CN-121980972-A - Intelligent process planning method and system integrating equipment manufacturing capability constraint

CN121980972ACN 121980972 ACN121980972 ACN 121980972ACN-121980972-A

Abstract

The invention provides a process intelligent planning method and system integrating equipment manufacturing capacity constraint, and belongs to the technical field of computer-aided process planning and intelligent manufacturing. The method comprises the steps of establishing a multidimensional manufacturing resource constraint model comprising processing capacity, movement space and clamping feasibility, solving a clamping scheme by adopting a multi-objective evolutionary algorithm by taking the constraint model as a boundary, constructing an attribute adjacency graph based on the determined clamping scheme, dynamically identifying processable features by adopting a graph neural network, fusing the constraint model and process knowledge, intelligently generating detailed technological rules comprising processing sequence, tool selection and cutting parameters by adopting a mixed particle swarm optimization algorithm, and taking the equipment manufacturing capacity as a front hard constraint of process planning to ensure that the generated technological scheme has high engineering feasibility.

Inventors

ZHENG ZUJIE
CHEN XUEFEN
SUN LIANG
JIANG RENZHENG
SHEN CAIXIA
SHEN FANGFANG
JIAO BINBIN
CAO XUEWEN
CHEN LONG
SHEN YI

Assignees

上海航天精密机械研究所

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (10)

1. An intelligent process planning method for fusing equipment manufacturing capacity constraints is characterized by comprising the following steps: Step 1, establishing a multi-dimensional manufacturing resource constraint model, wherein the multi-dimensional manufacturing resource constraint model comprises a processing capacity constraint model, a motion space constraint model and a clamping feasibility constraint model; Step 2, taking the machining capacity constraint model, the motion space constraint model and the clamping feasibility constraint model which are established in the step 1 as boundaries, inputting an MBD model of a part, establishing a clamping scheme multi-target optimization model, solving the clamping scheme multi-target optimization model by adopting a decomposition-based multi-target evolutionary algorithm, and outputting a Pareto optimal solution set for selection of process personnel to obtain a determined clamping scheme; Step 3, for the clamping scheme determined in the step 2, carrying out coordinate transformation on the MBD model, representing the transformed MBD model as an attribute adjacency graph, carrying out feature learning on the attribute adjacency graph by adopting a graph neural network, decomposing the complex attribute adjacency graph into a plurality of basic subgraphs through a graph segmentation algorithm, carrying out feature recognition, and checking whether each recognized feature is processable under the current clamping pose; And 4, establishing a process route optimization model aiming at the processing characteristic set of each process identified in the step 3, solving the process route optimization model by adopting a mixed particle swarm optimization algorithm, and outputting an optimal process scheme.
2. The process intelligence planning method of fusing equipment manufacturing capability constraints of claim 1, wherein the process capability constraint model MC is represented as a five-tuple set: Wherein, the Denoted j-th machine tool; is the maximum cutting power; ={ , ,..., is the set of available tools of number n, Is a feed speed range; is the main shaft rotating speed range; The motion space constraint model Representing the range of motion of each axis of the machine tool and the interference conditions with the workpiece/fixture: Wherein, the 、、 A X, Y, Z shaft travel limit; For a range of pivot angles, defining a tool reachability determination function R (p, o, , ) For any point p and cutter shaft direction o on the workpiece, if a cutter exists So that the cutting point is positioned in the working space of the machine tool and has no interference, the function value is 1, otherwise, the function value is 0; the clamping feasibility constraint model comprises positioning stability constraint and clamping deformation constraint: The positioning stability constraint is based on a static equilibrium condition, and for a given number of m positioning point sets l= { , ,..., -To satisfy: Where i is the sequence index of m, Is the counter-force of the branch at the locating point, For the weight of the workpiece, In order for the cutting force to be high, 、、 The clamping deformation constraint adopts a finite element method to calculate the maximum deformation of the workpiece under the action of the clamping force The requirements are as follows: Wherein, the For the displacement at the workpiece surface point p, S is the workpiece surface point set, Is the maximum deformation allowed.
3. The intelligent process planning method for fusing equipment manufacturing capacity constraints according to claim 2, wherein in step 2, the multi-objective optimization model of the clamping scheme is as follows: Decision variable, clamping frequency N, positioning reference surface set { of each clamping , ,..., By rotating matrix Translation vector The pose of the represented workpiece coordinate system relative to the machine tool coordinate system; objective function: Wherein, the method comprises the steps of, To minimize the number of clips; In order to minimize the reference plane switching cost, To maximize the total number of single clamping processable features; Constraint conditions, in the optimization process, carrying out feasibility verification of the following solutions based on the constraint model established in the step 1: positioning stability constraints, that is, for each clamping, there is a clamping force vector So that the workpiece meets the static balance condition and the counter force is not negative; Clamping deformation constraint: wherein Maximum deformation under the ith clamping; Tool accessibility constraint, for each clamping feature set to be processed Wherein at each point p on each feature, there is an arbor direction o such that R (p, o, , )=1。
4. A process intelligence planning method according to claim 3 in which the process of the decomposition-based multi-objective evolutionary algorithm in step 2 comprises: Coding scheme using hybrid coding, the chromosome is composed of N gene segments, each segment containing reference plane coding And attitude angle coding ), In order to be the roll angle, Is used as a pitch angle of the light beam, Is a yaw angle; weight vector generation, namely generating weight vectors which are uniformly distributed , ,..., ) Each weight vector corresponds to one sub-problem; chebyshev decomposition, namely decomposing the multi-objective problem into K single-objective sub-problems, wherein the objective function of the q-th sub-problem is as follows: Wherein, the =( , , ) Taking the optimal value of each target in the current population as a reference point; Representing the q-th weight vector; representing the weight component of the qth weight vector on the ith target; is the i-th objective function value corresponding to the decision variable x; The neighborhood definition, wherein each weight vector defines a neighborhood B (q) and comprises T weight vectors closest to the Euclidean distance of the neighborhood B (q); evolutionary operation, namely adopting simulated binary crossover and polynomial variation; Constraint processing, namely processing infeasible solutions by adopting a penalty function method; the algorithm termination condition is that the maximum iteration number is reached Or population convergence, and outputting a Pareto optimal solution set for selection by process personnel.
5. The intelligent process planning method for fusing equipment manufacturing capability constraints of claim 4, wherein in step 2, the decomposition-based multi-objective evolutionary algorithm employs a dynamic neighborhood strategy and an adaptive weight adjustment mechanism, comprising: wherein g is the current iteration number, For the maximum number of iterations to be performed, And Initial and final neighborhood sizes, respectively; Self-adaptive weight adjustment, namely self-adaptively adjusting weight vectors according to the distribution condition of the current population: Wherein, the In order for the rate of learning to be high, As a diversity loss function with respect to weight distribution, Representing the gradient of the loss function versus the weight; / the weight component of the ith target of the g/g+1th generation.
6. The intelligent process planning method for integrating equipment manufacturing capability constraints according to claim 5, wherein in step 3, for the ith clamping scheme determined in step 2, the ith clamping scheme is determined by a rotation matrix Translation vector The expressed pose of the workpiece is used for carrying out coordinate transformation on the MBD model, and the new coordinates of any point p on the workpiece after transformation are as follows: the GIN graph isomorphic network is adopted as a basic model of feature recognition, and embedded representation of each node is learned through multi-layer graph convolution ; The layer I graph convolution operation is defined as: complex sub-formation for intersecting features Wherein, the method comprises the steps of, For the initial attribute of layer i graph roll-up node v, As a neighborhood node, Is a set of neighborhood nodes for node v, In order for the parameters to be able to be learned, Is a multi-layer perceptron; decomposing the complex attribute adjacency graph into a plurality of basic subgraphs by adopting a graph segmentation algorithm, respectively identifying the basic subgraphs, checking whether each identified feature F can be processed under the current clamping pose, and { a sampling point set on the feature , ,..., -To satisfy: , is an arbitrary sampling point on the feature, Is in the direction of the cutter shaft, Is a cutter shaft direction set.
7. The intelligent process planning method for merging equipment manufacturing capability constraints according to claim 6, wherein in step 4, the process route optimization model is as follows: decision variables: Feature processing sequence by arrangement The representation is made of a combination of a first and a second color, For the feature index of the kth process, The number of index for processing features; Cutter selection of each feature From a set of available tools In selecting a tool ; Cutting parameters of each feature The cutting parameters of the cutting tool include spindle rotation speed, feeding speed, cutting depth and cutting width; objective function: Wherein, the The cutting time of the feature j is determined by the cutting parameters and the machining path length; For the tool changing time, if adjacent features use different tools, the tool changing time is generated; Is the auxiliary time; The method is the use cost of the machine tool; Cost is lost for the tool.
8. The intelligent process planning method for fusing equipment manufacturing capability constraints of claim 7, wherein in step 4, the coding scheme of the hybrid particle swarm optimization algorithm adopts a three-layer coding structure: The length of the operation layer is | The arrangement of I represents the feature processing sequence; Tool layer length | An integer vector of I, which represents the cutter number selected for each feature; Parameter layer length is | | 3, A real matrix representing the cutting parameters (n, f, ); Particle updating, namely replacing a speed-position updating formula of the traditional PSO by adopting a genetic operator; the cross operation, namely, sequentially crossing an operation layer and crossing two points of a cutter layer and a parameter layer; the mutation operation comprises the steps of adopting exchange mutation for an operation layer, adopting random reset for a cutter layer and adopting Gaussian disturbance for a parameter layer: Wherein, the Is a new parameter; is the original cutting parameter; fitness function: Wherein, the 、 As the weight coefficient of the light-emitting diode, Is a penalty factor; For the total processing time corresponding to the decision variable x, The violation amount for the j-th constraint; the termination condition is that the maximum iteration times or continuous multiple generations are reached without improvement; The algorithm outputs the optimal technological scheme including machining sequence, cutter selection and cutting parameters, and distributes machining allowance of each characteristic according to the rough machining, semi-finishing and finishing strategies to generate complete procedure/step content.
9. The process intelligence planning method for fusing equipment manufacturing capability constraints of claim 8, wherein in step 4, the hybrid particle swarm optimization algorithm comprises an adaptive parameter adjustment mechanism: Inertial weight self-adaptive adjustment: Wherein w (g) is the inertia weight of g generation, And (3) with Respectively the maximum inertia weight and the minimum inertia weight, g is the current iteration number, For the maximum iteration times, sigma is a random disturbance coefficient, and N (0, 1) is a standard normal distribution random number; Cross probability and variation probability adaptive adjustment: Wherein, the Cross probability for the g generation; The maximum crossover probability is obtained, and alpha is the attenuation coefficient; Variation probability of the g generation; 、 the minimum and maximum mutation probabilities are respectively.
10. A process intelligence planning system incorporating equipment manufacturing capability constraints, characterized in that a process intelligence planning method incorporating equipment manufacturing capability constraints according to any one of claims 1 to 9 is employed, comprising: the method comprises the steps of M1, establishing a multi-dimensional manufacturing resource constraint model, wherein the multi-dimensional manufacturing resource constraint model comprises a processing capacity constraint model, a motion space constraint model and a clamping feasibility constraint model; The module M2 takes the processing capacity constraint model, the motion space constraint model and the clamping feasibility constraint model established by the module M1 as boundaries, inputs an MBD model of a part, establishes a clamping scheme multi-target optimization model, adopts a decomposition-based multi-target evolutionary algorithm to solve the clamping scheme multi-target optimization model, and outputs a Pareto optimal solution set for selection of process personnel to obtain a determined clamping scheme; the module M3 is used for carrying out coordinate transformation on the MBD model according to the clamping scheme determined by the module M2, representing the transformed MBD model as an attribute adjacency graph, carrying out feature learning on the attribute adjacency graph by adopting a graph neural network, decomposing the complex attribute adjacency graph into a plurality of basic subgraphs through a graph segmentation algorithm, carrying out feature recognition, and checking whether each recognized feature is processable under the current clamping pose; And the module M4 is used for establishing a process route optimization model aiming at the processing characteristic set of each procedure identified by the module M3, solving the process route optimization model by adopting a mixed particle swarm optimization algorithm, and outputting an optimal process scheme.

Description

Intelligent process planning method and system integrating equipment manufacturing capability constraint Technical Field The invention relates to the technical field of Computer Aided Process Planning (CAPP) and intelligent manufacturing, in particular to a process intelligent planning method and system fusing equipment manufacturing capacity constraint. Background Computer Aided Process Planning (CAPP) technology is a bridge connecting product design with actual manufacturing, and its level of intelligence directly affects product quality, production cost, and production cycle. Currently, intelligent process planning methods focus on mathematical optimization solutions based on product geometric models and processing features, and theoretical optimal process routes are generated through feature recognition, process reasoning and path optimization. However, such methods tend to ignore physical limitations of manufacturing resources and engineering practical constraints, resulting in low applicability of the planning results in a practical production environment, and a low proportion of being directly applicable to production. In actual manufacturing processes, particularly in the field of aerospace manufacturing, the solution space of the process is often severely constrained by existing manufacturing equipment. The manufacturing equipment is typically predetermined and the structural member requires the completion of the process on a particular piece or pieces of equipment. Under the condition that equipment is basically determined, a process route and specific process parameters which meet quality requirements and have cost advantages are planned by taking actual processing capacity of equipment as a boundary, so that the key technical problem to be solved is urgent. In the prior art, part of research attempts to introduce manufacturing resource constraint in the process planning process, but a post-verification mode is mostly adopted, namely, a process scheme is generated first and then feasibility verification is carried out, and the scheme is frequently modified repeatedly due to low efficiency of the scheme of 'planning before verification'. In addition, the modeling of the manufacturing equipment by the existing method is rough, a multidimensional constraint model comprising processing capacity, movement space and clamping feasibility cannot be systematically established, and deep fusion of the constraint model and a process generation process cannot be realized. Therefore, development of an intelligent method for using equipment manufacturing capability as hard constraint and running constraint models through the whole process of process planning is needed, so that the generated process scheme is not only theoretically optimal, but also has the possibility of successful implementation at one time in a physical level. Patent application document CN121328980A discloses a reconfigurable flexible job shop scheduling optimization method with secondary clamping constraint, and relates to the technical field of intelligent manufacturing and production optimization. The method comprises the following steps of 1) establishing a mixed integer linear programming model considering a secondary clamping reconfigurable flexible job shop scheduling problem with the aim of minimizing maximum finishing time and minimizing the number of workers, 2) designing a three-section coding mode and a decoding mode corresponding to the mixed integer linear programming model based on procedure sequencing, machine selection and worker selection, and 3) solving an optimal scheduling scheme of the mixed integer linear programming model by adopting an improved multi-objective genetic algorithm based on the three-section coding mode and the decoding mode. However, the present patent cannot completely solve the existing technical problems, and cannot meet the needs of the present invention. Disclosure of Invention Aiming at the defects in the prior art, the invention aims to provide a process intelligent planning method and system for fusing equipment manufacturing capacity constraint. The intelligent process planning method for fusing equipment manufacturing capacity constraint provided by the invention comprises the following steps: Step 1, establishing a multi-dimensional manufacturing resource constraint model, wherein the multi-dimensional manufacturing resource constraint model comprises a processing capacity constraint model, a motion space constraint model and a clamping feasibility constraint model; Step 2, taking the machining capacity constraint model, the motion space constraint model and the clamping feasibility constraint model which are established in the step 1 as boundaries, inputting an MBD model of a part, establishing a clamping scheme multi-target optimization model, solving the clamping scheme multi-target optimization model by adopting a decomposition-based multi-target evolutionary algorithm, and outputting a P