noted that the terms "upper", "lower", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. Unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.
[0102] It should be noted that, in this application, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0103] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features applied for herein.
Claims
1. A method for automatic optimization of the cross section of a bar system member, characterized in that: It includes: Set the design goal of the verification result and use it as a construction condition to construct a loss function for the verification result; Based on the loss function, obtaining the gradient of the loss function; Initializing the cross-sectional parameters of the bar system component, and obtaining the internal force exerted on the bar system component based on the design load, and obtaining the calculation result based on the cross-sectional parameters and the internal force exerted; Based on the gradient of the loss function, the learning rate, the check result and the section parameters, the updated section parameters are obtained, and the check result is obtained again until the iteration end condition is met, and the final section parameters are used as the optimization results of the bar system components.
2. The method for automatically optimizing the cross section of a bar system member according to claim 1, characterized in that: The calculation results include the cross-sectional area of the bar system component, the maximum cross-sectional stress and the maximum width-to-thickness ratio of the plate section included in the bar system component; The design objectives of the calculation result include: a first design objective, a second design objective and a third design objective; Among them, the first design goal is that the maximum stress of the section does not exceed the first allowable value, the second design goal is that the maximum width-to-thickness ratio of the section does not exceed the second allowable value, and the third design goal is that the cross-sectional area of the bar system component reaches a minimum value under the premise of meeting the first design goal and the second design goal.
3. The method for automatic optimization of bar member cross section according to claim 1, characterized in that: Based on the gradient of the loss function, the learning rate, the check result and the cross-sectional parameters, the updated cross-sectional parameters are obtained, including: Substitute the check result into the gradient of the loss function and find the product of the learning rate and the gradient of the loss function; The product is subtracted from the cross-sectional parameter to obtain an updated cross-sectional parameter.
4. The method for automatically optimizing the cross section of a bar system member according to claim 1, characterized in that: The learning rate ranges from 0.0001 to 0.
001.
5. The method for automatically optimizing the cross section of a bar system member according to claim 4, characterized in that: The learning rate is set to 0.0001.
6. The method for automatically optimizing the cross section of a bar system member according to claim 1, characterized in that: The iteration termination condition includes that the change amplitude of the loss function is less than an amplitude threshold.
7. The method for automatically optimizing the cross section of a bar system member according to claim 6, characterized in that: The amplitude threshold is 1×10 6 .
8. The method for automatically optimizing the cross section of a bar system member according to claim 1, characterized in that: The iteration end condition includes that the number of iterations reaches a threshold number.
9. The method for automatically optimizing the cross section of a bar system member according to claim 8, characterized in that: The number threshold is 15 times.
10. An automatic optimization system for the cross section of a bar system member, characterized in that: It includes: The first module is used to: set the design target of the inspection result and use it as a construction condition to construct a loss function about the inspection result; The second module is used to: obtain the gradient of the loss function based on the loss function; The third module is used to: initialize the cross-sectional parameters of the bar system component, obtain the internal force exerted on the bar system component based on the design load, and obtain the calculation result based on the cross-sectional parameters and the internal force exerted; The fourth module is used to obtain updated cross-sectional parameters based on the gradient of the loss function, the learning rate, the check result and the cross-sectional parameters, and return to obtain the check result until the iteration end condition is met, and use the final cross-sectional parameters as the optimization result of the bar system component.
