Milling tool path generation method and system based on machining guide line offset

Abstract

The application belongs to the technical field of numerical control machining, and discloses a milling tool path generation method and system based on machining guide line offset, wherein the method comprises the following steps: S1, obtaining a simplified grid model of a part to be machined; S2, determining a machining area on the simplified grid model, and converting a tool path generation problem of the machining area into a tool path point access sequence problem; S3, solving the tool path point access sequence problem of the machining area, and generating a machining guide line; and S4, performing offsetting, elongation, cutting and connection processing on the machining guide line, obtaining an offset line, and taking the offset line as a moving line in a tool path of the part to be machined to generate the tool path. The machining guide line is generated based on the optimal access sequence of discrete path points, the tool machining moving line generated based on the machining guide line has less redundant cutting, and the machining efficiency is higher; and the tool path generation efficiency is higher.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field related to numerical control machining, and more particularly to a milling tool path generation method and system based on machining guide line offset. BACKGROUND

[0002] Numerical control technology and numerical control machine tools are the foundation of the manufacturing industry revolution. Most mechanical parts widely demanded in actual engineering are 2.5D parts, and the machining of most complex parts can also be completed by 2.5D rough machining of the blank and 3D-5D fine machining. Therefore, 2D cavity milling is the most commonly used numerical control milling type in the machining process of ordinary parts. The tool path of complex cavity milling usually has different generation schemes, and different tool path generation schemes will affect the machining time and machining cost of the part during the machining process. Therefore, in order to reduce the machining cost and improve the machining efficiency, it is necessary to design an efficient tool path generation scheme according to the features of the part to be machined, especially the complex cavity milling features.

[0003] At present, for two-dimensional cavity milling tasks, common tool path generation methods include zigzag cutting mode and offset ring cutting mode. If these methods are used directly, the generated tool path has a large amount of redundant cutting, the machining efficiency is low, and for complex cavity milling, the existing tool path generation method also has the problems of complex calculation and low path generation efficiency. In the actual machining process, it is necessary to design the tool path according to the features of the part to be machined. Therefore, a more efficient tool path generation method is needed to make up for the defects of traditional methods in terms of machining efficiency. SUMMARY

[0004] In view of the above defects or improvement needs of the prior art, the present application provides a milling tool path generation method and system based on machining guide line offset, which is used to solve the problems that the tool path generated by the existing tool path generation method for two-dimensional cavity milling tasks has a large amount of redundant cutting, the machining efficiency is low, and the tool path generation method has complex calculation and low path generation efficiency. The method and system utilize the simplified grid model of the part to be machined, design the tool path according to the features of the part to be machined, have less redundant cutting, and have high machining efficiency. At the same time, the machining guide line is generated based on the simplified grid model, so that the path generation has lower calculation complexity, and the efficiency of tool path generation is higher.

[0005] To achieve the above purpose, according to one aspect of the present application, a milling tool path generation method based on machining guide line offset is provided, comprising:

[0006] S1, obtaining a simplified grid model of a part to be machined, the grid length of the simplified grid model being greater than the cutting width of the tool;

[0007] S2, determining a to-be-processed region on the simplified grid model, taking a grid center point of the simplified grid model as a tool path point, and converting a tool path generation problem of the to-be-processed region into a tool path point access sequence problem;

[0008] S3, solving the tool path point access sequence problem of the to-be-processed region, obtaining an optimal access sequence of the tool path point of the to-be-processed region, and generating a machining guide line;

[0009] S4, performing offsetting, elongation, cutting, and connection processing on the machining guide line, obtaining an offset line, and taking the offset line as a moving line in a tool path of the to-be-processed part to generate the tool path.

[0010] According to another aspect of the present application, a milling tool path generation system based on machining guide line offsetting is provided, comprising:

[0011] a grid establishing module configured to obtain a simplified grid model of a to-be-processed part, the grid length of the simplified grid model being greater than the cutting width of a tool;

[0012] a path point determining module configured to determine a to-be-processed region on the simplified grid model, take a grid center point of the simplified grid model as a tool path point, and convert a tool path generation problem of the to-be-processed region into a tool path point access sequence problem;

[0013] a guide line generating module configured to solve the tool path point access sequence problem of the to-be-processed region, obtain an optimal access sequence of the tool path point of the to-be-processed region, and generate a machining guide line;

[0014] a path generating module configured to perform offsetting, elongation, cutting, and connection processing on the machining guide line, obtain an offset line, and take the offset line as a moving line in a tool path of the to-be-processed part to generate the tool path.

[0015] Overall, compared with the prior art, the milling tool path generation method and system based on machining guide line offsetting provided by the present application have the following advantages:

[0016] 1. The tool path generation problem is converted into a discrete path point access sequence solving problem by establishing a simplified grid model, the optimal access sequence of the discrete path point is taken as a machining guide line, the machining guide line has a less repeated path point access process, the tool path generation moving line based on the machining guide line has less redundant cutting, and the machining efficiency is higher;

[0017] 2. The method generates machining guide lines first, and then obtains offset lines through offset, extension, cutting and connection processing of the machining guide lines, which expands the machining range based on the machining guide lines, can realize effective cutting of the machining area, and has low calculation complexity and high efficiency of tool path generation because the calculation process of the machining guide lines generated based on the simplified grid model is simple due to the large grid length and small number of grid center points of the simplified grid model;

[0018] 3. The machining guide lines are generated by combining the reinforcement learning method, which can generate machining guide lines autonomously according to the characteristics of the machining part, and generate the final tool path based on the machining guide lines, so that the path generation process can reduce the dependence on manual work and improve the intelligent degree while ensuring the machining quality;

[0019] 4. The Monte Carlo method, i.e. the multi-layer value simulation algorithm, is used to solve the tool path generation problem, which can quickly and accurately calculate the value of a state to guide the agent to complete the reinforcement learning task; compared with the Monte Carlo algorithm and the time difference algorithm, the method avoids the instability caused by random sampling; compared with the dynamic programming algorithm, the method reduces the value of the state that will never appear in the calculation process; and the method does not store a large amount of state value, and does not need a large amount of memory space and complex search operations on the environment. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is a flowchart of the milling tool path generation method based on machining guide line offset provided by the application;

[0021] Figure 2 is a flowchart of the method;

[0022] Figure 3 is a grid conversion process schematic diagram in the implementation process of the application;

[0023] Figure 4 is a grid merging process schematic diagram in the implementation process of the application;

[0024] Figure 5 is a machining guide line generation schematic diagram in the implementation process of the application;

[0025] Figure 6 is a machining guide line offset, connection and cutting schematic diagram in the implementation process of the application;

[0026] Figure 7 is a tool path extension schematic diagram in the implementation process of the application;

[0027] Figure 8is a tool path connection schematic diagram in the implementation process of the present application;

[0028] Figure 9 is a numerical control machining program generation schematic diagram in the implementation process of the present application;

[0029] Figure 10 is a simulation case diagram in the implementation process of the present application;

[0030] Figure 11 is a simulation experiment result diagram in the implementation process of the present application. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application. In addition, the technical features involved in each embodiment of the present application described below can be combined with each other as long as they do not conflict with each other.

[0032] Please refer to Figure 1 The embodiment one provides a milling tool path generation method based on machining guide line biasing, which comprises the following steps:

[0033] S1, obtaining a simplified grid model of a part to be machined, the grid length of the simplified grid model being greater than the cutting width of a tool;

[0034] S2, determining a machining area on the simplified grid model, taking the grid center point of the simplified grid model as a tool path point, and converting the tool path generation problem of the machining area into a tool path point access sequence problem;

[0035] S3, solving the tool path point access sequence problem of the machining area, obtaining the optimal access sequence of the tool path point of the machining area, and generating a machining guide line;

[0036] S4, biasing, extending, cutting and connecting the machining guide line, obtaining a biasing line, and taking the biasing line as a moving line in the tool path of the part to be machined to generate a tool path.

[0037] The embodiment establishes a simplified grid model of the part to be machined. Since the grid length of the simplified grid model is large and greater than the cutting width, the grid center points of the simplified grid model are relatively small. After the simplified grid model is established, the corresponding machining area on the simplified grid model is determined according to the features of the part to be machined. Therefore, the tool path generation problem of the machining area is the discrete path point access sequence problem of the corresponding grid center points of the machining area. The discrete node access sequence problem can be solved by using various existing algorithms. The optimal access sequence of the tool path points of the machining area is obtained by solving the problem. The optimal access sequence is the machining guide line.

[0038] The machining guide line is not the moving route of the final tool, but can guide the moving route of the tool. Based on the machining guide line, the offset route can be obtained by offsetting, extending, cutting and connecting processing. The offset route is located on the side of the machining guide line and is consistent with the trend of the machining guide line. The offset route obtained by offsetting, extending, cutting and connecting processing of the machining guide line can more comprehensively cover the machining area, and finally realize effective milling of the machining area.

[0039] Figure 2 The overall flowchart of the embodiment of the application is shown. The embodiment first divides the three-dimensional part into several layers by layered milling, and discretizes each layer into a square grid plane to obtain an initial grid model, that is, the part to be machined is represented by the initial grid model. Then, by a convolution-like operation, the features of adjacent grids in the grid plane are combined into one feature to obtain a small-scale grid model of the target part, that is, a simplified grid model, so as to convert the machining guide line generation problem into a reinforcement learning task on the small-scale grid plane. Then, an MDP model of the part machining guide line generation task is established based on the small-scale grid model according to the process requirements of the numerical control milling process. Then, the reinforcement learning task is solved by using a multi-layer value simulation method to obtain the decision sequence of the agent, so as to obtain the machining guide line of the part. Then, the generated machining guide line is mapped to the initial grid model of the target part, and the machining guide line is offset by an equal distance to both sides thereof in a segmented offset manner to remove the materials on both sides of the machining guide line. The offset generated curve is extended and cut to obtain a smooth and connected equal-distance offset curve to generate the moving route in the tool path. Finally, the optimal traversal sequence of the tool between different moving routes is solved to generate the connecting line between the feed points of the tool, so that the non-cutting movement distance of the tool between the feed points is minimized, thereby obtaining the final milling tool path. For example, the tool lifting and feeding sequence problem can be abstracted as a TSP problem, which is solved by a dynamic programming algorithm.

[0040] The entire implementation process is as follows:

[0041] 1、Through the way of layered milling, the three-dimensional part is divided into several layers, and each layer is discretized into a square grid plane, obtaining the initial grid model of the part. Through the operation similar to convolution, the features of adjacent grids in the grid plane are merged into one feature, obtaining the small-scale grid model of the target part, i.e. the simplified grid model. The grid conversion process is shown in Figure 3 As shown in the figure, for a given part milling task, it is necessary to first distinguish between the machining area and the non-machining area, determine the part contour according to the machining allowance Δ and the tool radius R, and divide the part contour into the outer contour and the inner contour. The closed area between them is the machining area, and the area other than this is the non-machining area.

[0042] Then, according to the size of the part and the cutting width, an empty grid area, i.e. the initial grid model, is constructed, in which the length of each sub-square grid is equal to the cutting width, and the cutting width is equal to the tool diameter. Next, the part features are mapped to the empty grid area. The grid contained in the machining area is the machining grid, and the grid other than this is the non-machining grid. According to the machining grid, the tool path points are constructed, obtaining the grid model of the part. Then, the features of adjacent grids in the initial grid model of the part are merged, for example Figure 3 The lower two figures in and Figure 4 As shown in the figure, four adjacent grids are merged into one grid. When all the grids in the four adjacent grids are cutting points, then the merged grid is a cutting point, otherwise the merged grid is a non-cutting point. That is, when the four adjacent grids are all contained in the machining area, the merged grid is a machining grid, and the center point of the grid is a tool path point.

[0043] By merging the features of the grids, the size of the initial grid model can be reduced, which can greatly reduce the calculation amount of the path access sequence problem. The small-scale grid model obtained after merging, i.e. the simplified grid model, retains most of the features of the target part, but its accuracy in representing the part is lower. The path generated by solving the small-scale grid model can reflect the general movement direction of the tool to some extent, and can be used as the machining guide line in the tool path generation process.

[0044] S1 specifically comprises: obtaining an initial grid model of a part to be machined, the grid length of the initial grid model being equal to the cutting width of a tool; and merging the grids in the initial grid model in groups to obtain the simplified grid model.

[0045] S1 specifically comprises: dividing the part to be machined into several machining layers through layered milling; and obtaining the simplified grid model of any machining layer.

[0046] Correspondingly, S2 specifically comprises: determining a contour line of any of the to-be-processed layers according to the preset machining allowance and the tool radius, determining a to-be-processed area of any of the to-be-processed layers based on the contour line; taking the center points of the grids of the simplified grid model contained in the to-be-processed area as tool path points, and converting the tool path generation problem of any of the to-be-processed layers into a tool path point access sequence problem on the grid plane.

[0047] In Figure 3 In the illustrated embodiment, the part contour is divided into an outer contour and an inner contour, and the enclosed area between the outer contour and the inner contour is the to-be-processed area. The outer contour is formed by inwardly retracting the outer boundary of the part, and the amount of inward retraction is the sum of the machining allowance Δ and the tool radius R; the inner contour is formed by outwardly expanding the inner boundary of the part, and the amount of outward expansion is the sum of the machining allowance Δ and the tool radius R. In other embodiments, the part contour can only include the inner contour or the outer contour, and the specific form is not limited.

[0048] 2, In the embodiment, S3 specifically comprises: S31, converting the tool path point access sequence problem of the to-be-processed area into a reinforcement learning task, establishing a Markov decision process model (MDP model) of the reinforcement learning task, wherein the agent in the MDP model is the tool, the environment is the state of the grid center points in the simplified grid model, the state space of the tool in the environment and the corresponding action space are defined, the immediate reward function of the tool action is set, after the tool performs an action according to the action space corresponding to the state space, the environment and the state space change and an immediate reward value corresponding to the immediate reward function can be obtained; S32, solving the MDP model based on the immediate reward value of the tool action, and generating the machining guide line.

[0049] Specifically, according to the process requirements of the numerical control milling process, the MDP model of the part machining guide line generation task is established based on the small-scale grid model, i.e., the simplified grid model.

[0050] The MDP model mainly includes the following parts:

[0051] (1) Agent and environment. The tool for processing a given part corresponds to the agent of the MDP model, and the part features and processing area represented by the planar grid in the simplified grid model correspond to the environment of the MDP model. That is, the environment in this embodiment is the grid center point state of the part area and the grid center point state of the to-be-processed area in the simplified grid model. Since there is only one element of the discrete tool path point in the environment, a feature matrix M can be created to represent the current state of the environment, the elements in the feature matrix correspond to the grid center points in the environment, and different values represent different states of the grid center points in the environment. For example, 0 represents a cut path point, that is, an area that has completed cutting, 1 represents an uncut path point, that is, an area that has not completed cutting, and -1 represents a non-processing path point, that is, an area that cannot be cut, and the values of the elements in the feature matrix are constantly updated with the interaction between the agent and the environment.

[0052] (2) State space and action space. The state of the tool not only contains the position, movement direction and other information of the tool, but also contains the state information of the surrounding environment, and the state space S of the tool in the environment is defined as:

[0053] S = [d, s0, s1, s2, s3, s4, s5, s6, s7, x, y];

[0054] In the formula, d represents the current movement direction of the tool, and takes 0-7, 0-7 represent 8 adjacent directions in a clockwise rotation of one circle from the top, respectively; s 0～7 represents the environment state information of the 8 adjacent directions around the discrete tool path point where the tool is located, s 0～7 is the environment state information of the 8 adjacent directions around the discrete tool path point where the tool is located in a clockwise rotation of one circle from the top, respectively; x and y represent the horizontal coordinate and vertical coordinate of the discrete tool path point where the tool is located, respectively. The action space A of the tool is defined as:

[0055] A = [0, 1, 2, 3, 4, 5, 6, 7, 8];

[0056] In the formula, 0-7 represent the movement of the tool from the current tool path point to the adjacent tool path points in 8 directions; 8 represents the tool lifting and feeding action. In order to reduce the number of tool lifting and feeding, the tool can only perform the lifting and feeding movement when there is no adjacent uncut path point in the 8 directions of the current tool path point, and select an uncut path point and a direction with an uncut path point to feed. At the same time, in order to reduce the search space of the agent, the action space of some special states should be processed. For example, when the tool is located at the boundary path point of the to-be-processed area, some actions will cause the tool to overcut, resulting in failure of part processing, so the unactionable actions are excluded from the action space of these states to obtain the actionable action space.

[0057] (3) Instantaneous reward function. The instantaneous reward function R(s, a) of the agent is composed of three types of reward functions, namely, the reward function of feed direction change, the reward function of tool lifting and feeding action, and the reward function of tool movement, which is the sum of the three reward functions, used to ensure the smoothness and continuity of the machining guide line and reduce the redundant path of the machining guide line. The reward function of feed direction change and the reward function of tool movement are designed to generate a tool path that meets the geometric process requirements and ensure the smoothness and continuity of the tool path. The reward function of tool lifting and feeding action is designed to reduce the tool idle stroke and shorten the total length of the tool path, thereby improving the machining efficiency. In any state of the agent, the instantaneous reward value that can be obtained by all actions in the action space can be calculated by the instantaneous reward function. Although the instantaneous reward value does not directly determine the decision of the agent, it can reflect the immediate income of a certain action to a certain extent and has a certain guiding effect on the decision of the agent.

[0058] In order to accurately represent the influence of different feed direction change amounts on machining efficiency in the instantaneous reward function, the reward function of feed direction change is obtained by the following method: performing machining experiments to measure the time cost of feed direction change caused by different feed direction angle change amounts at different feed speeds, and the measurement results are shown in Table 1; performing binary polynomial fitting on the experimental data to obtain the relationship between the time cost of feed direction change and the feed direction angle change amount and the feed speed; and obtaining the reward function of feed direction change based on the relationship. The time cost of feed direction change is the difference between the machining time when the feed direction changes and the machining time when the feed direction does not change.

[0059] The machining experiment can be specifically as follows: at any feed speed, perform a plurality of machining experiments with different feed direction angle change amounts and compare them with comparison experiments with unchanged feed direction angle; wherein the total length of the machining path in the plurality of machining experiments and the comparison experiments is the same, and the feed direction in any group of machining experiments with different feed direction angle change amounts can change at least once according to the same angle, so that the difference between the machining time measured in any group of machining experiments with different feed direction angle change amounts and the machining time measured in the comparison experiment, divided by the number of angle changes, is the time cost when the feed direction change amount in the experiment is the angle change amount in the experiment at the given feed speed.

[0060] According to the experimental data of time cost caused by different feed speeds under the same feed direction angle change amount, it can be known that when the feed direction change angle is constant, the time cost is approximately a linear function of the feed speed. According to the experimental data of time cost caused by different feed direction angle change amounts under the same feed speed, it can be known that when the feed speed is constant, the time cost of feed direction change is approximately a quadratic function of the angle of feed direction change. Therefore, the time cost f(v, ω) can be fitted using the following formula, wherein v represents the feed speed, and ω represents the angle of feed direction change.

[0061] f(v, ω) = p 00 + p 10 v + p 01 ω + p 11 vω + p 02 ω 2 ;

[0062] According to the fitting results of the experimental data, the coefficients in the formula are:

[0063] p 00 = -3.684 (-6.326, -1.041);

[0064] p 10 = 0.002091 (0.0008873, 0.003295);

[0065] p 01 = -0.08314 (-0.1272, -0.03907);

[0066] p 11 = 9.615e -5 (8.48e -5 , 0.0001075);

[0067] p 02 = 0.00063 (0.0004209, 0.0008391);

[0068] Table 1 Experimental data of time cost (ms) under different feed speeds (mm / min) and feed direction change angles (°)

[0069]

[0070]

[0071] The reward function R1(s, a) of the feed direction change gives the agent, i.e., the tool, a certain punishment (negative reward value) according to the angle of the feed direction change, limits the number and angle of the feed direction change of the tool, and is defined as shown in the following formula:

[0072] R1(s, a) = -0.1 * f(v, ω);

[0073] Assuming that the feed speed of the cutter is constant during milling, 2000 mm / min, R1(s, a) can be defined as follows:

[0074]

[0075] where Angle(s, a) is the angle between the current feed direction and the next feed direction, and the value range is [0°, 180°].

[0076] The reward function R2(s, a) of the lifting and feeding action gives the cutter a negative reward value according to the lifting and feeding, so as to limit the number of times of lifting and feeding of the cutter, that is, a larger negative reward value is given to the lifting and feeding of the cutter, which is defined as shown in the following formula:

[0077]

[0078] The reward function R3(s, a) of the cutter movement is: when the cutter performs effective cutting, a positive reward value is given to the cutter to guide the cutter to continuously perform effective cutting; when the cutter movement produces idle stroke, a negative reward value is given to the cutter to reduce the idle stroke of the cutter. It is defined as shown in the following formula:

[0079]

[0080] The final immediate reward function of the agent is composed of the above three types of reward functions, which is defined as shown in the following formula:

[0081] R(s, a) = R1(s, a) + R2(s, a) + R3(s, a).

[0082] (4) State transition probability and termination state. Since this embodiment does not consider the influence of machine tool error, the state transition probability is constant, that is, given a certain state and action, the immediate reward, the next state and whether the environment ends the result is determined. The path completion reward value can also be set, when the agent traverses all the cutter path points, that is, when all the element values in the feature matrix M are less than or equal to 0, it means that the milling task of the given part is completed, and the interaction process should be ended, and a larger reward value such as 10 is given to the agent. The maximum number of decisions of the agent can also be set to avoid meaningless exploration of the agent. For example, the maximum number of decisions can be twice the number of all cutter path points of the environment. When the number of decisions of the agent is greater than twice the number of all cutter path points of the environment, if the agent has not traversed all the cutter path points at this time, the exploration of this round should be ended in advance to prevent the agent from falling into a dead loop in the process of interacting with the environment.

[0083] 3. In this embodiment, S32 specifically includes: using a multi-layer value simulation algorithm to solve the MDP model of the reinforcement learning task, obtaining the decision sequence of the intelligent agent, and thus obtaining the processing guide line, such as Figure 5 As shown in the figure. Due to the immediate reward function, the agent prioritizes actions with fewer turns, less redundant cutting, and fewer tool lifts. Furthermore, when there are no tool path points to be cut among the agent's eight adjacent tool path points, the agent will perform a tool lift and feed action, searching for an uncut tool path point as the new tool position. These two points ensure that, in the strategy determined by the immediate reward, the agent performs effective cutting in most cases, generating a path that traverses all tool path points.

[0084] Specifically, when the tool is in any state, the state value of any state is calculated according to the sum of the instantaneous reward values ​​of all actions on the tool path formed from any state as the starting point; t At the beginning, the strategy of determining the action with the immediate reward value is simulated, and the state value V(s t )for:

[0085]

[0086] Among them, V(s t ) represents state s t The state value of Indicates state s t The tool path based on which the state value is calculated is determined by simulating the tool path using the immediate reward value, i.e., the strategy of determining the action using the immediate reward value is simulated. t is the instantaneous reward value obtained by the agent interacting with the environment at time t, γ is the reward discount factor, and π i Represents a strategy for determining actions based on immediate reward values:

[0087]

[0088] Among them, R(s t ,a t ) represents state s t Next, perform action a t The instant reward value that can be obtained.

[0089] However, this path only determines the action strategy (π i ) is not an excellent tool path, and its machining efficiency may be relatively low, but the strategy of determining the action based on the immediate reward value (π i ) The value of the state generated by the simulation can reflect the optimal strategy π to a certain extent * The status value.

[0090] In order to obtain the optimal strategy π * The state value is determined using a multi-layer value simulation algorithm. The action value of any action performed by the tool in the current state is defined as the sum of the immediate reward value of that action and the state value of the next sub-state corresponding to the action performed from the current state. The first sub-state of the current state is the set of states after performing one possible action from the current state. In this algorithm, when determining the action for a state, a greedy strategy is used to select the action with the highest action value in the next sub-state:

[0091]

[0092] Among them, π g represents a strategy of selecting the action with the largest action value according to the greedy algorithm, Q(s t ,a t ) represents state s t Next, perform action a t Calculate the action value of each possible action. The action value of each possible action is calculated. The possible actions are all actions in the possible action space of the current state. Then, use a greedy algorithm to select the action with the highest action value. For example, if action 5 is a possible action in the current state, and the state after action 5 is a state in the first sub-state of the current state, the state value of the state after action 5 needs to be calculated. Then, calculate the immediate reward value of action 5 to obtain the action value of action 5. The current state selects the next action based on the action value of each possible action.

[0093] There are two ways to calculate the state value V(s) of the first layer sub-state t+1 ), the first is the strategy to determine the action based on the immediate reward value (π i ) is simulated, that is, starting from the first layer sub-state, the action of each step is selected with the strategy of the action with the largest immediate reward value to obtain the simulation path; then the sum of the immediate reward values ​​of all actions is calculated based on the simulated path. The state value judgment strategy of this sub-state is the strategy of determining the action according to the immediate reward value (π i ). Starting from the first layer of sub-states, the action strategy is determined according to the immediate reward value (π i ) to simulate and guide the agent to complete the environmental task, thereby obtaining the state value of the first-level sub-state, as shown in the following formula:

[0094]

[0095] The state value of the first layer sub-state V(s t+1 ) Determine the action strategy based on the immediate reward (π i ) is used for simulation, which is the multi-layer value simulation algorithm when m=1.

[0096] The second scheme is to select the action with the maximum action value according to the greedy strategy (π g ) again, guide the agent to perform simulation, until the agent completes the environment task, so as to obtain the state value of the first layer sub-state, as shown in the following formula:

[0097]

[0098] In the second scheme, the state value V(s t+2 ) of the next layer sub-state of any state in the path simulation starting from the first layer sub-state needs to be calculated. The next layer sub-state is the state corresponding to the execution of one action at any state. Similar to the first layer sub-state, the state value of the next layer sub-state of any state in the path simulation of the first layer sub-state can also be calculated using two different strategies. The first strategy is to use the immediate reward value to determine the action strategy (π i ), that is, to use the action selection strategy (π i ) with the maximum immediate reward value to determine the action to perform simulation path to calculate the state value of the next layer sub-state of any state in the path simulation of the first layer sub-state, and then perform simulation according to the state value of the next layer sub-state, so as to obtain the state value of the first layer sub-state. The state value of the next layer sub-state of any state in the first layer sub-state path simulation is determined from the sub-state, and the path is simulated and determined using the immediate reward value, and then the state value of the next layer sub-state is obtained based on the simulated path, that is, the state value determination strategy of the next layer sub-state of any state in the first layer sub-state path simulation is the action selection strategy (π i ) according to the immediate reward value. This scheme is a multi-layer value simulation algorithm when m=2.

[0099] The second strategy is to use the greedy strategy (π g ) again to select the action with the maximum action value, but this requires calculating the state value of the next layer sub-state of any state in the path simulation of the next layer sub-state of any state in the path simulation of the first layer sub-state.

[0100] The above is the main process of the multi-layer value simulation algorithm. Multi-layer means that from a certain layer sub-state, the action selection strategy (π i ) according to the immediate reward value is used to simulate the path for value calculation. That is, the m-layer value simulation algorithm is:

[0101] When m=0, the action with the maximum immediate reward value is selected from the starting state to form the tool path, that is, the action with the maximum immediate reward value is directly executed in the current state;

[0102] When m=1, the action with the maximum action value in the current state is selected from the starting state to form the tool path, wherein the action value of the action in the current state is calculated based on the state value of the first layer sub-state of the current state, and the tool path used for calculating the state value of the first layer sub-state of the current state is the tool path determined by using the immediate reward value simulation.

[0103] When m=n (n>1), the action with the maximum action value in the current state is selected from the starting state to form the tool path, wherein the action value of the action in the current state is calculated based on the state value of the first layer sub-state of the current state, the tool path used for calculating the state value of the first layer sub-state of the current state is the tool path determined by using the n-1 layer value simulation algorithm, any action in the tool path corresponding to the state value of the first layer sub-state of the current state is selected based on the action value, the tool path used for calculating the state value of the next layer sub-state based on which the action is selected is the tool path determined by using the n-2 layer value simulation algorithm, and the recursion is performed in sequence until the tool path used for calculating the state value of the sub-state is the tool path determined by using the 0 layer value simulation algorithm, i.e., the tool path determined by using the immediate reward value simulation.

[0104] Specifically, m=2, at this time, the two-layer value simulation algorithm is specifically:

[0105] The action with the maximum action value in the current state is selected from the starting state to form the tool path, wherein the tool path used for calculating the state value of the first layer sub-state of the current state is the tool path determined by using the 1 layer value simulation algorithm, any action in the tool path corresponding to the state value of the first layer sub-state of the current state is selected based on the action value, and the tool path used for calculating the state value of the next layer sub-state based on which the action is selected is the tool path determined by using the 0 layer value simulation algorithm, i.e., the tool path determined by using the immediate reward value simulation.

[0106] Optionally, considering the calculation accuracy and calculation efficiency, m=2 is set, i.e., the two-layer value simulation algorithm is used to solve the MDP model, the decision sequence of the agent is obtained, and thus the machining guide line is generated. The pseudo code of the multi-layer value simulation algorithm with m=2 is shown in the following table. In order to obtain accurate state values during simulation, attention must be paid to future reward values, and therefore the reward discount factor γ=1.0 is set.

[0107]

[0108]

[0109] The milling tool path generation method based on the machining guide line offset provided in the embodiment, by combining with the reinforcement learning, considers the influencing factors of the machining efficiency in the machining process when generating the machining guide line, the tool can select a reasonable movement direction based on the current milling workpiece state and the tool movement state, and therefore the generated machining guide line has less redundant path. The tool path generated based on the machining guide line offset has less redundant cutting and higher machining efficiency. The multi-layer value simulation method is used to calculate the machining guide line of the workpiece to be machined. When the multi-layer value simulation method is used, a small-scale grid model, i.e. a simplified grid model, is used, and the number of grids is small, and therefore the calculation amount is small and the calculation efficiency is high.

[0110] In other embodiments, other methods can also be used to solve the tool path point access sequence problem of the machining area to generate the machining guide line, for example, the genetic algorithm, the ant colony algorithm, etc. The genetic algorithm, the ant colony algorithm, etc. are commonly used to solve the access sequence problem of the discrete path points, and the specific calculation process is not described herein.

[0111] 4. In the embodiment, S4 specifically includes: mapping the machining guide line to the initial grid model, and performing at least one of offsetting, lengthening, cutting and connecting processing on the machining guide line, so that the offset line and the machining guide line cover all the grids of the machining area corresponding to the initial grid model. That is, the machining guide line is placed in the initial grid model according to the path point coordinates, the center point of the grid contained in the machining area in the initial grid model is the tool path point corresponding to the initial grid model, and the machining guide line and the offset line can finally pass through all the tool path points corresponding to the initial grid model, i.e. cover all the grids of the machining area corresponding to the initial grid model, by setting the offset distance of the machining guide line and the processing times of offsetting, etc. At the same time, due to the setting of the grid length of the initial grid model, the complete and effective milling of the machining area can be completed when the tool passes through all the tool path points corresponding to the initial grid model.

[0112] Optionally, the grid length of the simplified grid model is N times of the tool cutting width, and N is an integer greater than or equal to 2. Therefore, by performing at least one of offsetting, lengthening, cutting and connecting processing on the machining guide line, the machining guide line and the offset line can cover all the grids of the machining area corresponding to the initial grid model, and the complete and effective milling of the machining area can be completed. At least one of offsetting, lengthening, cutting and connecting processing is performed on the machining guide line, i.e. offsetting, lengthening, cutting and connecting processing is first performed on the machining guide line to obtain the offset line corresponding to the machining guide line, and then offsetting, lengthening, cutting and connecting processing is performed again on the offset line corresponding to the machining guide line to obtain the offset line again, and so on.

[0113] Further, the acquiring the offset line in S4 specifically comprises: offsetting the machining guide line to both sides by using the segmented offset manner to form offset line segments on both sides; connecting the offset line segments on any one side at the disconnected parts, cutting the offset line segments on any one side at the intersecting parts, and extending the offset line segments on any one side at the end points, and then connecting the end points of the offset line segments on both sides to obtain the continuous offset line.

[0114] In the example of simplifying the grid length of the grid model to be twice the tool cutting width, i.e., combining four adjacent grids in the initial grid model as a group to form a simplified grid model, the generated machining guide line is mapped to the initial grid model, and the machining guide line is offset to both sides by using the segmented offset manner to remove the material on both sides of the machining guide line. As shown in Figure 6 Figure 6 The first figure in the figure is the machining guide line mapped to the initial grid model, and the second figure is the offset line segment formed by offsetting the machining guide line to both sides. The offset machining guide line at the corner of the machining guide line will appear the situation of the curves being apart from each other and intersecting with each other, so the curves apart from each other in the offset line segment are connected in the form of a polyline segment, and the intersecting curves are cut to obtain the smooth and connected offset curves, as shown in the third figure in Figure 6 Figure 7 As shown in the figure, the offset line generated by offsetting the machining guide line will exist the grid that is not cut at the start point and the end point of the machining guide line. Therefore, the start point and the end point of the generated offset line are extended to ensure that all machining grids are cut by the tool. At the same time, the end point of the first offset line is connected with the end point of the second offset line, and the generated two offset lines are connected into one tool path. The two paths generated by offsetting the same continuous tool path can be connected because the two paths generated by offsetting the same continuous tool path are very close to each other. After being connected, it is equivalent to starting from the start point of one offset line and returning to the start point of another offset line.

[0115] Figure 6 and Figure 7 In the example shown in the figure, the grid length of the simplified grid model can be an even multiple of the tool cutting width. When the grid length of the simplified grid model is an even multiple of the tool cutting width, the machining guide line does not pass through the grid center point of the initial grid model because the grid center point of the simplified grid model corresponds to the grid edge of the initial grid model. Therefore, the machining guide line can be offset at least once to make the offset line pass through all tool path points of the initial grid model, and the machining guide line is deleted. The offset line passing through the grid center point of the initial grid model is used as the moving line in the tool path.

[0116] ​​In other embodiments, the grid length of the simplified grid model can also be an odd multiple of the tool cutting width. When the grid length of the simplified grid model is an odd multiple of the tool cutting width, the grid center points of the simplified grid model coincide with the grid center points of the initial grid model. Therefore, the machining guide line can be offset so that the offset line and the machining guide line jointly pass through all tool path points of the initial grid model. At this time, the machining guide line and the offset line passing through the grid center points of the initial grid model can be simultaneously used as the movement line in the tool path.

[0117] It can be seen that, regardless of whether the grid length of the simplified grid model is an even multiple or an odd multiple of the cutting width, the offset line and / or the machining guide line can cover all grid center points of the corresponding to-be-machined region of the initial grid model by adjusting the offset distance and the number of offset processing times of the machining guide line. That is, a complete and effective tool path can be generated based on the machining guide line by the offset method. Therefore, the embodiment does not limit the specific setting of the grid length of the simplified grid model.

[0118] Further, in the environment of reinforcement learning, the non-machinable region is characterized as a discrete grid, and the discrete grid cannot accurately represent the boundary of the workpiece. Therefore, an additional tool path is generated using the boundary line of the workpiece to machine the boundary region of the workpiece. That is, the tool path generated in S4 further includes: using the contour line as the movement line in the tool path of the to-be-machined part to generate the tool path. This enables the tool to machine the boundary of the to-be-machined part along the contour line. The contour line can include an inner contour and / or an outer contour, and the specific setting is not limited according to the characteristics of the to-be-machined part.

[0119] 5, In the embodiment, S4 further includes: based on the movement line in the tool path of the to-be-machined part, establishing a set of tool feed points on the movement lines; and solving an optimal traversal order of the set of tool feed points as the machining order between the movement lines to obtain the final tool path, so that the total movement distance of traversing the set of tool feed points is the shortest. For example, the optimal traversal order of the tool can be solved by a dynamic programming algorithm to generate a connection line between the tool feed points. The purpose of connecting the tool path is to determine the optimal traversal order of the tool so that the non-cutting movement distance of the tool between the tool feed points is minimized.

[0120] Further, to simplify the problem of solving the optimal traversal order of the tool between different movement lines, the grid length of the simplified grid model can be set to be 2 Mtimes the machining guide line, and a plurality of lines in the inner and outer contour lines of the part to be machined. These tool movement lines are not directly connected, so the tool path connecting line is needed to connect the lifting point and the feed point of the non-connected tool movement line to form a complete milling tool path. The inner and outer contour lines of the target part are closed curves, so their feed points and lifting points coincide. The lifting points and feed points of the offset line generated by the offset of the machining guide line do not coincide, but they are close to each other, such as Figure 8 (a) in the middle, so they are regarded as the same point.

[0121] On this basis, the tool feed point set on the plurality of movement lines is established, and the feed point set is specifically composed of a plurality of starting points corresponding to the plurality of movement lines. The tool path connection problem can be described as: determining the traversal order of the tool in a series of discrete feed points (node1, node2, …, node n ) and making the non-cutting movement distance of the tool between these feed points shortest. As shown in the example of (b) in Figure 8 , there are four feed points: node1, node2, node3 and node4. The distance between these points can be obtained by calculating the Euclidean distance. Suppose there is a starting point S, and the distance between it and all feed points is 0. The purpose of tool path connection is to determine the optimal traversal order of the tool to minimize the non-cutting movement distance of the tool between the feed points.

[0122] This problem can actually be transformed into finding the shortest loop that traverses each feed point once and returns to the starting point S, which is a typical TSP problem. The reward function designed when solving the machining guide line limits the lifting and feeding behavior of the tool, reducing the number of lifting and feeding of the tool. Therefore, the number of nodes of this TSP problem is small, and the dynamic programming algorithm can be used to obtain the optimal traversal order of the tool. Finally, the tool path is composed of three parts: the movement line generated by the offset of the machining guide line, the inner and outer contour lines of the part, and the tool path connection line. The obtained tool path and its connection line are converted into machining G code to generate a numerical control machining program, as shown in Figure 9 .

[0123] 6, The beneficial effects of this embodiment: To verify the effectiveness of the milling tool path generation method based on the offset of the machining guide line proposed in this embodiment, use Figure 10The simulation experiment is carried out on the cavity workpiece shown in Table 2. First, the part is layered by layer milling, and each layer is converted into a grid model; then the grid model of each layer is used to generate a machining guide line, and the generated machining guide line is offset to generate a tool path; finally, the tool path generated by each layer is post-processed to obtain the numerical control machining program of the part.

[0124] The tool path generated by the method proposed in this embodiment is compared with the tool path generated by three common cutting modes (reciprocal, follow component, and follow periphery) in NX to prove the effectiveness and advancement of the method proposed in this embodiment. The tool paths generated by the four methods are simulated in the cutting simulation software CIMCO Edit 8.02.23, and the cutting results are shown in Table 3. Figure 11 Figure 11 In Table 3, (a) represents the complex cavity milling tool path generation method based on machining guide line offset, (b) represents the NX reciprocal cutting mode, (c) represents the NX follow component cutting mode, and (d) represents the NX follow periphery cutting mode. It can be observed that the tool paths generated by the four methods can complete the milling of the target part. The machining time, cutting time, trajectory length, and cutting distance of the tool paths generated by different methods are counted using the cutting simulation software CIMCO Edit 8.02.23, and the counting results are shown in Table 3. The improvement ratio is calculated based on the machining time. It can be seen that under the condition that the part model, machining allowance, feed speed, cutting width, tool diameter, and other cutting conditions are consistent, the cutting distance and trajectory length of the tool path generated by the complex cavity milling tool path generation method based on machining guide line offset proposed in this embodiment are shorter, the redundant cutting is less, and the machining efficiency is higher.

[0125] Table 2 Simulation experiment parameters

[0126]

[0127] Table 3 Simulation experiment results

[0128]

[0129] ​Furthermore, the second embodiment of the present invention provides a milling tool path generation system based on machining guide line offset, which is used to implement any of the methods described above. The system can be understood by reference to the above methods, and the system includes: a grid establishment module, which is used to obtain a simplified grid model of the part to be machined, and the grid side length of the simplified grid model is greater than the cutting width of the tool; a path point determination module, which is used to determine the area to be machined on the simplified grid model, and use the grid center point of the simplified grid model as the tool path point to convert the tool path generation problem of the area to be machined into a tool path point access sequence problem; a guide line generation module, which is used to solve the tool path point access sequence problem of the area to be machined, obtain the optimal access sequence of the tool path points of the area to be machined, and generate machining guide lines; a path generation module, which is used to offset, extend, cut and connect the machining guide lines, obtain offset lines, and use the offset lines as the moving lines in the tool path of the part to be machined to generate a tool path.

[0130] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for generating a milling tool path based on machining guide line offset, characterized in that: include: S1, obtaining a simplified mesh model of a part to be processed, wherein the mesh side length of the simplified mesh model is greater than the cutting width of the tool; S2, determining a region to be processed on the simplified grid model, using the grid center point of the simplified grid model as a tool path point, and converting the tool path generation problem of the region to be processed into a tool path point access sequence problem; S3, solving the problem of access sequence of tool path points in the area to be processed, obtaining the optimal access sequence of tool path points in the area to be processed, and generating a processing guide line; S4, offsetting, extending, cutting, and connecting the machining guide lines to obtain an offset line, and using the offset line as a moving line in a tool path of the part to be machined to generate a tool path; S3 specifically includes: S31, converting the tool path point access sequence problem of the area to be processed into a reinforcement learning task, establishing a Markov decision process model (MDP model) of the reinforcement learning task, wherein the intelligent agent in the MDP model is the tool, and the environment is the state of the grid center point in the simplified grid model. The state space of the tool in the environment and the corresponding action space are defined, and an immediate reward function for the tool action is set. After the tool performs an action in the action space corresponding to the state space, the environment and state space change and a corresponding immediate reward value can be obtained according to the immediate reward function. S32, solving the MDP model based on the instant reward value of the tool action to generate the processing guide line; Among them, S31 defines the state space of the tool in the environment S for: ； Where, d Indicates the current movement direction of the tool; s 0~7 Represents the environmental status information in 8 adjacent directions around the discrete tool path point where the tool is located; x and y Respectively represent the abscissa and ordinate of the discrete tool path point where the tool is located; Tool motion space A for: ； Where 0 to 7 represent the movement from the current tool path point to the adjacent tool path points in 8 directions, and 8 represents the tool lifting and feeding action.

2. The method for generating a milling tool path based on machining guide line offset according to claim 1, wherein: S1 specifically includes: Dividing the part to be processed into several layers to be processed by layered milling; Obtaining the simplified mesh model of any layer to be processed; Accordingly, S2 specifically includes: Determine a contour line of any of the layers to be processed according to a preset machining allowance and a tool radius, and determine a to-be-processed area of ​​any of the layers to be processed based on the contour line; The center point of the grid of the simplified grid model contained in the area to be processed is used as the tool path point, and the tool path generation problem of any layer to be processed is converted into a tool path point access sequence problem on the grid plane.

3. The method for generating a milling tool path based on machining guide line offset according to claim 1, wherein: The immediate reward function of the tool action is specifically the sum of the reward function for changing the feed direction, the reward function for lifting the tool and the reward function for moving the tool. Among them, the reward function for changing the feed direction gives the tool a negative reward value according to the angle of the feed direction change, so as to limit the number and angle of the tool feed direction change; the reward function for lifting the tool and feeding the tool gives the tool a negative reward value according to the tool lifting and feeding, so as to limit the number of times the tool is lifted and fed; the reward function for tool movement is: when the tool performs effective cutting, the tool is given a positive reward value, and when the tool movement produces an empty stroke, the tool is given a negative reward value to reduce the tool's empty stroke.

4. The method for generating a milling tool path based on machining guide line offset according to claim 1, wherein: S32 specifically includes: solving the MDP model using a multi-layer value simulation algorithm to obtain a decision sequence of the intelligent agent, thereby generating the processing guide line; Specifically, when the tool is in any state, the state value of any state is calculated based on the sum of the immediate reward values ​​of all actions on the tool path formed from any state as the starting point; the action value of the tool performing any action in the current state is the sum of the immediate reward value of any action and the state value of the next sub-state corresponding to the execution of any action from the current state; the first sub-state of the current state is the state set after executing one feasible action from the current state; m The layer value simulation algorithm is: when When , the actions with the largest immediate reward value are selected from the starting state to form the tool path; when When , the action with the largest action value in the current state is selected in sequence from the starting state to form a tool path, wherein the action value of the action in the current state is calculated based on the state value of the first-level sub-state of the current state, and the tool path based on which the state value of the first-level sub-state of the current state is calculated is the tool path determined by simulation using the immediate reward value; when When the tool path is formed, the action with the largest action value in the current state is selected from the starting state in sequence. The action value of the action in the current state is calculated based on the state value of the first layer sub-state of the current state. The tool path based on the state value calculation of the first layer sub-state of the current state is used n -1 layer value simulation algorithm solves the tool path determined by the state value of the first layer sub-state of the current state. Any step in the tool path corresponding to the state value of the current state is selected based on the action value. The tool path based on which the state value of the next layer sub-state based on which any step is selected is used. n -The tool path determined by the 2-layer value simulation algorithm is recursively solved until the state value of the sub-state is calculated based on the tool path determined by the 0-layer value simulation algorithm, that is, the tool path determined by the immediate reward value simulation.

5. The method for generating a milling tool path based on machining guide line offset according to claim 1, wherein: S1 specifically includes: Obtaining an initial mesh model of the part to be processed, wherein the mesh side length of the initial mesh model is equal to the cutting width of the tool; Merging the meshes in the initial mesh model into a group to obtain the simplified mesh model; Accordingly, S4 specifically includes: The processing guide line is mapped to the initial grid model, and the processing guide line is offset, extended, cut and connected at least once, so that the offset line and the processing guide line cover all grids of the to-be-processed area corresponding to the initial grid model.

6. The method for generating a milling tool path based on machining guide line offset according to claim 1, wherein: Obtaining the bias circuit in S4 specifically includes: The processing guide line is equidistantly offset to both sides by using a segmented offset method, so as to form offset line segments on both sides respectively; The bias line segments on either side are connected at the disconnected parts, cut at the intersecting parts, and extended at the end parts, and then the end points of the bias line segments on both sides are connected to obtain the continuous bias line.

7. The method for generating a milling tool path based on machining guide line offset according to claim 2, wherein: Generating toolpaths in S4 also includes: The contour line is also used as a moving line in the tool path of the part to be processed to generate a tool path.

8. The method for generating a milling tool path based on machining guide line offset according to claim 7, wherein: Generating toolpaths in S4 also includes: Based on the movement lines in the tool path of the part to be processed, establishing a set of entry points of the tool on multiple movement lines; With the goal of minimizing the total moving distance of the feed point set, the optimal traversal sequence of the feed point set is solved as the processing sequence of the tool between multiple moving routes, and then the final tool path is obtained.

9. A milling tool path generation system based on machining guide line offset, characterized in that: The method for generating a milling tool path based on machining guide line offset for implementing any one of claims 1 to 8 comprises: A mesh building module is used to obtain a simplified mesh model of the part to be processed, wherein the mesh side length of the simplified mesh model is greater than the cutting width of the tool; a path point determination module, configured to determine a region to be processed on the simplified grid model, use a grid center point of the simplified grid model as a tool path point, and transform a tool path generation problem for the region to be processed into a tool path point access sequence problem; A guide line generation module is used to solve the problem of the access sequence of the tool path points in the area to be processed, obtain the optimal access sequence of the tool path points in the area to be processed, and generate a processing guide line; The path generation module is used to offset, extend, cut and connect the processing guide line, obtain the offset line, and use the offset line as the moving line in the tool path of the part to be processed to generate the tool path.

Images