Blade arc grinding process

Abstract

The application provides a blade arc grinding process and belongs to the technical field of blade processing. The blade arc grinding process comprises the following steps: fixing the bottom of the blade through a lower fixing mechanism to facilitate the placement of the blade; fixing the upper part of the blade through an upper fixing mechanism to facilitate the grinding of the blade; installing the grinding wheel body through an installation mechanism to facilitate the replacement of the grinding wheel body; moving the grinding wheel body through a driving mechanism to facilitate the grinding of the blade by the grinding wheel body; and placing the blade in the fixed groove through the lower fixing mechanism, and the elasticity of the two return springs in each automatic clamping assembly enables the clamping block to automatically fix the bottom of the blade, so that the blade is placed more stably.

Description

Technical Field

[0001] This invention belongs to the field of blade processing technology, specifically relating to a blade arc grinding process. Background Technology

[0002] Pump blades all have an outer radius (R) arc surface. In the past, this outer radius arc surface was ground using an external grinder on a specially designed grinding fixture, which resulted in low processing efficiency and could not meet the needs of mass production of blades. With the significant increase in blade production, it is necessary to find a processing fixture that can grind dozens of blades at once to improve efficiency and meet market demand.

[0003] A search revealed that Chinese patent CN2684999 discloses "a blade arc grinding device, which has a grinding machine center hole on each of the two sides of the fixture body, and a groove for placing the blade to be processed on the fixture body. A cover plate is connected to the groove on the fixture body by screws, and a locking bolt for clamping the blade is provided on the cover plate. An operating handle is connected to the fixture body by bolts. It has the advantages of simple structure, low investment, low processing cost, and easy and convenient operation. First, the blade is clamped in the groove of the fixture body, and the fixture body is installed between the headstock and the center of the worktable of the cylindrical grinding machine to start the grinding work. During the grinding work, the operator uses the operating handle to make the fixture body swing repeatedly around the center line of the center point, so that the roughness and contour of the arc surface of the ground blade can meet the requirements."

[0004] However, although the above technical solution can process multiple blades at the same time, its fixing method is not stable enough, which can easily cause damage to the blades during grinding. Moreover, when grinding the blades, the grinding wheel needs to be hand-held, which is inefficient. Summary of the Invention

[0005] The purpose of this invention is to provide a blade arc grinding process, aiming to solve the problems of existing technologies that can process multiple blades simultaneously, but whose fixing methods are not stable enough, easily causing damage to the blades during grinding, and whose grinding wheels need to be handheld for grinding, resulting in low efficiency. Furthermore, another objective of this invention is to improve the accuracy and efficiency of blade arc grinding by employing advanced intelligent control technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A blade arc grinding process includes:

[0008] S1. The bottom of the blade is fixed by the lower fixing mechanism to facilitate the placement of the blade;

[0009] S2. The upper part of the blade is fixed by the upper fixing mechanism to facilitate the grinding of the blade;

[0010] S3. The grinding wheel body is installed through the installation mechanism to facilitate the replacement of the grinding wheel body;

[0011] S4. The grinding wheel body is moved by the drive mechanism so that the grinding wheel body can grind the blades.

[0012] In a preferred embodiment of the present invention, in step S1, the lower fixing mechanism includes:

[0013] Base plate;

[0014] A fixing groove is formed in the top of the base plate; and

[0015] An automatic clamping assembly, comprising multiple sets, is provided for fixing multiple blades within the fixing groove.

[0016] As a preferred embodiment of the present invention, each group of the automatic clamping components includes:

[0017] Two spring grooves are provided, which are oppositely formed on the inner walls of the two sides of the fixing groove;

[0018] Two return springs are provided, each fixed to a separate inner wall of a spring groove that is far apart.

[0019] Two clamping blocks are provided, which are respectively fixed to the adjacent ends of the two return springs;

[0020] The limiting component has two sets, which are respectively located in the two spring grooves, and are used to limit the linear movement of the clamping block.

[0021] As a preferred embodiment of the present invention, each group of the limiting components includes:

[0022] The first limiting groove has two sections, which are respectively formed on the inner walls of the two sides of the spring groove; and

[0023] The first limiting block has two parts, which are respectively fixed to the two sides of the clamping block and slidably connected in the two first limiting grooves.

[0024] In a preferred embodiment of the present invention, in step S2, the upper fixing mechanism includes:

[0025] Two fixing plates are provided, both of which are fixed to the top of the base plate;

[0026] A bidirectional lead screw is rotatably connected between the two fixed plates, and one end of each of the two bidirectional lead screws movably passes through one side of one of the fixed plates and extends outward.

[0027] Two clamping plates are provided, each threadedly connected to both sides of the circumferential surface of the two bidirectional lead screws via lead screw nuts; and

[0028] A power assembly, located on one side of the fixed plate, is used to drive the two bidirectional lead screws to rotate synchronously.

[0029] As a preferred embodiment of the present invention, the power assembly includes:

[0030] The second motor is fixed to one side of one of the fixed plates, and the output end of the second motor movably passes through one side of the fixed plate and is fixed to the end of one of the bidirectional lead screws;

[0031] Two I-beam wheels are provided, each fixed to the extended end surface of one of the two bidirectional lead screws; and

[0032] A belt is connected between the circumferential surfaces of the two I-beam pulleys.

[0033] In a preferred embodiment of the present invention, in step S3, the mounting mechanism includes:

[0034] A rectangular frame is disposed on the upper side of the base plate;

[0035] The skateboard is slidably connected to the inside of the rectangular frame;

[0036] The second limiting component is connected to the slide plate to enable it to move linearly.

[0037] Multiple mounting bases are provided, all of which are fixed to the bottom of the slide plate;

[0038] Multiple locking grooves are provided, each opened on one side of one of the mounting seats;

[0039] The mounting slider is provided in multiple forms and is respectively fixed to the inner walls of both sides of the multiple engagement grooves;

[0040] The grinding wheel body has multiple parts, which are respectively embedded and slidably connected in multiple engagement grooves;

[0041] Multiple mounting grooves are provided, each located at one end of a plurality of mounting sliders, and the plurality of mounting sliders are slidably connected within the plurality of mounting grooves; and

[0042] A fixing component, disposed on the mounting base, is used to fix multiple grinding wheel bodies into the engagement groove.

[0043] As a preferred embodiment of the present invention, the fixing component includes:

[0044] A socket that extends through both ends of the plurality of said mounting bases;

[0045] The insertion rod is movably inserted into the insertion hole;

[0046] A screw, fixed to one end of the insertion rod; and

[0047] A hexagonal nut is threaded onto the screw.

[0048] As a preferred embodiment of the present invention, the second limiting component includes:

[0049] The second limiting groove is provided in two parts, which are respectively opened on the inner walls of the two sides of the rectangular frame;

[0050] Two second limiting blocks are provided, which are respectively fixed to the two ends of the slide plate and slidably connected to the two second limiting grooves; and

[0051] Two limiting rods are provided, both fixed between the inner walls of the two sides of the rectangular frame, and the sliding plate is slidably connected to the circumferential surface of the two rectangular frames.

[0052] In a preferred embodiment of the present invention, in step S4, the driving mechanism comprises a horizontal driving component and a vertical driving component, wherein:

[0053] The lateral drive component includes:

[0054] The gantry frame is fixed to the rectangular frame;

[0055] A first motor is fixed to the top of the gantry frame, and the output end of the first motor movably passes through the bottom of the gantry frame and extends downward.

[0056] A half gear is fixed to the circumferential surface of the first motor;

[0057] A ring rack is fixed to the top of the slide plate and intermittently meshes with the half gear;

[0058] The vertical drive assembly consists of four cylinders. The bottoms of the four cylinders are fixed to the top four corners of the base plate, and the tops of the extended ends of the four cylinders are fixed to the bottom four corners of the rectangular frame.

[0059] Step S4 also includes: monitoring the shape and grinding progress of the blade in real time using sensors, and adjusting the grinding process according to the actual shape and requirements of the blade using a convolutional neural network algorithm to ensure that the arc shape of the blade is accurately achieved. The specific process is as follows:

[0060] Step a: During the grinding process of the grinding wheel body (32) grinding the blade (38), the sensors installed on the grinding wheel body (32) include a laser scanner and an ultrasonic sensor, which continuously measure the actual shape of the blade and the grinding progress, and transmit the data to the controller. During the blade grinding process, the ultrasonic sensor is used to monitor the distance change on the blade surface to estimate the grinding progress; the laser scanner is used to measure the distance between different points on the blade, and then use these distance data to construct the three-dimensional shape of the blade.

[0061] Step b: The controller receives sensor data and uses it to calculate the actual shape of the blade in real time; the sensor data is filtered, denoised, and corrected.

[0062] Step c: Before grinding begins, the operator or system user needs to define the target shape of the blade; this can be done by inputting parameters or by pre-storing standard blade shapes. The target shape is usually described by mathematical formulas or curves.

[0063] Step d: Compare the actual blade shape with the target shape and calculate the error between them; the error represents the difference between the actual blade shape and the target shape, which is the basis for controlling the grinding process;

[0064] Step e: A convolutional neural network control method is used to adjust the grinding parameters based on the error calculation results, so that the actual blade shape gradually approaches the target shape; the grinding parameters include grinding speed, grinding depth, and grinding trajectory; the convolutional neural network structure is as follows:

[0065] Input layer: The input layer takes the leaf shape data as input; it represents the shape of the leaf as an image, where each pixel represents a part of the leaf, and the value of the pixel reflects the height or thickness of the leaf;

[0066] Convolutional layer: Convolutional layers are used to detect local features of a leaf, including curvature, indentation, and convexity; a convolutional layer consists of multiple convolutional kernels, each of which performs a convolution operation on the input data, and the convolution operation can capture local features;

[0067] Pooling layers: Pooling layers are used to reduce data dimensionality, decrease computational complexity, and make the network insensitive to changes in the position of the image;

[0068] Fully connected layer: The fully connected layer takes data from the pooling layer, flattens it, and connects it to the output layer of the neural network. The fully connected layer is used to learn complex relationships between data.

[0069] Output layer: The output layer generates control signals to adjust grinding parameters. The number of nodes in the output layer depends on the dimension of the control signals.

[0070] Convolution operation:

[0071] Convolution operations involve applying a convolution kernel to different locations in the input data to produce a convolutional feature map. Assuming the input data is I, the convolution kernel is K, and the output feature map is O, the convolution operation is as follows:

[0072] O(i,j)=∑ m∑ nI(i+m,j+n) K(m,n)

[0073] Where O(i, j) represents the pixel value of the output feature map, i and j represent the coordinates of the pixel, and m and n represent the coordinates of the convolution kernel. * indicates a convolution operation;

[0074] Pooling operations:

[0075] Pooling operations are used to reduce the dimensionality of data. Common pooling operations include max pooling and average pooling. Max pooling is performed as follows:

[0076] O(i,j)=max m,nI(i+m,j+n)

[0077] Here, i and j represent the coordinates of the pixels, m and n represent the coordinates of the pooling window, and max represents taking the maximum value within the window;

[0078] The above input data I:

[0079] [[1, 2, 3, 4, 5],

[0080] [6, 7, 8, 9, 10],

[0081] [11, 12, 13, 14, 15],

[0082] [16, 17, 18, 19, 20],

[0083] [21, 22, 23, 24, 25]]

[0084] Convolution kernel K:

[0085] [[0.1, 0.2],

[0086] [0.3, 0.4]]

[0087] Convolution operation result O:

[0088] [[10.2, 14.4, 18.6],

[0089] [30.2, 34.4, 38.6],

[0090] [50.2, 54.4, 58.6]]

[0091] The output O of the convolution operation is then passed to the pooling layer, and the output of the pooling layer is then passed to the fully connected layer and the output layer to generate control signals.

[0092] Step f: The control algorithm continuously monitors sensor data and constantly adjusts grinding parameters to minimize errors until the actual blade shape approaches or reaches the target shape;

[0093] Step g: The control algorithm analyzes the initial state of the blade and the desired final shape, and then plans the optimal grinding path; specifically, it includes the following processes:

[0094] g.1. First, it is necessary to clearly define the goal and constraints of the problem. In blade grinding, the goal is to find a path to minimize wear and improve grinding efficiency. The constraints include the physical limitations of the grinding machine, the material properties of the blade, and the desired final shape.

[0095] g.2. Model the problem in mathematical form so that it can be solved using optimization algorithms. The problem can be modeled in the following ways:

[0096] Define the path: Divide the blade surface into a grid, with each grid cell representing a possible grinding point;

[0097] Define the state: Each mesh cell can represent a state, which includes the current shape of the blade, the required final shape, and material property information;

[0098] Define path cost: Define a cost function for movement between each state, which can include grinding cost, time cost, and energy cost;

[0099] Objective function: The objective function typically includes minimizing the total cost to simultaneously satisfy the objectives of minimizing wear and improving grinding efficiency;

[0100] g.3. Select an optimization algorithm suitable for the problem to find the optimal path;

[0101] First, define the objective function and the cost function: Suppose we want to grind a blade from its initial shape S to its final shape F, where S and F can both be represented as sets of curves or points:

[0102] Define the objective function J, which includes the combination of wear cost W, time cost (grinding time T), and energy cost (energy E required for grinding).

[0103] J = W + αT + βE

[0104] Here, α and β are weighting parameters used to adjust the impact of grinding time T and grinding energy E on the final cost; then, the path planning problem is transformed into an optimization problem, namely minimizing the objective function J, which can be achieved by selecting the optimal path to minimize J;

[0105] In the grinding path planning problem, the simulated annealing algorithm is used to search for the optimal path to meet the objectives of minimizing wear and improving grinding efficiency. The specific implementation steps are as follows:

[0106] g.3.1. Randomly generate or select an initial path as the current solution. The higher the initial temperature, the higher the probability of accepting bad solutions, which helps to escape local minima. Set the initial temperature T, which is usually a high value. Set termination conditions, such as the maximum number of iterations or reaching a certain temperature.

[0107] g.3.2. In each iteration, the current solution is slightly perturbed to generate a new candidate solution, and the difference in objective function value between the current solution and the candidate solution is calculated: the change in cost ΔJ;

[0108] If ΔJ is less than 0, accept the candidate solution as the new current solution;

[0109] If ΔJ is greater than 0, the candidate solution is accepted with a certain probability. The acceptance probability is affected by temperature T and ΔJ, and is usually calculated according to the Metropolis criterion: P = e( ΔJ / T); where P is the acceptance probability, e is the base of the natural logarithm, ΔJ is the cost change, and T is the current temperature; gradually decrease the temperature T, usually according to a predefined cooling strategy. The rate of temperature decrease is usually a key parameter that determines the performance of the algorithm; repeat the above steps until the termination condition is met.

[0110] g.3.3. When the termination condition is met, the algorithm ends and returns the final path as the optimization result;

[0111] g.4. Perform the optimization iteration process to find the best path. The optimization algorithm will search in the state space in an attempt to find the path that minimizes the objective function.

[0112] g.5. Once the optimal path is found, it is implemented in the actual blade grinding, which involves controlling the grinding machine to grind according to the planned path.

[0113] Compared with the prior art, the beneficial effects of the present invention are:

[0114] 1. In this solution, the lower fixing mechanism enables the blade to be placed in the fixing slot. The elasticity of the two reset springs in each automatic clamping assembly allows the clamping block to automatically fix the bottom of the blade, making the blade placement more stable.

[0115] 2. In this solution, the upper fixing mechanism is provided, in which the power component can drive the two clamping plates to move inward simultaneously to fix the blade, making the blade more firmly fixed. At the same time, through the cooperation of the upper fixing mechanism and the lower fixing mechanism, the upper and lower parts of the blade can be fixed at the same time, making the blade more stable during grinding and less prone to tilting or breakage.

[0116] 3. In this solution, the installation mechanism allows the grinding wheel body to be easily disassembled from the mounting base, facilitating the replacement of the grinding wheel body. Damaged grinding wheel bodies or those with different curvatures can be easily replaced, making it easier to grind the blades to the corresponding angle according to actual needs, thus making it more practical.

[0117] 4. In this solution, the first motor is started, and its output end can drive the half gear to move, so that the ring rack and the half gear reciprocate when they mesh intermittently, thereby causing the slide plate to reciprocate back and forth. The slide plate can drive the grinding wheel body to reciprocate through the mounting base, realizing automatic reciprocating grinding, which is more convenient.

[0118] 5. In this solution, the provided cylinder can drive the rectangular frame to move, thereby causing the slide plate to move downward. The slide plate can drive the grinding wheel body to move downward through the mounting base, so that the grinding wheel body can move downward slowly during grinding, which can better perform the grinding operation.

[0119] 6. By monitoring the shape of the blade and the grinding progress in real time through sensors, the convolutional neural network algorithm can adjust the grinding process according to the actual shape and requirements of the blade, ensuring that the arc shape of the blade is accurately realized. Furthermore, the intelligent control algorithm can analyze the initial state of the blade and the required final shape, and then plan the optimal grinding path to minimize wear and improve grinding efficiency. Attached Figure Description

[0120] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0121] Figure 1 This is a first-view perspective perspective view of the present invention;

[0122] Figure 2 This is a partial exploded view of the lateral drive component of the present invention;

[0123] Figure 3 This is a partial perspective view of the mounting mechanism of the present invention;

[0124] Figure 4 This is a partial perspective view of the lower fixing mechanism of the present invention;

[0125] Figure 5 This is a cross-sectional view of the base plate of the present invention;

[0126] Figure 6 This is a partial perspective view of the upper fixing mechanism of the present invention;

[0127] Figure 7 This is a second-view perspective perspective view of the present invention;

[0128] Figure 8 This is a bottom view of the present invention;

[0129] Figure 9 This is a third-view perspective view of the present invention.

[0130] In the diagram: 1. Base plate; 2. Fixing plate; 3. Two-way lead screw; 4. Lead screw nut; 5. Clamping plate; 6. Belt; 7. I-beam pulley; 8. Cylinder; 9. Rectangular frame; 10. Limiting rod; 11. Gantry frame; 12. First motor; 13. Half gear; 14. Ring rack; 15. Slide plate; 16. Second limiting block; 17. Second limiting groove; 18. Storage groove; 19. Third limiting groove; 20. Storage plate; 21. Third limiting block; 22. Fixing groove; 23. Spring groove; 24. First limiting groove; 25. Clamping block; 26. First limiting block; 27. Separator block; 28. Return spring; 29. ​​Insertion hole; 30. Insertion rod; 31. Mounting base; 32. Grinding wheel body; 33. Mounting slide groove; 34. Mounting slider; 35. Engaging groove; 36. Screw; 37. Hexagonal nut; 38. Blade; 39. Second motor. Detailed Implementation

[0131] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0132] Example 1

[0133] Please see Figures 1-9 The present invention provides the following technical solutions:

[0134] A blade arc grinding process includes:

[0135] S1. The bottom of the blade 38 is fixed by the lower fixing mechanism to facilitate the placement of the blade 38;

[0136] S2. The upper part of the blade 38 is fixed by the upper fixing mechanism to facilitate the grinding of the blade 38;

[0137] S3. The grinding wheel body 32 is installed through the installation mechanism to facilitate the replacement of the grinding wheel body 32;

[0138] S4. The grinding wheel body 32 is moved by the drive mechanism so that the grinding wheel body 32 can grind the blade 38.

[0139] Specifically, in step S1, the lower fixing mechanism includes:

[0140] Base plate 1;

[0141] A fixing groove 22 is formed at the top of the base plate 1; and

[0142] An automatic clamping assembly, which has multiple sets, is used to fix multiple blades 38 in the fixing groove 22.

[0143] In this embodiment, the fixing groove 22 is located at the center of the top of the base plate 1 and is used to place multiple blades 38 at the same time. With the automatic clamping component, the bottom of the blade 38 can be automatically fixed when the blade 38 is placed in the fixing groove 22.

[0144] Specifically, each set of automatic clamping components includes:

[0145] Two spring grooves 23 are provided, which are oppositely opened on the inner walls of the two sides of the fixing groove 22;

[0146] Two return springs 28 are provided, which are respectively fixed on the inner walls of two spring grooves 23 that are far apart;

[0147] Two clamping blocks 25 are provided, which are respectively fixed to the close ends of the two return springs 28;

[0148] The limiting component has two sets, which are located in two spring grooves 23 respectively, and are used to limit the linear movement of the clamping block 25.

[0149] In this embodiment, the upper sides of the two clamping blocks 25 in each automatic clamping assembly are inclined, so that when the blade 38 is inserted between the two clamping blocks 25, it can enter between the two clamping blocks 25. The elasticity of the return spring 28 holds the two clamping blocks 25 together. When the blade 38 is inserted into the clamping block 25, the two clamping blocks 25 clamp the bottom of the blade 38 and fix it. The limiting mechanism makes the two clamping blocks 25 move linearly in the spring groove 23 respectively, and it is not easy to tilt.

[0150] Specifically, each set of limit components includes:

[0151] The first limiting groove 24 has two sections, which are respectively formed on the inner walls of the two sides of the spring groove 23; and

[0152] Two first limiting blocks 26 are provided, which are respectively fixed to the two sides of the clamping block 25 and slidably connected in the two first limiting grooves 24.

[0153] In this embodiment: when the clamping block 25 moves in the spring groove 23, it drives the first limiting block 26 to slide in the first limiting groove 24. The first limiting groove 24 is not connected to the fixed groove 22, so that the first limiting block 26 will not detach from the first limiting groove 24. This makes it less likely for the clamping block 25 to deviate or tilt when sliding, and can better clamp the blade 38.

[0154] Preferably, a partition block 27 is provided between every two blades 38, and multiple partition blocks 27 are located on the bottom wall of the fixing groove 22 to facilitate the insertion of blades 38, making it easier to insert blades 38 and preventing two adjacent blades 38 from contacting each other, thus reducing the likelihood of interference.

[0155] Preferably, the bottom of the base plate 1 is provided with a storage groove 18, and the inner walls of both sides of the storage groove 18 are provided with third limiting grooves 19. The storage plate 20 is embedded in the storage groove 18. The storage plate 20 is "L" shaped. Both sides of the storage plate 20 are fixed with third limiting blocks 21. The third limiting blocks 21 are embedded in the third limiting grooves 19, so that the storage plate 20 can slide in the storage groove 18 and is not easy to slide out. The top of the storage plate 20 is left with a gap between it and the top wall of the storage groove 18, which can be used to store things, making it more practical.

[0156] Specifically, in step S2, the upper fixing mechanism includes:

[0157] There are two fixing plates 2, both of which are fixed to the top of the base plate 1;

[0158] A bidirectional lead screw 3 is rotatably connected between two fixed plates 2, and one end of each bidirectional lead screw 3 movably passes through one side of one of the fixed plates 2 and extends outward.

[0159] Two clamping plates 5 are provided, each threadedly connected to both sides of the circumferential surface of two bidirectional lead screws 3 via lead screw nuts 4; and

[0160] The power assembly, located on one side of the fixed plate 2, is used to drive the two bidirectional lead screws 3 to rotate synchronously.

[0161] In this embodiment: the upper fixing mechanism is used to fix the upper part of the blade 38, excluding the upper surface of the blade 38. The upper surface of the blade 38 is higher than the clamping plate 5. Two screw nuts 4 are provided on each of the two clamping plates 5. Each bidirectional screw 3 is threadedly connected to the two screw nuts 4. When the bidirectional screw 3 rotates, the two screw nuts 4 move simultaneously in the direction of approaching or moving away from each other, so that multiple screw nuts 4 drive the two clamping plates 5 to move in the direction of approaching or moving away from each other. When moving inward, the upper parts of multiple blades 38 can be clamped, making the blade 38 more stable when being ground.

[0162] Specifically, the power components include:

[0163] The second motor 39 is fixed to one side of one of the fixing plates 2, and the output end of the second motor 39 movably passes through one side of the fixing plate 2 and is fixed to the end of one of the bidirectional lead screws 3;

[0164] Two I-beam wheels 7 are provided, each fixed to the extended end surface of one of the two bidirectional lead screws 3; and

[0165] The belt 6 is connected between the circumferential surfaces of the two I-beam pulleys 7.

[0166] In this embodiment: by setting up the power component, by starting the second motor 39, one of the bidirectional lead screws 3 is driven to rotate, the bidirectional lead screw 3 drives one of the I-beam pulleys 7 to rotate, and through the transmission of the belt 6, the other I-beam pulley 7 is rotated, which drives the other bidirectional lead screw 3 to rotate, so that the two bidirectional lead screws 3 rotate synchronously, thereby driving the clamping plate 5 to complete the clamping operation.

[0167] Specifically, in step S3, the installation mechanism includes:

[0168] Rectangular frame 9 is set on the upper side of base plate 1;

[0169] The skateboard 15 is slidably connected to the inside of the rectangular frame 9;

[0170] The second limiting component is connected to the slide plate 15 to enable it to move in a straight line.

[0171] Multiple mounting bases 31 are provided, all of which are fixed to the bottom of the slide plate 15;

[0172] Multiple locking slots 35 are provided and are respectively opened on one side of multiple mounting bases 31;

[0173] Multiple mounting sliders 34 are provided and are respectively fixed to the inner walls of both sides of multiple engagement slots 35;

[0174] The grinding wheel body 32 has multiple parts, which are respectively embedded and slidably connected in multiple engagement grooves 35;

[0175] Multiple mounting grooves 33 are provided, each located on both sides of a plurality of mounting sliders 34, and the plurality of mounting sliders 34 are slidably connected within the plurality of mounting grooves 33; and

[0176] A fixing component is provided on the mounting base 31 for fixing multiple grinding wheel bodies 32 into the engagement groove 35.

[0177] In this embodiment: the provided mounting mechanism allows for easy disassembly of the grinding wheel body 32 from the mounting base 31, facilitating the replacement of the grinding wheel body 32. It also facilitates the replacement of damaged or differently curved grinding wheel bodies 32, enabling the grinding of blades at corresponding angles according to actual needs, thus enhancing practicality. The grinding wheel body 32 is movably connected within the engaging groove 35. The matching of the mounting slider 34 and the mounting groove 33 ensures that the grinding wheel body 32 is embedded in the engaging groove 35 and will not fall out. Then, the insertion rod 30 passes through the insertion hole 29, abutting one end of multiple grinding wheel bodies 32, ensuring that the grinding wheel bodies 32 remain stationary within the engaging groove 35 and are fixed in place. The insertion rod 30 is then fixed within the insertion hole 29 by a fixing component, thereby securing the grinding wheel bodies 32. This stable fixation of multiple grinding wheel bodies 32 allows for better grinding of the blades 38.

[0178] Specifically, the fixed components include:

[0179] The socket 29 extends through both ends of the plurality of mounting bases 31;

[0180] The insertion rod 30 is movably inserted into the insertion hole 29;

[0181] Screw 36 is fixed to one end of insert rod 30; and

[0182] Hex nut 37 is threaded onto screw 36.

[0183] In this embodiment: after the insertion rod 30 passes through the insertion hole 29, one end of the insertion rod 30 has a tip that fits against one side of the mounting base 31, so that the insertion rod 30 does not pass through the insertion hole 29 as a whole. The other end of the insertion rod 30 is fixed with a screw 36. The screw 36 passes through the insertion hole 29 and is threaded onto the screw 36 by a hexagonal nut 37, which fixes the insertion rod 30 so that the insertion rod 30 cannot move inside the insertion hole 29, thereby fixing the grinding wheel body 32.

[0184] Specifically, the second limiting component includes:

[0185] The second limiting groove 17 is provided in two parts, which are respectively opened on the inner walls of the two sides of the rectangular frame 9;

[0186] Two second limiting blocks 16 are provided, which are respectively fixed to the two ends of the slide plate 15 and slidably connected to the two second limiting grooves 17; and

[0187] There are two limit rods 10, both of which are fixed between the inner walls of the two sides of the rectangular frame 9. The slide plate 15 is slidably connected to the circumferential surface of the two rectangular frames 9.

[0188] In this embodiment, the slide plate 15 can slide within the rectangular frame 9. With the limit rod 10 provided, the slide plate 15 can only slide in a straight line back and forth when it moves. At the same time, the slide plate 15 drives the second limit block 16 to move within the second limit groove 17, so that the position of the slide plate 15 is further limited and the movement is more stable.

[0189] Specifically, in step S4, the drive mechanism consists of a lateral drive component and a vertical drive component, wherein:

[0190] The lateral drive components include:

[0191] Gantry frame 11 is fixed to rectangular frame 9;

[0192] The first motor 12 is fixed to the top of the gantry frame 11, and the output end of the first motor 12 moves through the bottom of the gantry frame 11 and extends downward.

[0193] Half gear 13 is fixed to the circumferential surface of the first motor 12;

[0194] The ring rack 14 is fixed to the top of the slide plate 15 and intermittently meshes with the half gear 13;

[0195] The vertical drive assembly consists of four cylinders 8. The bottom of the four cylinders 8 is fixed to the top four corners of the base plate 1, and the top of the extended ends of the four cylinders 8 is fixed to the bottom four corners of the rectangular frame 9.

[0196] In this embodiment: the first motor 12 is started, and its output end can drive the half gear 13 to move, so that the ring rack 14 and the half gear 13 reciprocate when they mesh intermittently, thereby causing the slide plate 15 to reciprocate back and forth. The slide plate 15 can drive the grinding wheel body 32 to reciprocate through the mounting base 31, realizing automatic reciprocating grinding, which is more convenient; the cylinder 8 can drive the rectangular frame 9 to move, thereby causing the slide plate 15 to move downward. The slide plate 15 can drive the grinding wheel body 32 to move downward through the mounting base 31, so that the grinding wheel body 32 can move downward slowly during grinding, which can better perform grinding operations.

[0197] It should be noted that cylinder 8, first motor 12, and second motor 39 are all existing technologies, and the corresponding models can be selected according to actual needs. Further details will not be provided.

[0198] The principle and specific steps of the blade arc grinding process provided by this invention are as follows:

[0199] S1. First, place multiple blades 38 in the fixing groove 22 between two opposing clamping blocks 25, insert the blades 38 downwards, and then clamp the bottom of the blades 38 by the elasticity of the return spring 28.

[0200] S2. At this time, start the second motor 39, which drives one of the double-acting lead screws 3 to rotate. The double-acting lead screw 3 drives one of the I-beam pulleys 7 to rotate. Through the transmission of the belt 6, the other I-beam pulley 7 rotates, which drives the other double-acting lead screw 3 to rotate. Thus, the two double-acting lead screws 3 rotate synchronously, and the two lead screw nuts 4 move towards each other at the same time. The lead screw nuts 4 drive the two clamping plates 5 to move towards each other, which can clamp the upper part of multiple blades 38.

[0201] S3. At this time, insert multiple grinding wheel bodies 32 into multiple engaging slots 35, aligning the grinding wheel bodies 32 with the blades 38. The grinding wheel bodies 32 can be selected according to the actual needs, with the corresponding arc of the grinding wheel slightly higher than the top of the blades 38. Then, insert the insertion rod 30 into the insertion hole 29 and thread the hexagonal nut 37 onto the screw 36 to fix the insertion rod 30, so that the insertion rod 30 is fixed in the insertion hole 29. The insertion rod 30 fixes the multiple grinding wheel bodies 32 into the engaging slots 35.

[0202] S4. Start the first motor 12 in the transverse drive assembly. Its output end can drive the half gear 13 to move, so that the ring rack 14 and the half gear 13 intermittently mesh and reciprocate, thereby causing the slide plate 15 to reciprocate back and forth. The slide plate 15 can drive multiple grinding wheel bodies 32 to reciprocate back and forth through multiple mounting seats 31 to realize reciprocating grinding operation.

[0203] S5. Finally, start cylinder 8, which can drive the rectangular frame 9 to move, thereby causing the slide plate 15 to move downward. When the grinding wheel body 32 moves back and forth, it slowly moves downward to realize the automatic grinding of the blade 38, so that the blade 38 can be ground to the corresponding arc according to the actual needs, which is more convenient.

[0204] By using sensors to monitor the shape and grinding progress of the blades in real time, and by using convolutional neural network algorithms to adjust the grinding process according to the actual shape and requirements of the blades, the arc shape of the blades can be accurately achieved.

[0205] Step 1: During the grinding process, sensors installed on the equipment include a laser scanner and ultrasonic sensors. The actual shape of the blade and the grinding progress are continuously measured, and the data is transmitted to the controller. Ultrasonic sensors can be used to measure distances or surface features of an object. During blade grinding, ultrasonic sensors can be used to monitor distance changes on the blade surface to estimate the grinding progress; laser rangefinders can be used to measure distances at different points on the blade, and then this distance data is used to construct the three-dimensional shape of the blade.

[0206] Step 2: The controller receives sensor data and uses it to calculate the actual shape of the blade in real time; this usually involves filtering, denoising and correcting the sensor data to ensure the accuracy and stability of the data.

[0207] Step 3: Before grinding begins, the operator or system user needs to define the target shape of the blade; this can be done by inputting parameters or by pre-storing standard blade shapes. The target shape is usually described by mathematical formulas or curves.

[0208] Step 4: The intelligent control algorithm compares the actual blade shape with the target shape and calculates the error between the two. The error represents the difference between the actual blade shape and the target shape, which is the basis for controlling the grinding process.

[0209] Step 5: The control algorithm adjusts the grinding parameters based on the error calculation results to gradually bring the actual blade shape closer to the target shape. These parameters may include grinding speed, grinding depth, grinding trajectory, etc. The adjustment can be achieved using a convolutional neural network control method to ensure the accuracy and stability of the grinding process.

[0210] Input layer: The input layer takes the leaf shape data as input; the shape of the leaf can be represented as an image, where each pixel represents a part of the leaf, and the value of the pixel reflects the height or thickness of the leaf.

[0211] Convolutional layers: Convolutional layers are used to detect local features of a leaf, such as curvature, indentations, and convexities. A convolutional layer consists of multiple convolutional kernels, each of which performs a convolution operation on the input data. Convolutional operations can capture local features.

[0212] Pooling layers: Pooling layers are used to reduce data dimensionality, decrease computational complexity, and make the network insensitive to changes in image location. Common pooling operations include max pooling and average pooling.

[0213] Fully connected layers: Fully connected layers take data from pooling layers, flatten it, and connect it to the output layer of the neural network. Fully connected layers are used to learn complex relationships between data.

[0214] Output Layer: The output layer generates control signals to adjust grinding parameters, such as grinding speed or depth. The number of nodes in the output layer depends on the dimension of the control signals.

[0215] Convolution operation:

[0216] Convolution operations involve applying a convolution kernel to different locations in the input data to produce a convolutional feature map. Assume the input data is I, the convolution kernel is K, and the output feature map is O. The convolution operation is shown below:

[0217] O(i,j)=∑ m∑ nI(i+m,j+n) K(m,n)

[0218] Where O(i, j) represents the pixel value of the output feature map, i and j represent the coordinates of the pixel, and m and n represent the coordinates of the convolution kernel. * indicates the convolution operation.

[0219] Pooling operations:

[0220] Pooling operations are used to reduce the dimensionality of data. Common pooling operations include max pooling and average pooling. Max pooling is performed as follows:

[0221] O(i,j)=max m,nI(i+m,j+n)

[0222] Here, O(i, j) represents the pixel value after pooling, i and j represent the pixel coordinates, and m and n represent the coordinates of the pooling window. max represents taking the maximum value within the window.

[0223] Input data I:

[0224] [[1, 2, 3, 4, 5],

[0225] [6, 7, 8, 9, 10],

[0226] [11, 12, 13, 14, 15],

[0227] [16, 17, 18, 19, 20],

[0228] [21, 22, 23, 24, 25]]

[0229] Convolution kernel K:

[0230] [[0.1, 0.2],

[0231] [0.3, 0.4]]

[0232] Convolution operation result O:

[0233] [[10.2, 14.4, 18.6],

[0234] [30.2, 34.4, 38.6],

[0235] [50.2, 54.4, 58.6]]

[0236] Next, we can pass the output O of the convolution operation to the pooling layer, and then pass the output of the pooling layer to the fully connected layer and the output layer to generate control signals. The specific mathematical formulas and parameters of these operations need to be determined based on the actual neural network architecture and task requirements.

[0237] Step 6: The intelligent control algorithm continuously monitors sensor data and adjusts grinding parameters to minimize errors. This is a closed-loop control process until the actual blade shape approaches or reaches the target shape.

[0238] Intelligent control algorithms can analyze the initial state of the blade and the desired final shape, and then plan the optimal grinding path to minimize wear and improve grinding efficiency.

[0239] Optimizing grinding path planning is a complex problem aimed at finding a path that grinds a blade from its initial state to the desired final shape in the shortest possible time, while minimizing wear and maximizing grinding efficiency. This typically involves the use of path planning and optimization algorithms. Below is a simplified implementation example, including the basic mathematical expression and algorithmic approach:

[0240] 1. Define the problem:

[0241] First, it is necessary to clearly define the problem's objectives and constraints. In blade grinding, the objective is to find a path that minimizes wear (such as loss of blade material) and improves grinding efficiency (such as reducing grinding time). Constraints may include the physical limitations of the grinding machine, the material properties of the blade, and the desired final shape.

[0242] 2. Modeling problem:

[0243] The problem is modeled in mathematical form so that it can be solved using optimization algorithms. Typically, a problem can be modeled in the following way:

[0244] Define the path: Divide the blade surface into a grid, with each grid cell representing a possible grinding point.

[0245] Define state: Each mesh cell can represent a state, which includes information such as the current shape of the blade, the required final shape, and material properties.

[0246] Define path cost: Define a cost function for movement between each state, which can include grinding cost, time cost, energy cost, etc.

[0247] Objective function: The objective function typically includes minimizing the total cost to simultaneously satisfy the objectives of minimizing wear and improving grinding efficiency.

[0248] 3. Select an optimization algorithm:

[0249] Choosing an optimization algorithm suitable for the problem is crucial for finding the optimal path. Common optimization algorithms include genetic algorithms, simulated annealing, and gradient descent. The choice of these algorithms depends on the complexity of the problem and its constraints.

[0250] 4. Optimization process:

[0251] The optimization algorithm performs an iterative process to find the optimal path. It searches the state space, attempting to find a path that minimizes the objective function.

[0252] 5. Implementation path:

[0253] Once the optimal path is found, it is implemented in the actual blade grinding. This may involve controlling the grinding machine to grind along the planned path.

[0254] Derivation of mathematical formulas:

[0255] The most important steps in the above process are defining the objective function and the cost function. Below is a simple example illustrating the possible forms of the objective function:

[0256] Suppose we want to grind a blade from its initial shape S to its final shape F, where S and F can both be represented as a set of curves or points.

[0257] Define an objective function J, which can include a combination of wear cost W (material loss), time cost T (grinding time), and energy cost E (energy required for grinding):

[0258] J = W + αT + βE

[0259] Here, α and β are weighting parameters used to adjust the impact of time cost and energy cost on the final cost.

[0260] The path planning problem can then be transformed into an optimization problem, namely, minimizing the objective function J. This can be achieved by selecting the optimal path to minimize J.

[0261] In the grinding path planning problem, the simulated annealing algorithm is used to search for the optimal path to minimize wear and improve grinding efficiency. The following are the basic implementation steps of the simulated annealing algorithm:

[0262] 1. Initialization:

[0263] An initial path is randomly generated or selected as the current solution. A higher initial temperature increases the probability of accepting a bad solution, which helps to escape local minima, but may also lead to more randomness.

[0264] Set the initial temperature T, which is usually a relatively high value.

[0265] Set termination conditions, such as the maximum number of iterations or reaching a certain temperature.

[0266] 2. Iterative process:

[0267] In each iteration, a small perturbation is applied to the current solution to generate a new candidate solution.

[0268] Calculate the difference in objective function values ​​between the current solution and the candidate solutions (e.g., the change in cost ΔJ).

[0269] If ΔJ is less than 0, accept the candidate solution as the new current solution.

[0270] If ΔJ is greater than 0, the candidate solution is accepted with a certain probability. The acceptance probability is affected by temperature T and ΔJ, and is usually determined according to the Metropolis law.

[0271] Criterion calculation: P=e ( ΔJ / T)

[0272] Where P is the acceptance probability, e is the base of the natural logarithm, ΔJ is the cost change, and T is the current temperature.

[0273] The temperature T is gradually decreased, typically following a predefined cooling strategy. The rate of temperature decrease is usually a key parameter that determines the algorithm's performance.

[0274] Repeat the above steps until the termination condition is met.

[0275] 3. Termination:

[0276] The algorithm terminates when the termination condition is met (such as the maximum number of iterations or the minimum temperature is reached).

[0277] Return the final path as the optimization result.

[0278] Key parameters of the simulated annealing algorithm:

[0279] Initial temperature T: The choice of initial temperature can affect the performance of the algorithm. Generally, a higher initial temperature increases the probability of accepting bad solutions, which helps to escape local minima, but may also lead to more randomness. The initial temperature can be adjusted according to the characteristics of the problem.

[0280] Cooling strategy: The cooling strategy determines how the temperature changes over time. Common cooling strategies include exponential decay and linear decay. The choice of cooling rate affects the algorithm's performance and convergence speed.

[0281] Acceptance probability: The acceptance probability is calculated based on the Metropolis criterion, where ΔJ and T are the key parameters. Higher temperatures (T) make it easier to accept inferior solutions, while the acceptance probability gradually decreases as the temperature decreases.

[0282] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A blade arc grinding process, characterized in that, The utility model relates to a kind of leaf grinding machine, including: S1, the fixation of blade (38) bottom is realized by lower fixed mechanism, to facilitate the placement of blade (38), wherein, The lower fixed mechanism includes: Bottom plate (1); Fixed groove (22) is opened in the top of the bottom plate (1);And Automatic clamping assembly is provided with multiple groups, for fixing multiple the blade (38) in the fixed groove (22); S2, the fixation of blade (38) upper part is realized by upper fixed mechanism, to facilitate the grinding of blade (38), wherein, The upper fixed mechanism includes: Fixed plate (2) is provided with two, and is fixed on the top of the bottom plate (1); Two-way screw rod (3) is rotatably connected between two fixed plates (2), and one end of two two-way screw rods (3) is movably penetrated through one side end of one of the fixed plates (2) and extends outward; Clamping plate (5) is provided with two, respectively through screw nut (4) thread connection on the circumference surface two sides of two two-way screw rods (3);And Power assembly is arranged on the side of the fixed plate (2), for driving two two-way screw rods (3) synchronous rotation; S3, the installation of grinding wheel main body (32) is realized by mounting mechanism, to facilitate the replacement of grinding wheel main body (32), wherein, The mounting mechanism includes: Rectangular frame (9) is arranged on the upper side of the bottom plate (1); Slide plate (15) is slidably connected to the inner side of the rectangular frame (9); Second limiting component is connected with the slide plate (15) to realize its linear motion; Mounting seat (31) is provided with multiple, and is fixed on the bottom of the slide plate (15); Engagement groove (35) is provided with multiple, respectively opened in one side end of multiple mounting seats (31); Mounting sliding block (34) is provided with multiple, respectively fixed on the two side inner walls of multiple engagement grooves (35); Grinding wheel main body (32) is provided with multiple, respectively embedded and slidably connected in multiple engagement grooves (35); Mounting sliding groove (33) is provided with multiple, respectively opened in two side ends of multiple mounting sliding blocks (34), and multiple mounting sliding blocks (34) are slidably connected in multiple mounting sliding grooves (33);And Fixing assembly is arranged on the mounting seat (31), for fixing multiple grinding wheel main bodies (32) in the engagement groove (35); The fixing assembly includes: The hole (29) is penetrated through the two side ends of multiple mounting seats (31); Insert rod (30) is movably inserted into the hole (29); Screw rod (36) is fixed on one end of the insert rod (30);And Hexagonal nut (37) is threadedly connected on the screw rod (36); S4, the movement of grinding wheel main body (32) is realized by driving mechanism, to facilitate the grinding of blade (38) by grinding wheel main body (32), wherein, The driving mechanism is composed of horizontal driving assembly and vertical driving assembly, wherein: The horizontal driving assembly includes: Portal frame (11) is fixed on the rectangular frame (9); The first motor (12) is fixed to the top of the gantry (11), and the output end of the first motor (12) moves through the bottom of the gantry (11) and extends downward. Half gear (13) is fixed to the circumferential surface of the first motor (12); A ring rack (14) is fixed to the top of the slide plate (15) and intermittently meshes with the half gear (13); The vertical drive assembly consists of four cylinders (8). The bottoms of the four cylinders (8) are respectively fixed to the top four corners of the base plate (1), and the tops of the extended ends of the four cylinders (8) are respectively fixed to the bottom four corners of the rectangular frame (9).

2. A blade arc grinding process according to claim 1, characterized in that Each of the aforementioned automatic clamping components includes: Two spring grooves (23) are provided, which are opened opposite to each other on the inner walls of the two sides of the fixing groove (22); Two return springs (28) are provided, which are respectively fixed on the inner walls of the two spring grooves (23) that are far apart; Two clamping blocks (25) are provided, which are respectively fixed to the adjacent ends of the two return springs (28); The limiting component has two sets, which are located in the two spring grooves (23) respectively, and are used to limit the linear movement of the clamping block (25).

3. A blade arc grinding process according to claim 2, characterized in that Each set of the limiting components includes: The first limiting groove (24) has two sections, which are respectively opened on the inner walls of the two sides of the spring groove (23); and Two first limiting blocks (26) are provided, which are respectively fixed to the two sides of the clamping block (25) and slidably connected in the two first limiting grooves (24).

4. A blade arc grinding process according to claim 3, characterized in that The power assembly includes: The second motor (39) is fixed to one side of one of the fixing plates (2), and the output end of the second motor (39) movably passes through one side of the fixing plate (2) and is fixed to the end of one of the bidirectional lead screws (3); Two I-beam wheels (7) are provided, each fixed to the extended end surface of one of the two bidirectional lead screws (3); and The belt (6) is connected between the circumferential surfaces of the two I-beams (7).

5. A blade arc grinding process as claimed in claim 1, wherein, The second limiting component includes: The second limiting groove (17) is provided in two places, which are respectively opened on the inner walls of the two sides of the rectangular frame (9); Two second limiting blocks (16) are provided, which are respectively fixed to the two ends of the slide plate (15) and slidably connected in the two second limiting grooves (17); and Two limit rods (10) are provided, both fixed between the inner walls of the two sides of the rectangular frame (9), and the slide plate (15) is slidably connected to the circumferential surface of the two rectangular frames (9).

6. A blade arc grinding process as claimed in claim 1, wherein, Step S4 also includes: monitoring the shape and grinding progress of the blade in real time using sensors, and adjusting the grinding process according to the actual shape and requirements of the blade using a convolutional neural network algorithm to ensure that the arc shape of the blade is accurately achieved. The specific process is as follows: Step a: During the grinding process of the grinding wheel body (32) grinding the blade (38), the sensors installed on the grinding wheel body (32) include a laser scanner and an ultrasonic sensor, which continuously measure the actual shape of the blade and the grinding progress, and transmit the data to the controller. During the blade grinding process, the ultrasonic sensor is used to monitor the distance change on the blade surface to estimate the grinding progress; the laser scanner is used to measure the distance between different points on the blade, and then use these distance data to construct the three-dimensional shape of the blade. Step b: The controller receives sensor data and uses it to calculate the actual shape of the blade in real time; the sensor data is filtered, denoised, and corrected. Step c: Before grinding begins, the operator or system user needs to define the target shape of the blade; this is done by inputting parameters or by pre-stored standard blade shapes, and the target shape is described by mathematical formulas or curves. Step d: Compare the actual blade shape with the target shape and calculate the error between them; the error represents the difference between the actual blade shape and the target shape, which is the basis for controlling the grinding process; Step e: A convolutional neural network control method is used to adjust the grinding parameters based on the error calculation results, so that the actual blade shape gradually approaches the target shape; the grinding parameters include grinding speed, grinding depth, and grinding trajectory; the convolutional neural network structure is as follows: Input layer: The input layer takes the leaf shape data as input; it represents the shape of the leaf as an image, where each pixel represents a part of the leaf, and the value of the pixel reflects the height or thickness of the leaf; Convolutional layer: Convolutional layers are used to detect local features of the leaf, including curvature, indentation and convexity; a convolutional layer consists of multiple convolutional kernels, each of which performs a convolution operation on the input data, and the convolution operation can capture local features; Pooling layers: Pooling layers are used to reduce data dimensionality, decrease computational complexity, and make the network insensitive to changes in the position of the image; Fully connected layer: The fully connected layer takes data from the pooling layer, flattens it, and connects it to the output layer of the neural network. The fully connected layer is used to learn complex relationships between data. Output layer: The output layer generates control signals to adjust grinding parameters. The number of nodes in the output layer depends on the dimension of the control signals. Step f: The control algorithm continuously monitors sensor data and constantly adjusts grinding parameters to minimize errors until the actual blade shape approaches or reaches the target shape; Step g: Analyze the initial state of the blade and the desired final shape, and then plan the optimal grinding path; specifically, this includes the following process: g.

1. First, it is necessary to clearly define the goal and constraints of the problem. In blade grinding, the goal is to find a path to minimize wear and improve grinding efficiency. The constraints include the physical limitations of the grinding machine, the material properties of the blade, and the desired final shape. g.

2. Model the problem in mathematical form so that it can be solved using optimization algorithms. The problem is modeled in the following way: Define the path: Divide the blade surface into a grid, with each grid cell representing a possible grinding point; Define state: Each mesh cell can represent a state, which includes the current shape of the blade, the required final shape, and material property information; Define path cost: Define a cost function for movement between each state, which includes grinding cost, time cost, and energy cost; Objective function: The objective function includes minimizing the total cost to simultaneously satisfy the objectives of minimizing wear and improving grinding efficiency; g.

3. Select an optimization algorithm suitable for the problem to find the optimal path; First, define the objective function and the cost function: Suppose we want to grind a blade from its initial shape S to its final shape F, where S and F can both be represented as a set of curves or points; Define the objective function J: J = W + αT + βE; Where W is the wear cost, T is the grinding time, E is the energy required for grinding, and α and β are weighting parameters used to adjust the impact of grinding time T and grinding energy E on the final cost; then, the path planning problem is transformed into an optimization problem, that is, minimizing the objective function J, which is achieved by selecting the best path to minimize J; In the grinding path planning problem, the simulated annealing algorithm is used to search for the optimal path to meet the goals of minimizing wear and improving grinding efficiency. g.

4. Perform the optimization iteration process to find the best path. The optimization algorithm will search in the state space in an attempt to find the path that minimizes the objective function. g.

5. Once the optimal path is found, it is implemented in the actual blade grinding, which involves controlling the grinding machine to grind according to the planned path.

Images