Robot mirror image processing track planning method for double-robot shot blasting equipment

Abstract

The invention relates to the technical field of robot collaborative processing, in particular to a robot mirror image processing track planning method for double-robot shot blasting equipment, and the method comprises the following steps: discretizing and sampling the surface of a workpiece to form track points, and calculating an effective topological energy receiving rate; constructing a shot peening jet volume energy field model comprising a kinetic energy flux density function and a jet pneumatic interference boundary model; calculating the reference energy flux of the master robot by using the kinetic energy flux density function and the effective topological energy receiving rate, and establishing an energy flux topological balance equation for constraining the effective energy flux of the slave robot to be equal to the reference energy flux; and the target state of the slave robot is solved based on the equation, if aerodynamic interference is detected based on the jet aerodynamic interference boundary model, time domain sliding mode avoidance is executed, and if kinematics singularity is detected, cross-parameter substitution compensation is executed. According to the method, equivalent strengthening machining on the asymmetric curved surface is achieved through physical-level energy topology balance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot collaborative processing. More particularly, the present application relates to a robot mirror image processing trajectory planning method of a double-robot shot blasting device. BACKGROUND

[0002] In the field of aerospace and high-end equipment manufacturing, for the surface strengthening treatment of large and complex curved thin-walled parts such as wing panels and integral beams, double-robot mirror image shot blasting has become a key process. The core logic of this process is to use two robots to synchronously follow and spray at the spatial mirror image positions on both sides of the workpiece, and to use the impact forces generated by the two-sided jets to cancel each other out inside the workpiece, thereby preventing macroscopic bending or distortion of the thin-walled part due to unilateral force. The current mainstream trajectory planning method generally uses a geometric position mirror image strategy, that is, according to the CAD model of the workpiece, the tool coordinate system of the master robot is directly transformed by a mirror matrix to generate the target trajectory of the slave robot, trying to achieve mechanical balance through geometric symmetry.

[0003] However, in actual engineering applications, the existing planning method based on pure geometric mirror image has serious limitations when dealing with variable-curvature heterogeneous components. First of all, geometric symmetry does not equal energy symmetry. Actual large skin-type workpieces often have asymmetric curvature characteristics, such as the leading edge of an aircraft wing, which is convex on one side and may be a gently concave surface on the other side. According to the principles of gas jet dynamics, the convex structure will cause the jet beam to diverge at the moment of impact, resulting in a decrease in effective kinetic energy flux per unit area; while the concave structure will have a focusing effect on the jet beam, resulting in an increase in local energy density. If only geometric position symmetry is maintained, there will be a significant difference in the actual surface strengthening energy received by the workpiece on both sides, and this microscopic energy imbalance will lead to an asymmetric residual stress field inside the workpiece after accumulation, ultimately causing the workpiece to deform after the tool is removed, which cannot meet the high-precision forming requirements.

[0004] Secondly, the existing planning algorithm lacks dynamic perception ability for aerodynamic flow field interference. Shot blasting jet is not an ideal ray, but a conical flow field accompanied by high-pressure turbulent entrainment effect. When the double robot processes the edge, hole or narrow cavity area of the workpiece, although the robot's mechanical arm does not collide, the high-speed gas flow fields on both sides may intersect and interfere in space. This aerodynamic interference will destroy the stability of the jet, resulting in substandard shot blasting coverage or uncontrollable turbulence, and the planning method based on geometric coordinates cannot identify the conflict of this flow field physical boundary.

[0005] In addition, the kinematic singularity point is a hard constraint for limiting the collaborative work of the dual robots. In the process of tracking a complex curved surface track, in order to forcibly satisfy the geometric mirror relationship, the robot is often forced to enter the joint limit position or the singular pose region. Under the prior art system, once encountering a singularity point, the machining can only be stopped or deviated from the established path to avoid, which directly leads to the discontinuity or missing of the machining, and seriously affects the surface integrity. SUMMARY

[0006] To solve the problems of energy receiving imbalance of heterogeneous curved surfaces, aerodynamic interference of flow field and kinematic singularity limitation caused by relying on geometric mirror in the prior art, the present application proposes a robot mirror machining track planning method of a dual-robot shot blasting device, which comprises the following steps: Discretely sampling the workpiece surface to form a series of track points, and calculating the effective topological energy receiving rate in combination with the curvature and angle characteristics of the track points; Constructing a shot blasting jet volume energy field model containing a kinetic energy flux density function and a jet aerodynamic interference boundary model; Calculating the reference energy flux of the master robot by using the kinetic energy flux density function and the effective topological energy receiving rate, and establishing an energy flux topological balance equation, which constrains the effective energy flux of the slave robot to be equal to the reference energy flux; Based on the energy flux topological balance equation, the target state of the slave robot is solved inversely. In the solving process, if it is detected that the regions overlap based on the jet aerodynamic interference boundary model, a time domain sliding mode avoidance strategy is activated to calculate the execution lag time, and if it is detected that the slave robot is in a kinematic singularity domain, a cross-parameter substitution compensation strategy is activated to calculate the compensation amount of the non-geometric process parameters; According to the target state of the slave robot, a control instruction stream is generated to drive the dual-robot collaborative work to realize the consistency of the effective impact kinetic energy flux on both sides of the workpiece.

[0007] The present application breaks the traditional geometric symmetry, i.e. stress symmetry, from the physical nature, directly solves the energy receiving imbalance problem caused by heterogeneous curved surfaces at the track planning level by introducing the reference energy flux as a unified dimension, and ensures the continuity and stability of the dual-robot work in a complex limited environment through the integrated interference avoidance and singularity compensation mechanism.

[0008] Preferably, the calculation relationship of the kinetic energy flux density function is:

[0009] In the formula, i is the index number of the track point, is the kinetic energy flux density value at the i-th track point, ​​Indicates the first The radial distance values ​​of each trajectory point from the central axis of the jet. Indicates corresponding to the first The effective divergence radius of the jet beam at the target distance of each trajectory point. This represents the energy transfer efficiency coefficient. This indicates the shot peening air pressure value. This represents the mass flow rate of the projectile. This indicates the numerical value of high-pressure air density.

[0010] This invention no longer simplifies the jet as a homogeneous ray, but models it as a volumetric energy field conforming to a Gaussian distribution. It accurately quantifies the nonlinear decay law of energy density with respect to target distance and radial distance, providing an aerodynamic model benchmark for subsequent accurate energy balance calculations.

[0011] Preferably, the formula for calculating the effective topological energy receiver rate is:

[0012] In the formula, Indicates the first Effective topological energy receiver rate of each trajectory point Indicates the first The jet impact angle value at each trajectory point Indicates the first The average principal curvature value of each trajectory point. Indicates the first The Gaussian curvature values ​​of each trajectory point Indicates the first The target distance value corresponding to each trajectory point.

[0013] This invention establishes a mapping relationship between the microscopic geometric features of a workpiece and its macroscopic physical energy, which can automatically identify the divergence loss caused by convex surfaces and the convergence gain caused by concave surfaces. It is the core calibration parameter for realizing physical-level mirror machining.

[0014] Preferably, the establishment of the energy flux topological balance equation specifically involves: The energy flux function is obtained by multiplying the kinetic energy flux density function by the effective topological energy receiving rate. The energy flux of the main robot is calculated using the energy flux function and recorded as the reference energy flux. The energy flux of the slave robot is set to be equal to the reference energy flux, thereby establishing an equation that includes the geometric position variables and process parameter variables of the slave robot as the energy flux topological balance equation.

[0015] The application converts the abstract mirror processing process requirements into mathematically reversible constraint equations, so that the system can maintain the dynamic balance of energy input by adjusting other variables in the case of geometric position cannot be strictly symmetrical.

[0016] Preferably, the jet aerodynamic interference boundary model defines an interference radius that expands with the increase of target distance; The detection based on the jet aerodynamic interference boundary model comprises: calculating the Euclidean distance between the target point of the master robot and the target point of the slave robot, and comparing the Euclidean distance with the sum of the interference radii of the master and slave robots determined based on the jet aerodynamic interference boundary model; if the Euclidean distance is less than the sum of the interference radii, it is determined that there is a risk of aerodynamic interference, and the time domain sliding mode avoidance strategy is triggered.

[0017] Preferably, the time domain sliding mode avoidance strategy specifically comprises: keeping the spatial path coordinates of the slave robot unchanged; calculating the minimum time difference required for the interference regions of the master and slave robots to be staggered on the time axis, and adding the minimum time difference to the time stamp of the slave robot as a hysteresis time to realize the processing mode of spatial overlap but time separation.

[0018] The application utilizes the time dimension to solve the spatial conflict, and realizes the physical isolation of the flow field through microsecond-level time displacement without changing the shot covering path.

[0019] Preferably, the cross-parameter substitution compensation strategy converts the geometric position error into the compensation amount of non-geometric process parameters by using an energy sensitivity matrix, and the calculation relationship of the energy sensitivity matrix is:

[0020] In the formula, is the initial set air pressure and flow value of the slave robot for the first trajectory point; is the jet beam effective divergence radius value corresponding to the safe boundary target distance ; is the energy sensitivity matrix of the first trajectory point.

[0021] Preferably, if it is detected that the slave robot is in the kinematic singular domain, the cross-parameter substitution compensation strategy is activated to calculate the compensation amount of non-geometric process parameters, which specifically comprises: calculating the condition number of the Jacobian matrix of the target pose of the slave robot, and determining whether the condition number exceeds a preset safety threshold; if yes, it is determined that the slave robot is in the kinematic singular domain and the geometric position of the slave robot is locked at the safe boundary; The energy loss is caused by the difference between the actual energy flux caused by the locking position and the benchmark energy flux; the energy loss is mapped to the shot peening mass flow rate increment and the shot peening gas pressure increment of the slave robot by using the pseudo-inverse matrix of the energy sensitivity matrix, and the shot peening mass flow rate increment and the shot peening gas pressure increment are called compensation amounts of non-geometric process parameters.

[0022] The application realizes the control logic that can compensate, when the robot cannot reach the ideal geometric position due to mechanical structure limitation, the energy loss caused by too long distance or angle deviation is automatically compensated by increasing the gas pressure or flow rate, and the working ability boundary of the robot is greatly expanded.

[0023] Preferably, the method further comprises performing surface integrity constraints, in particular: The surface roughness value after single impact is estimated based on the contact mechanics theory; If the surface roughness value exceeds the preset upper limit of the process, the single high-energy density scanning track is divided into multiple low-energy density reciprocating scanning tracks, and the scanning speed is re-planned under the premise of keeping the total cumulative energy flux unchanged.

[0024] Preferably, the method further comprises performing data preprocessing: Before the track planning, the original point cloud data of the workpiece surface is obtained by using a three-dimensional laser scanner; The original point cloud data is denoised and smoothed, and the normal vector, principal curvature and Gaussian curvature of each discrete point are calculated based on the processed point cloud data.

[0025] The application has the following beneficial effects: The application precisely describes the different response characteristics of heterogeneous surfaces to jet flow energy by constructing a kinetic energy flux density function and an effective topological energy receiving rate model. By solving the energy flux topological balance equation, the target distance or process parameters of the slave robot are dynamically adjusted according to the local curvature, and the effective strengthening energy received by both sides of the workpiece is strictly consistent, which fundamentally eliminates the distortion and deformation of large thin-walled parts caused by uneven stress input.

[0026] Further, for the problem that the robot is easy to fall into a singular posture in complex curved surface machining, the application proposes a cross-parameter substitution compensation strategy based on the energy sensitivity matrix. When the robot cannot reach the ideal geometric position due to mechanical structure limitation, the algorithm can automatically calculate the energy loss and map it to the compensation increment of the gas pressure or flow rate. This control logic of process compensation in insufficient space enables the dual-robot system to maintain high-quality working ability in the limited space close to the joint limit.

[0027] Further, by introducing the jet aerodynamic interference boundary model and combining the time domain sliding mode avoidance strategy, the invention can make the turbulent flow field of the double-machine jet deviate in time without changing the spatial processing path, through high-precision time axis sliding mode control. This not only avoids the fluctuation of shot intensity caused by flow field mutual interference, but also ensures the spatial continuity and integrity of the processing trajectory. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 is a step flow chart of a robot mirror image processing trajectory planning method of a double-robot shot blasting equipment provided by the embodiment of the invention; Figure 2 is a robot processing trajectory self-adaptive adjustment comparison chart provided by the embodiment of the invention. DETAILED DESCRIPTION

[0029] Please refer to Figure 1 which shows a step flow chart of a robot mirror image processing trajectory planning method of a double-robot shot blasting equipment provided by the embodiment 1, and the method comprises the following steps: S1: discretely sampling the workpiece surface to form a series of trajectory points, and calculating the effective topological energy receiving rate in combination with the curvature and angle characteristics of the trajectory points.

[0030] It should be noted that large-scale aviation thin-walled parts usually have complex free-form surface characteristics, such as sharp bending of the wing leading edge or gentle transition at the wing root. The difference in such geometric characteristics means that the reflection and absorption efficiency of the jet is completely different when it hits different positions. If this topological difference is ignored and only the geometric distance is consistent, it will lead to excessive dissipation of energy at the convex part, and excessive accumulation of energy at the concave part, which is extremely easy to cause the workpiece to be distorted after the fixture is removed. Therefore, the primary task of this step is to realize the digital feature extraction of the workpiece surface.

[0031] Preferably, as an example, discretely sampling the workpiece surface to form a series of trajectory points, and calculating the effective topological energy receiving rate in combination with the curvature and angle characteristics of the trajectory points, comprises: First, data acquisition and data preprocessing are performed, which is specifically implemented as follows: a high-precision three-dimensional laser scanner is used to scan the workpiece surface to obtain original point cloud data. Since there may be noise points in the original point cloud data due to environmental light reflection, the original point cloud data is filtered and denoised and Laplace smoothed. Based on the cleaned point cloud data after processing, the normal vector, principal curvature and Gaussian curvature of each data point are calculated by using differential geometry algorithm.

[0032] Then, the cleaned point cloud data after processing is discretely sampled to form a series of ordered trajectory points.

[0033] Finally, for each trajectory point, the effective topological energy receiving rate is calculated by the following relation:

[0034] wherein, represents the effective topological energy receiving rate of the i-th trajectory point; represents the jet impact angle value of the i-th trajectory point; represents the average principal curvature value of the i-th trajectory point; represents the Gaussian curvature value of the i-th trajectory point; represents the target distance value corresponding to the i-th trajectory point. It can be understood that when the workpiece surface is convex, is positive, and as the convexity increases, the value increases, and the value of the denominator term will be significantly greater than 1, and will increase accordingly, which will result in the calculated receiving rate decreasing, which means that the convex structure accelerates the spatial divergence of the jet energy, resulting in a decrease in the proportion of actual energy intercepted per unit area.

[0035] When the workpiece surface is concave, is negative, and as the concavity deepens, the absolute value increases, and the value of the denominator term will be less than 1, resulting in an increase in the receiving rate . This means that the concave structure has a focusing and converging effect on the jet beam, resulting in an increase in the local energy density.

[0036] When increases, i.e., the jet is more inclined, the molecular term decreases, resulting in a decrease in the receiving rate , which represents the loss of vertical component energy.

[0037] Through the above processing, the risk of geometric symmetry but energy asymmetry can be automatically described, providing a data basis for subsequent precise compensation. S2: Construct a shot jet volume energy field model containing kinetic flux density function and jet aerodynamic interference boundary model.

[0038] S2: Construct a shot jet volume energy field model containing kinetic flux density function and jet aerodynamic interference boundary model.

[0039] S2: Construct a shot jet volume energy field model containing kinetic flux density function and jet aerodynamic interference boundary model.

[0040] ​​It should be noted that in traditional dual-robot trajectory planning, the shot peening jet is often simplified into a geometric ray with no thickness, ignoring the volumetric effect and non-uniform energy distribution of the jet itself. This simplification means that the system cannot detect energy attenuation at the jet edge or predict aerodynamic interference between the two jets in space. If conventional methods are used, this can easily lead to insufficient reinforcement at the workpiece edge or turbulent collision between the two jets in narrow areas, causing processing instability. Therefore, to address these potential problems, a volumetric energy field model based on the physical flow field needs to be introduced.

[0041] Preferably, as an example, a shot peening jet volumetric energy field model is constructed, including the kinetic flux density function and the jet aerodynamic disturbance boundary model, comprising: Using shot peening process parameters as input, a kinetic energy flux density function is constructed using aerodynamic principles. This kinetic energy flux density function satisfies the following relationship:

[0042] In the formula, Indicates the first The kinetic energy flux density values ​​at each trajectory point; Indicates the energy transfer efficiency coefficient; This indicates the shot peening air pressure value of the robot. This represents the mass flow rate of the robot's projectiles. This indicates the numerical value of high-pressure air density; Indicates the corresponding to the first The effective divergence radius of the jet beam is the value of the target distance at each trajectory point. Indicates the first The radial distance of each trajectory point from the central axis of the jet is 0 during centering machining; Pi is a constant. It is a natural exponential function.

[0043] Understandably, when the target distance... As the volume increases, the jet beam widens, leading to Increase, thus increasing the denominator term It will increase exponentially, thus affecting the overall score. This significantly reduces the energy density, thus effectively describing the diffusion phenomenon where energy becomes increasingly scarce over greater distances.

[0044] when Increase, that is, when the sampling point deviates from the center of the jet, the numerator in the exponential term increases. To become a negative number with a larger absolute value, The function value drops sharply and approaches 0, thus effectively describing the Gaussian distribution characteristics where energy is mainly concentrated in the jet center and the edge energy decays rapidly.

[0045] Furthermore, in order to define the flow field disturbance, it is also necessary to define the jet aerodynamic disturbance boundary model, specifically: for the first... A trajectory point, defining its interference radius. satisfy: ,in is the turbulence expansion coefficient.

[0046] The volumetric energy field model of the shot peening jet is constructed by combining the volumetric energy field model of the shot peening jet and the kinetic energy flux density function.

[0047] S3: Calculate the reference energy flux of the master robot using the kinetic energy flux density function and the effective topological energy receiving rate, and establish an energy flux topological balance equation. The energy flux topological balance equation constrains the effective energy flux of the slave robot to be equal to the reference energy flux.

[0048] It's important to note that the core objective of dual-machine mirror machining is mechanical equilibrium. In heterogeneous surface scenarios, the master robot's operating state is fixed. However, if the slave robot simply replicates the master robot's geometric parameters, the different curvatures on both sides will lead to an imbalance in the actual input stress fields. To achieve physical equilibrium, it's essential to first determine how much effective energy the master robot is delivering.

[0049] Preferably, as an example, the reference energy flux of the main robot is calculated using the kinetic energy flux density function and the effective topological energy receiving rate, and an energy flux topological balance equation is established. This energy flux topological balance equation constrains the effective energy flux of the slave robot to be equal to the reference energy flux, including: First, the product of the effective topological energy receiving rate and the kinetic energy flux density function is used as the energy flux function; Because the master and slave robots operate synchronously, when the slave robot plans the... When the main robot is at the nth trajectory point, it is positioned on the surface of the workpiece. One main trajectory point.

[0050] Next, the process parameters of the main robot will be... Geometric features of the main trajectory points Substituting into the energy flux function, the baseline energy flux of the main robot at the i-th trajectory point is calculated as follows:

[0051] In the formula, The main robot in the The baseline energy flux value for each trajectory point. These are the preset shot peening air pressure data and preset shot peening mass flow rate values ​​of the main robot, respectively. The shot peening air pressure value and shot peening mass flow rate are set values ​​in the process controller. the target distance value of the main robot at the first trajectory point is obtained by using an existing robot off-line programming algorithm, taking the coordinates and normal vector of the point cloud data as the processing object, and performing geometric offset calculation along the normal direction according to a preset shot process standard operation distance; the effective jet beam divergence radius value of the target distance value of the main robot at the first trajectory point, the effective jet beam divergence radius value being obtained by using existing fluid dynamics simulation software to simulate the jet impact area diameter of a specific type of nozzle at different target distances, and performing calculation according to a jet diffusion linear regression equation fitted from simulation data; the jet impact angle value of the main robot at the first trajectory point, the jet impact angle value being determined by the angle between the axis obtained by robot inverse kinematics and the normal line; the average principal curvature and the Gaussian curvature values of the first trajectory point of the machining surface corresponding to the main robot, respectively. Subsequently, the energy flux topology balance equation is constructed from the effective energy flux of the slave robot being equal to the reference energy flux value, specifically:

[0052] In the formula, P is the shot gas pressure value of the slave robot for the first trajectory point to be solved, and m is the shot mass flow value of the slave robot for the first trajectory point to be solved;

[0053]

[0054] ​​​​​​​​​​​​​​​​​​​​​It can be understood that, assuming the slave robot is working on a deep concave area, its receiving rate will increase significantly according to the foregoing logic. At this time, if the slave robot adopts the same parameters as the master robot, the product on the left side of the equation will be much larger than the right side , resulting in over-spraying. In order to maintain the equation, the equation forces the original energy intensity term value of the slave robot to be reduced. The system will automatically derive the instructions to reduce the air pressure, shot mass flow or increase the target distance through reverse solving.

[0055] In summary, the method can automatically deduce the required compensation action according to the local shape difference of the workpiece, so as to ensure that the effective energy injected into the material inside is always constant regardless of the shape of the workpiece.

[0056] S4: Reverse solving the target state of the slave robot based on the energy flux topological balance equation, and performing detection based on the jet aerodynamic interference boundary model in the solving. If the region overlap is detected, the time domain sliding mode avoidance strategy is activated to calculate the hysteresis time. If the slave robot is detected in the kinematic singular domain, the cross-parameter substitution compensation strategy is activated to calculate the compensation amount of the non-geometric process parameters.

[0057] Preferably, as an example, reverse solving the target state of the slave robot based on the energy flux topological balance equation, and performing detection based on the jet aerodynamic interference boundary model in the solving. If the region overlap is detected, the time domain sliding mode avoidance strategy is activated to calculate the hysteresis time. If the slave robot is detected in the kinematic singular domain, the cross-parameter substitution compensation strategy is activated to calculate the compensation amount of the non-geometric process parameters, comprising: Solving the energy flux topological balance equation by using a numerical optimization algorithm such as Newton-Raphson iteration method. In the solving process for the first trajectory point, the system performs the following two detection and correction strategies in parallel: 1. Aerodynamic interference detection and time domain sliding mode avoidance strategy avoidance.

[0058] Calculate the Euclidean distance between the TCP position of the master robot corresponding to the first trajectory point and the TCP position of the slave robot corresponding to the first trajectory point.

[0059] At the same time, according to the jet aerodynamic interference boundary model, the interference radii of the master robot and the slave robot at the first trajectory point are calculated respectively and .

[0060] If , it is determined that there is a risk of aerodynamic interference at the first trajectory point.​

[0061] At this point, the system activates the time-domain sliding mode avoidance strategy and calculates the execution lag time. And add that time to the time from the robot's first... Timestamp of each trajectory point This enables a processing mode that overlaps spatially but separates temporally.

[0062] It is understandable that when the distance When the distance is less than a threshold, it indicates that the physical distance between the two robots is too close in space. A time lag is then introduced to address this. This allows for the elimination of processing interference through temporal misalignment when spaces overlap.

[0063] 2. Kinematic singularity detection and cross-parameter substitution compensation strategy: substitution compensation.

[0064] In solving the problem from the robot's perspective on the first... Target distance of each trajectory point If the calculated geometric pose causes the condition number of the Jacobian matrix of the robot to exceed a preset safety threshold, then the robot is determined to be in a kinematically singular domain. For example, the safety threshold is taken as... .

[0065] At this point, the robot's target distance will be forcibly locked at the safe boundary target distance. place, This represents the maximum non-singular distance that the robotic arm can reach in this posture.

[0066] because Forced to be fixed as This leads to the actual energy flux calculated by the equation. Less than the benchmark value This results in an energy deficit. .

[0067] To eliminate this loss, a cross-parameter substitution compensation strategy is initiated, using the energy sensitivity matrix to calculate the compensation amount from the robot's process parameters. The specific calculation method is as follows:

[0068] Among them, the The relationship between the energy sensitivity matrix of each trajectory point is:

[0069] In the formula, In order to target the Each trajectory point needs to be compensated for the incremental value of the shot peening air pressure from the robot; In order to target the Each trajectory point needs to be compensated for the incremental value of the shot peening mass flow rate from the robot; is the pseudo-inverse matrix of matrix , used to calculate the optimal compensation solution; is the energy loss value at the i-th trajectory point; is the initial set pressure and flow value of the robot for the i-th trajectory point; is the effective jet beam divergence radius value corresponding to the safety boundary target distance , and is the energy sensitivity matrix of the i-th trajectory point; the first row elements of the matrix represent the partial derivative of the kinetic energy flux density with respect to the pressure; the second row elements of the matrix represent the partial derivative of the kinetic energy flux density with respect to the flow. , are collectively referred to as compensation values of non-geometric process parameters. It can be understood that when the robot is forced to stop at a far safety position due to singularity constraints , the jet divergence radius becomes larger, resulting in smaller element values in matrix

[0070] , i.e., lower sensitivity, which means that the energy gain brought by unit pressure becomes weaker. When using pseudo-inverse for operation, a larger pressure increment will be automatically calculated. This means that the farther the distance, the lower the sensitivity, and the larger the pressure compensation of the system must be. This logic precisely ensures that even if the robot cannot reach the ideal position, it can still send enough energy to the workpiece surface by greatly increasing the pressure to achieve lossless equivalence of the energy field. S5: generating a control instruction stream according to the target state of the slave robot, and driving the dual-robot cooperative operation to realize consistency of effective impact kinetic energy flux on both sides of the workpiece. Preferably, as an example, generating a control instruction stream according to the target state of the slave robot, and driving the dual-robot cooperative operation to realize consistency of effective impact kinetic energy flux on both sides of the workpiece, includes:

[0071] First, integrity constraints are executed, specifically: Based on the Hertz contact theory, the surface roughness value after a single impact is estimated using the solved compensation process parameters

[0072] . If the surface roughness value exceeds the preset upper limit of the process, it means that the pressure applied to compensate for the energy is too large, which may damage the surface. At this time, a multi-pass disassembly strategy is triggered: the originally planned single-pass high-energy scan is disassembled into several reciprocating scanning trajectories with low energy density.

[0073]

[0074] ​​​It can be understood that the strategy ensures that the total injection energy remains unchanged by reducing the peak force of a single impact but increasing the number of scans.

[0075] Then, the control instruction stream subjected to integrity constraints is output to drive the dual robot to perform machining. At this time, the effective impact kinetic energy flux on both sides of the workpiece remains consistent throughout the entire machining process, thereby eliminating the deformation source.

[0076] Figure 2 For comparison of adaptive adjustment of robot machining trajectory, the gray wavy line in the figure indicates the actual surface cross-sectional profile of the workpiece to be machined, the red dashed line indicates the robot end motion trajectory generated by using the traditional geometric mirror method, and it can be seen that the trajectory only maintains a constant spacing in geometry and does not consider the influence of curvature on the energy field. The green solid line indicates the optimized trajectory generated by using the present application. The green arrows respectively indicate the active adjustment actions of the present application in different curvature regions.

[0077] As can be seen from the image, in the concave area on the left side of the workpiece, the green trajectory presents an active upward trend compared to the red trajectory, so as to reduce the surface energy density and prevent over-spraying.

[0078] In the convex area on the right side of the workpiece, the green trajectory presents an active downward trend compared to the red trajectory, so as to increase the surface energy density and compensate for the loss of divergence. The figure directly proves that the present application breaks the traditional geometric equidistance restriction and realizes the reconstruction of the trajectory based on the physical energy field.

[0079] Thus far, the present embodiment is completed.

[0080] The above only describes the preferred embodiments of the present application and should not be used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for planning the robot mirror processing trajectory of a dual-robot shot peening equipment, characterized in that, include: The workpiece surface is discretized and sampled to form a series of trajectory points. The effective topological energy receiving rate is calculated by combining the curvature and angular characteristics of the trajectory points. Construct a volumetric energy field model for a shot peening jet that includes the kinetic flux density function and the aerodynamic disturbance boundary model of the jet; The reference energy flux of the main robot is calculated using the kinetic energy flux density function and the effective topological energy receiving rate, and an energy flux topological balance equation is established. The energy flux topological balance equation constrains the effective energy flux of the slave robot to be equal to the reference energy flux. Based on the energy flux topological balance equation, the target state of the robot is solved in reverse. If regional overlap is detected based on the jet aerodynamic interference boundary model, the time-domain sliding mode avoidance strategy is activated to calculate the execution lag time. If the robot is detected to be in the kinematic singular domain, the cross-parameter substitution compensation strategy is activated to calculate the compensation amount of non-geometric process parameters. Based on the control command stream generated from the target state of the robot, the two robots work together to achieve consistency in the effective impact kinetic energy flux on both sides of the workpiece.

2. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 1, characterized in that, The formula for calculating the kinetic energy flux density function is as follows: In the formula, The index number of the trajectory point. Indicates the first The kinetic flux density values ​​at each trajectory point. Indicates the first The radial distance values ​​of each trajectory point from the central axis of the jet. Indicates corresponding to the first The effective divergence radius of the jet beam at the target distance of each trajectory point. This represents the energy transfer efficiency coefficient. This indicates the shot peening air pressure value. This represents the mass flow rate of the projectile. This indicates the numerical value of high-pressure air density.

3. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 1, characterized in that, The formula for calculating the effective topological energy receiving rate is as follows: In the formula, Indicates the first Effective topological energy receiver rate of each trajectory point Indicates the first The jet impact angle value at each trajectory point Indicates the first The average principal curvature value of each trajectory point. Indicates the first The Gaussian curvature values ​​of each trajectory point Indicates the first The target distance value corresponding to each trajectory point.

4. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 1, characterized in that, The establishment of the energy flux topological balance equation is specifically as follows: The energy flux function is obtained by multiplying the kinetic energy flux density function by the effective topological energy receiving rate. The energy flux of the main robot is calculated using the energy flux function and recorded as the reference energy flux. The energy flux of the slave robot is set to be equal to the reference energy flux, thereby establishing an equation that includes the geometric position variables and process parameter variables of the slave robot as the energy flux topological balance equation.

5. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 1, characterized in that, The jet aerodynamic interference boundary model defines an interference radius that expands with increasing target distance; The detection based on the jet aerodynamic interference boundary model includes: calculating the Euclidean distance between the target point of the master robot and the target point of the slave robot, and comparing the Euclidean distance with the sum of the interference radii of the master and slave robots determined based on the jet aerodynamic interference boundary model. If the distance is less than the sum of the interference radii, it is determined that there is aerodynamic interference risk, and the time-domain sliding mode avoidance strategy is triggered.

6. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 1, characterized in that, The time-domain sliding mode avoidance strategy is specifically as follows: Keep the spatial path coordinates of the robot unchanged; The minimum time difference required to offset the interference areas of the master and slave robots on the time axis is calculated, and this minimum time difference is superimposed on the timestamp of the slave robot as the execution lag time, thereby realizing a processing mode with spatial overlap but temporal separation.

7. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 2, characterized in that, The cross-parameter substitution compensation strategy utilizes the energy sensitivity matrix to transform geometric position errors into compensation amounts for non-geometric process parameters. The calculation formula for the energy sensitivity matrix is ​​as follows: In the formula, To use robots to target the first Initial air pressure and flow rate values ​​for each trajectory point; To correspond to the target distance of the safety boundary The effective divergence radius of the jet at that location; For the first The energy sensitivity matrix of each trajectory point.

8. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 7, characterized in that, If the robot is detected to be in a kinematically singular domain, a cross-parameter substitution compensation strategy is activated to calculate the compensation amount for non-geometric process parameters, specifically including: Calculate the condition number of the Jacobian matrix of the target pose of the robot and determine whether the condition number exceeds a preset safety threshold. If so, determine that the robot is in a kinematic singular domain and lock the geometric position of the robot at a safe boundary. Calculate the difference between the actual energy flux caused by the locked position and the reference energy flux, i.e., the energy deficit; use the pseudo-inverse matrix of the energy sensitivity matrix to map the energy deficit to the shot peening mass flow rate increment and shot peening pressure increment of the robot, the shot peening mass flow rate increment and shot peening pressure increment are called the compensation amount of non-geometric process parameters.

9. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 1, characterized in that, The method further includes performing surface integrity constraints, specifically: Predict the surface roughness value after a single impact based on contact mechanics theory; If the surface roughness value exceeds the preset process allowable upper limit, the single high energy density scan trajectory will be broken down into multiple low energy density reciprocating scan trajectories, and the scan speed will be replanned while keeping the total cumulative energy flux unchanged.

10. The robot mirror machining trajectory planning method for a dual-robot shot peening equipment according to claim 1, characterized in that, The method further includes performing data preprocessing: Before trajectory planning, the original point cloud data of the workpiece surface is obtained using a 3D laser scanner; The original point cloud data is denoised and smoothed, and the normal vector, principal curvature and Gaussian curvature of each discrete point are calculated based on the processed point cloud data.

Images