Design method of split cabin clamping device based on expert system

Abstract

The invention relates to an expert system-based split cabin clamping device design method, which is characterized by comprising the following steps of: 1, constructing a split cabin clamping design expert system which comprises a user interface, a knowledge base, a rule base, an interpretation machine and an inference machine; 2, constructing and training a neural network prediction model; the neural network prediction model comprises an input layer, a middle layer and an output layer; 3, key design parameters of the split cabin to be machined are input through the user interface; fourthly, the inference engine conducts inference calculation on the input parameters, and a complete design scheme of the clamping device is generated; and 5, outputting the design scheme, wherein the scheme comprises the overall layout of the clamping device, the number of supporting points, the layout, the selection of a positioning surface and the range of clamping force. By constructing a knowledge base and an expert system reasoning mechanism, the intelligent design of the clamping device is realized, and the design efficiency is remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of mechanical design and automation technology, and specifically relates to a split-body clamping device design method based on an expert system. BACKGROUND

[0002] In the field of aerospace, split-body clamping devices are particularly important due to their weak rigidity and open structure. Traditional clamping device design methods rely heavily on the experience of designers and trial-and-error methods, resulting in long design cycles, low efficiency, and difficulty in meeting the high-precision clamping requirements of complex body structures. In addition, traditional design methods lack intelligent support, making it difficult to accurately optimize the clamping scheme during the design process, resulting in poor versatility and adaptability of the clamping device.

[0003] In recent years, with the rapid development of artificial intelligence technology, expert systems, as an intelligent system based on knowledge base and reasoning mechanism, have been gradually applied to the field of complex engineering design. However, there is currently no mature method that combines expert system technology with split-body clamping device design, and existing technologies still have the following problems: 1. Low design efficiency: The design process relies heavily on human experience for different specifications of the body, lacks intelligent design tools, and cannot quickly respond to production needs; 2. Lack of versatility: Traditional clamping devices are mostly one-size-fits-all, and when the diameter of the body at both ends, the length of the hull, and other parameters change, the original tooling needs to be significantly modified or scrapped, resulting in a significant increase in production costs and serious waste of resources; 3. Low knowledge reuse: The knowledge of successful cases, fault solutions, and other knowledge in clamping design is stored in unstructured documents, and there is no reusable system. New engineers need to relearn and accumulate to independently complete the design, resulting in high knowledge transfer costs.

[0004] 4. Lack of targeted protection for weak rigidity areas. Thin-walled bodies are prone to local elastic deformation or even permanent damage under clamping force, and traditional design methods cannot accurately predict and optimize the layout of support points and the size of clamping force, resulting in workpiece springback after processing and out-of-tolerance shape and position.

[0005] Therefore, there is an urgent need for a split-body clamping device design method based on an expert system to improve design efficiency, optimize clamping device performance, and enhance its versatility and adaptability. SUMMARY

[0006] The present application aims at overcoming the deficiencies of the prior art, and provides a split type cabin clamping device design method based on an expert system.

[0007] To solve the above technical problems, the present application provides a split type cabin clamping device design method based on an expert system, characterized in that it comprises the following steps: Step one: constructing a split type cabin clamping design expert system, which comprises a user interface, a knowledge base, a rule base, an explanation machine and a reasoning machine.

[0008] The user interface provides a communication interface between the user and the expert system. The user inputs design parameters through the interface, and the system outputs design results and intermediate process information through the interface.

[0009] The knowledge base is used to store the deterministic knowledge and case-based knowledge in the field of split type cabin clamping design. Specifically, it includes: (1) Geometric model library: stores three-dimensional models and shape parameters of various typical split type cabins.

[0010] (2) Clamping force experience library: stores the empirical values of clamping force under different size and shape cabin conditions.

[0011] (3) Case library: stores various successful clamping design scheme instances, including cabin size parameters, clamping schemes and use effect scores.

[0012] The rule base is used to store design rules refined based on the experience of field experts, which are the basis for automatic reasoning of the expert system. Mainly includes: (1) Support point layout rule: "Support points should be symmetrically arranged at both ends of the cabin", "The distribution density of support points in the weak stiffness area should not be less than 8 / m 2 ".

[0013] (2) Positioning surface selection rule: "Follow the six-point positioning principle and avoid over-positioning", "The flatness error of the positioning reference surface should be controlled within 0.02mm".

[0014] (3) Obstacle avoidance rule: ensure that the fixture assembly does not interfere with the main shaft of the machine tool, tool path, etc.

[0015] The reasoning machine obtains the reasoning result by using forward reasoning and neural network mechanism according to the initial parameters input by the user, matches the rule premise in the rule base, activates the applicable rules, and executes the conclusion part, so as to gradually deduce the clamping scheme.

[0016] The explanation machine can explain the behavior of the expert system to the user, including explaining the correctness of the reasoning conclusion and the reason for the system outputting other subsequent candidate solutions.

[0017] Step two: build and train a neural network prediction model. The neural network prediction model includes an input layer, an intermediate layer and an output layer. The input layer input parameters are cabin length, cabin diameter at both ends, key assembly surface precision requirement, average wall thickness, and each section taper. The output layer output parameters are support point number, support point layout, positioning surface selection, and clamping force recommendation range.

[0018] Generally, in order to achieve an accurate prediction model, the expected output value is input into the system in step two. In this system, the expected output value is the historical optimal value of the clamping device design of a certain shape and size of split-body cabin given by an expert, which is also the ideal data reached by the design scheme. After inputting the expected output value into the system, the error between the actual output value of the system and the expected output value (i.e. the optimal value) is calculated using the back propagation algorithm. If the error cannot meet the requirements, the network returns to the original path for weight adjustment until the convergence error meets the requirements, and the training is completed.

[0019] Step three: input the key design parameters of the split-body cabin to be processed through the user interface, including but not limited to cabin length, cabin diameter at both ends, key assembly surface precision requirement, average wall thickness, and each section taper.

[0020] Step four: the reasoning machine performs reasoning calculation on the input parameters to generate a complete design scheme of the clamping device. First, search in the knowledge base. If there is a historical case with a matching degree of ≥ 90% with the current input parameters, the system preferentially adopts the clamping scheme of the case and makes adaptive modifications based on the rule base to generate an initial scheme. If the matching degree is insufficient, the neural network mechanism is started. The reasoning machine pre-processes the input parameters into a feature vector required by the neural network, and calls the trained neural network prediction model in step two. The neural network predicts the number of support points, layout, positioning surface selection and clamping force range of the new type of cabin based on the learned internal rules. The generated new output parameters are input into the knowledge base. Then, the reasoning machine is started, and the rules in the rule base about specific structure selection of the clamp, detailed modeling, obstacle avoidance checking, etc. are called to convert the parameters predicted by the neural network into a specific and executable clamping device design scheme.

[0021] Step five: output the design scheme, which includes the overall layout of the clamping device, the number of support points, the layout, the positioning surface selection and the clamping force range. Design personnel can form a three-dimensional design model for production based on the final design scheme for the use of split-body cabin clamping.

[0022] The beneficial effects of the present application are as follows: 1. This invention achieves intelligent design of clamping devices by constructing a knowledge base and an expert system reasoning mechanism, which significantly improves design efficiency.

[0023] 2. By introducing a neural network, this invention has the capability to process novel split-type cabin designs that exceed the scope of existing knowledge bases, thus overcoming the limitations of traditional expert systems.

[0024] 3. This invention has high versatility and scalability, and can be applied to the clamping design of various complex cabins. Attached Figure Description

[0025] Figure 1 1 is a flowchart of the design of this invention; Figure 2 The diagram below shows the structure of the model built using a neural network in this invention.

[0026] Figure 3 The diagram below shows the layout of support points in a specific embodiment of the present invention.

[0027] Figure 4 1 is a schematic diagram of the positioning surface in a specific embodiment of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] This embodiment details a specific application of a design based on the method of the present invention. Figure 1 A flowchart of the design method for the expert system-based split-type cabin clamping device is shown.

[0030] Step 1: Construct an expert system for the clamping design of split-type cabins. The expert system includes a user interface, a knowledge base, a rule base, an interpreter, and an inference engine.

[0031] The user interface provides a communication interface between the user and the expert system. The user inputs design parameters through this interface, and the system outputs design results and intermediate process information through this interface.

[0032] The knowledge base is used to store deterministic and case-based knowledge in the field of split-type hull clamping design. Specifically, it includes: (1) Geometric model library: stores the three-dimensional models and shape parameters of various typical split-type cabins.

[0033] (2) Clamping force experience library: stores the experience values ​​of clamping force for different sizes and shapes of cabins.

[0034] (3) Case Library: Stores various successful clamping design examples, including cabin size parameters, clamping schemes and usage effect ratings.

[0035] The rule base stores design rules derived from the experience of domain experts. These rules form the basis for the expert system's automated reasoning. They mainly include: (1) Support point layout rules: "Support points should be symmetrically arranged at both ends of the cabin", "The distribution density of support points in the weak stiffness area should not be less than 8 points / m 2 ".

[0036] (2) Rules for selecting positioning surfaces: “Follow the six-point positioning principle and avoid over-positioning” and “The flatness error of the positioning reference surface must be controlled within 0.02mm”.

[0037] (3) Obstacle avoidance rules: Ensure that the fixture assembly does not interfere with the machine tool spindle, tool path, etc.

[0038] The inference engine obtains inference results based on the initial parameters input by the user using forward inference and neural network mechanisms. It then matches these results with the rule premises in the rule base, activates the applicable rules, and executes their conclusions, thereby gradually deriving the clamping scheme.

[0039] The explanation machine can explain the behavior of the expert system to the user, including explaining the correctness of the reasoning conclusions and the reasons why the system outputs other subsequent candidate solutions.

[0040] Step Two: Build and train the neural network prediction model. (Appendix) Figure 2 The diagram illustrates the structure of a model built using a neural network. The neural network prediction model includes an input layer, an intermediate layer, and an output layer. The input parameters for the input layer are the cabin length, diameters at both ends of the cabin, precision requirements for key assembly surfaces, average wall thickness, and taper of each section. The output parameters for the output layer are the number of support points, support point layout, selection of positioning surfaces, and recommended clamping force range.

[0041] To achieve an accurate prediction model, step two typically includes inputting the desired output value into the system. In this system, the desired output value is the historical optimal value of the clamping device design for a split-type hull of a certain shape and size, provided by experts. It is also the ideal data achieved by the design scheme. After inputting the desired output value into the system, the backpropagation algorithm is used to calculate the error between the actual output value and the desired output value (i.e., the optimal value). If the error does not meet the requirements, the network returns along the original path to adjust the weights until the convergence error meets the requirements, at which point training ends.

[0042] Step 3: Input the key design parameters of the split-type cabin to be processed through the user interface: length of each conical section 327mm, 300mm, 680mm, diameter of the large end 514mm, diameter of the small end 168mm, wall thickness 5mm, taper of each section 15.44°, 8.19°, 3.87°, required processing accuracy ±0.05mm, and flatness requirement of the mating surface 0.02.

[0043] Step Four: The inference engine performs reasoning calculations on the input parameters to generate a complete design scheme for the clamping device. First, it searches the knowledge base. If no historical cases with a matching degree ≥90% with the current input parameters are found, the neural network mechanism is activated. The inference engine preprocesses the input parameters into feature vectors required by the neural network and calls the neural network prediction model trained in Step Two. Based on the learned inherent rules, the neural network predicts key parameters such as the number and layout of support points, the selection of positioning surfaces, and the clamping force range for this new type of cabin. The generated new output parameters are input into the knowledge base. Subsequently, the inference engine starts, calling rules in the rule base regarding the specific structure selection of the clamp, detailed modeling, obstacle avoidance checks, etc., to transform the parameters predicted by the neural network into a specific and executable clamping device design scheme.

[0044] Step 5: Output the design scheme as follows: The clamping device base size is 1700mm × 800mm, with 8 support points. The center of the base is the zero point, and the cabin placement axis is the +y axis from the large end to the small end. The support point layout coordinates are (0, -520), (0, -100), (0, 285), (0, 540), (220, -530), (152, 65), (-220, -530), (-152, 65). The positioning surfaces are the mating surface, the large end face, and the lateral process chuck reference surface. The clamping force range is 80 N·m to 100 N·m. (Attachment) Figure 3 A schematic diagram of the support point layout of the clamping device under the aforementioned design scheme is shown, with appendix. Figure 4 A schematic diagram of the positioning surface of the clamping device under the described design scheme is shown. Designers use the final design scheme to create a 3D design model for production, intended for use in clamping open-type cabins.

[0045] It should be noted that, in this document, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0046] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for designing a split case clamping device based on an expert system, characterized by The method comprises the following steps: Step 1: constructing a split-body clamping design expert system, wherein the expert system comprises a user interface, a knowledge base, a rule base, an explanation machine and a reasoning machine; The user interface provides a communication interface between the user and the expert system; The user inputs the clamped body size parameters through the interface, and the expert system outputs the design results and intermediate process information through the interface; The knowledge base is used for storing the deterministic knowledge and case-based knowledge in the field of split-body clamping design; specifically comprising: (1) a geometric model library: storing three-dimensional models and shape parameters of various typical split bodies; (2) a clamping force experience library: storing empirical values of clamping force under different size and shape conditions of the body; (3) a case library: storing various successful clamping design scheme instances, including body size parameters, clamping schemes and use effect scores; The rule base is used for storing design rules refined based on the experience of experts in the field, which are the basis for automatic reasoning of the expert system; mainly including: (1) Support point layout rule: "Support points should be symmetrically arranged at both ends of the cabin body", "The distribution density of support points in the weak stiffness area should not be less than 8 / m 2 "; (2) positioning surface selection rules: "follow the six-point positioning principle and avoid over-positioning", "the flatness error of the positioning reference surface needs to be controlled within 0.02mm"; (3) obstacle avoidance rules: ensuring that the fixture assembly does not interfere with the main shaft of the machine tool, tool path, etc.; The reasoning machine obtains the reasoning result by using forward reasoning and neural network mechanism according to the size parameters of the body device input by the user, matches the rules in the rule base, activates the applicable rules, and executes the conclusion part, thereby gradually deducing the clamping scheme; The explanation machine can explain the behavior of the expert system to the user, including explaining the correctness of the reasoning conclusion and the reasons for the system outputting other subsequent candidate solutions; Step 2: constructing and training a neural network prediction model; the neural network prediction model comprises an input layer, an intermediate layer and an output layer; the input parameters of the input layer are body length, body diameter at both ends, key assembly surface precision requirement, average wall thickness and taper of each section; the output parameters of the output layer are support point number, support point layout, positioning surface selection and clamping force recommendation range; In order to achieve an accurate prediction model, the expected output value is input into the system in step 2; in this system, the expected output value is the historical optimal value of the clamping device design of a split body of a certain shape and size given by the expert, which is also the ideal data reached by the design scheme; after inputting the expected output value into the system, the error between the actual output value of the system and the expected output value, i.e. the optimal value, is calculated by using the back propagation algorithm; if the error cannot meet the requirements, the network returns to the original path for weight adjustment until the convergence error meets the requirements, and the training is completed; Step 3: inputting the key design parameters of the split body to be machined through the user interface, including but not limited to body length, body diameter at both ends, key assembly surface precision requirement, average wall thickness and taper of each section; Step four: the inference engine performs inference calculation on the input parameters to generate a complete design scheme of the clamping device; first, search in the knowledge base, if there is a historical case with a matching degree ≥ 90% with the current input parameters, the system preferentially adopts the clamping scheme of the case, and makes adaptive modification based on the rule base to generate an initial scheme; if the matching degree is insufficient, the neural network mechanism is started; the inference engine pre-processes the input parameters into a feature vector required by the neural network, and calls the trained neural network prediction model in step two; the neural network predicts the number of support points, layout, positioning surface selection and clamping force range of the new type of cabin body according to the learned internal law; input the generated new output parameters into the knowledge base; then, the inference engine is started, and the rules in the rule base about the specific structure selection, detailed modeling, obstacle avoidance checking and the like of the clamp are called, and the parameters predicted by the neural network are converted into a specific and executable clamping device design scheme; Step five: output the design scheme, which includes the overall layout, the number of support points, the layout, the positioning surface selection and the clamping force range of the clamping device; the designer can form a three-dimensional design model for production according to the final design scheme, which is used for open cabin body clamping.

Images