Stone inlaying water jet cutting method, device, equipment and medium

Abstract

The invention provides a stone inlaying water jet cutting method, device, equipment and medium, and relates to the technical field of abrasive water jet machining, the method analyzes key factors in abrasive water jet machining through deep learning, such as pressure, target distance, abrasive flow, cutting speed and angle, so as to optimize machining parameters and establish a prediction model of nozzle wear. The neural network is trained through experimental data, the influence of machining parameters on the cutting quality can be accurately predicted, and the optimal parameter combination is reversely deduced according to the target cutting depth. Meanwhile, the nozzle abrasion model can predict the nozzle abrasion loss, then the radius compensation of the machining contour is guided, and the cutting precision is ensured. According to the method, the problem of contour deviation caused by nozzle abrasion in traditional water jet cutting is effectively solved, the stone inlaying machining efficiency and quality are remarkably improved, the time and material cost is reduced, and wide application prospects are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of abrasive water jet machining, and particularly relates to a stone inlay water jet cutting method, device, equipment and medium. Background Art

[0002] Currently, in the modern stone processing field, with the continuous improvement of the requirements for processing accuracy and efficiency, traditional cutting technologies are gradually difficult to meet the needs of complex shapes and high-precision inlays. Traditional cutting methods, such as sawing and flame cutting, often have a heat affected zone, resulting in problems such as material deformation and cracks, and it is difficult to achieve fine contour cutting. Especially in the stone inlay process, these disadvantages are more obvious.

[0003] The stone inlay process has extremely high requirements for cutting accuracy. It not only requires precise contour dimensions but also a smooth cutting surface to ensure the inlay effect. When traditional cutting technologies cut complex shapes, they often require multiple adjustments and corrections, which not only increases the processing time and cost but also may lead to material waste. In addition, the heat affected zone generated during the cutting process by traditional cutting methods will change the physical properties of the stone and affect its final use performance.

[0004] In recent years, abrasive water jet technology, as a new type of cold machining technology, has gradually attracted attention. This technology combines high-speed water flow with abrasive particles to form a powerful cutting force, enabling high-precision cutting without heat influence. However, although abrasive water jet technology has significant advantages in cutting accuracy and surface quality, it still faces some challenges in practical applications. For example, the nozzle will gradually wear due to the impact of abrasive particles during long-term use, resulting in a decrease in cutting accuracy. In addition, the selection of processing parameters is also crucial for cutting quality and efficiency, but currently, there is a lack of effective optimization methods to precisely adjust these parameters to adapt to different processing requirements.

[0005] Therefore, how to effectively solve the problem of nozzle wear and optimize processing parameters to further improve the application effect of abrasive water jet technology in stone inlay has become the focus of current research. This not only requires in-depth research on the nozzle wear law but also the development of a method that can quickly and accurately predict nozzle wear and optimize processing parameters, so as to achieve efficient and high-quality stone inlay processing.

[0006] In view of this, the present application is proposed. Summary of the Invention

[0007] The present invention provides a stone inlay water jet cutting method, device, equipment and medium, which can at least partially improve the above problems.

[0008] To achieve the above object, the present invention adopts the following technical solutions: A stone inlay water jet cutting method, comprising: Obtain a stone inlay pattern to be processed, preprocess the stone inlay pattern to obtain processing trajectories of multiple different parts; Call a pre-trained neural network inverse prediction model to perform inverse processing on the processing trajectories, and screen out optimized processing parameters corresponding to the processing trajectories; Use a trained nozzle wear prediction model to perform wear prediction processing on the optimized processing parameters to obtain a predicted nozzle wear amount; Perform radius compensation processing on the corresponding processing trajectories according to the predicted nozzle wear amount to obtain a final processing contour, and cut the stone slab according to the final processing contour.

[0009] The present invention also provides a stone inlay water jet cutting device, comprising: A preprocessing unit for obtaining a stone inlay pattern to be processed, preprocessing the stone inlay pattern to obtain processing trajectories of multiple different parts; An inverse unit for calling a pre-trained neural network inverse prediction model to perform inverse processing on the processing trajectories, and screening out optimized processing parameters corresponding to the processing trajectories; A prediction unit for using a trained nozzle wear prediction model to perform wear prediction processing on the optimized processing parameters to obtain a predicted nozzle wear amount; A compensation unit for performing radius compensation processing on the corresponding processing trajectories according to the predicted nozzle wear amount to obtain a final processing contour, and cutting the stone slab according to the final processing contour.

[0010] The present invention also provides a stone inlay water jet cutting device, comprising: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the stone inlay water jet cutting method described in any one of the above is implemented.

[0011] The present invention also provides a readable storage medium, comprising: a computer program stored therein, and the computer program can be executed by a processor of a device where the storage medium is located to implement the stone inlay water jet cutting method described in any one of the above.

[0012] In summary, the above-mentioned stone inlay water jet cutting method optimizes the stone inlay water jet cutting process through neural network technology, improves the cutting quality and reduces the influence of nozzle wear on the machining accuracy. This method analyzes the influence weights of machining parameters (such as pressure, target distance, abrasive flow rate, cutting speed and angle) on the cutting quality through experiments, and uses neural network to establish a prediction model to realize the reverse deduction of the optimal machining parameters according to the target cutting depth. At the same time, a nozzle wear model is established based on deep learning to predict the nozzle wear amount to guide the radius compensation of the machining contour, so as to ensure the cutting accuracy. This method not only improves the machining efficiency and quality of stone inlay, but also significantly reduces the contour deviation caused by nozzle wear, reduces the time and material costs, and has high practicality and promotion value. Brief Description of the Drawings

[0013] Figure 1 is a schematic flow chart of the stone inlay water jet cutting method provided by the first embodiment of the present invention; Figure 2 is a flow block diagram of the stone inlay water jet cutting method provided by the first embodiment of the present invention; Figure 3 is a schematic network structure diagram of the neural network predicting the effective cutting depth provided by the embodiment of the present invention; Figure 4 is a schematic diagram of the neural network reverse deduction prediction model based on the BP neural network provided by the embodiment of the present invention; Figure 5 is a schematic network structure diagram of the nozzle wear prediction model provided by the embodiment of the present invention; Figure 6 is a module schematic diagram of the stone inlay water jet cutting device provided by the second embodiment of the present invention. Detailed Embodiments

[0014] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0015] Refer to Figures 1 to 3 As shown, the first embodiment of the present invention discloses a stone inlay water jet cutting method, which can be executed by a stone inlay water jet cutting device (hereinafter referred to as the cutting device), and particularly, by one or more processors in the cutting device to implement the following method: S1, obtain the stone inlay pattern to be processed, preprocess the stone inlay pattern, and obtain machining trajectories of multiple different parts; Specifically, step S1 includes: obtaining a stone inlay pattern to be processed, splitting the stone inlay pattern into multiple different component patterns, determining the basic information of each component, classifying each component pattern according to the color in the basic information, and at the same time, screening out matching stone slabs according to the basic information; Converting each of the component patterns into a processing trajectory, wherein each part of the processing trajectory is a closed and complete pattern.

[0016] Preferably, the basic information includes the size, shape and color of the component.

[0017] In this embodiment, first, obtain the stone inlay pattern to be processed, which is the starting point of the entire processing process. After obtaining the pattern, split it into multiple different component patterns for subsequent refined processing. For each component pattern, its basic information needs to be determined, and this basic information includes the size, shape, color, etc. of the component. These information are crucial for subsequent processing parameter settings and stone slab selection. Classify each component pattern according to the color in the basic information, so as to ensure that components of the same color can adopt a consistent processing strategy during processing, improving processing efficiency and quality.

[0018] Immediately afterwards, screen out matching stone slabs according to the basic information; different colors and textures of stone slabs are suitable for different component patterns. Through precise matching, the beauty of the design pattern can be restored to the greatest extent. Then, convert each of the component patterns into a processing trajectory, and here the processing trajectory is the actual working path of the water jet cutting machine. Each part of the processing trajectory must be a closed and complete pattern, so as to ensure the integrity and accuracy of cutting, avoid incomplete cutting or redundant cutting marks, and thus improve the overall quality of stone inlay.

[0019] Through the above steps, efficient preprocessing of the stone inlay pattern is achieved, laying a solid foundation for subsequent processing parameter optimization and nozzle wear law analysis. This process not only improves processing efficiency, but also enhances the beauty and quality of the final product through precise component splitting and classification, as well as matching suitable stone slabs. At the same time, the design of a closed and complete processing trajectory effectively avoids errors during cutting, further ensuring processing accuracy.

[0020] Please refer to Figure 4 , S2, call the pre-trained neural network inverse prediction model to perform inverse processing on the processing trajectory, and screen out the optimized processing parameters corresponding to the processing trajectory; Specifically, step S2 includes: obtaining a preset target effective cutting depth as the expected output of the neural network inverse prediction model, and constructing a loss function according to the error between the current prediction output result and the target effective cutting depth; Keep the structure and weights of the neural network inverse prediction model unchanged. Take the input pressure, target distance, cutting speed, abrasive flow rate, and cutting angle as variables and initialize them to random values or empirical values. Based on the backpropagation algorithm and minimizing the loss function, update the input parameters backward until the gap between the cutting depth obtained by iteration and the expected output is less than the preset value, and take the input parameter combination corresponding to this cutting depth as the optimized processing parameters.

[0021] Preferably, the optimized processing parameters include the pressure, cutting speed, abrasive flow rate, target distance, and cutting angle of the abrasive water jet.

[0022] In this embodiment, after completing the preprocessing of the stone inlay pattern and obtaining the processing trajectories of multiple different parts, enter the key processing parameter optimization stage. Use the pre-trained neural network inverse prediction model to perform inverse processing on the processing trajectories to screen out the optimized processing parameters corresponding to the processing trajectories. This process is the core link to improve cutting quality and efficiency.

[0023] Specifically, first, a preset target effective cutting depth needs to be set. This depth is determined according to the actual processing requirements and stone characteristics and will be used as the expected output of the neural network inverse prediction model. Then, construct a loss function based on the error between the current prediction output result and the target effective cutting depth. The construction of the loss function is the key to neural network learning. It quantifies the gap between the model prediction value and the actual expected value and provides a direction for subsequent parameter optimization.

[0024] Before parameter optimization, keep the structure and weights of the neural network inverse prediction model unchanged to ensure the stability and reliability of the model. Take the input pressure, target distance, cutting speed, abrasive flow rate, and cutting angle as variables and initialize them to random values or empirical values. These parameters are the key factors affecting the water jet cutting effect. By reasonable initialization, the convergence speed of the model can be accelerated and the optimization efficiency can be improved.

[0025] Next, based on the backpropagation algorithm and the principle of minimizing the loss function, update the input parameters. The backpropagation algorithm is an efficient optimization algorithm. It calculates the gradient of the loss function with respect to each parameter and updates the weights in the network backward, thereby gradually reducing the gap between the predicted value and the expected value. In each iteration, the model predicts the cutting depth according to the current parameter combination and calculates the error with the target depth. By continuously adjusting the parameters, when the gap between the cutting depth obtained by iteration and the expected output is less than the preset value, it is considered that the model has found a set of relatively ideal processing parameters. The input parameter combination corresponding to this set of parameters is the optimized processing parameters, including the pressure, cutting speed, abrasive flow rate, target distance, cutting angle, etc. of the abrasive water jet.

[0026] The beneficial effects of this process are remarkable. Through the neural network inverse prediction model, the optimized processing parameters corresponding to the processing trajectory can be quickly and accurately screened, greatly reducing the time cost of relying on experience and repeated trials in traditional methods. At the same time, the optimized processing parameters can ensure the stability and consistency of the cutting process, improve the cutting quality, and reduce cutting defects caused by improper parameters, such as surface roughness and dimensional deviation. In addition, this method can flexibly adjust parameters according to the characteristics and processing requirements of different stones, with strong adaptability and versatility.

[0027] Preferably, before calling the pre-trained neural network inverse prediction model to perform inverse processing on the processing trajectory, it further includes: Obtain abrasive water jet experimental data, perform denoising processing and normalization processing on the abrasive water jet experimental data, and divide it into a training set and a test set. Among them, the abrasive water jet experimental data includes effective cutting depth, pressure, target distance, cutting speed, abrasive flow rate, and cutting angle; Construct a forward neural network model, where the forward neural network model includes an input layer, one or two hidden layers, and an output layer. The input of the forward neural network model is pressure, target distance, cutting speed, abrasive flow rate, and cutting angle, and the output of the forward neural network model is the effective cutting depth; According to the training set, use the ReLU activation function, loss function, and optimizer to train the forward neural network model, and predict the forward neural network model according to the test set until the output result reaches a preset value to obtain the trained neural network inverse prediction model, where the loss function is the mean square error MSE, and the optimizer is the Adam optimizer.

[0028] In this embodiment, abrasive water jet experimental data is obtained. These data are obtained through actual water jet cutting experiments and cover multiple key parameters such as effective cutting depth, pressure, target distance, cutting speed, abrasive flow rate, and cutting angle. These experimental data are the basis for model training, and their quality and integrity directly affect the accuracy and reliability of the model. To improve the quality of the data, denoising processing and normalization processing are performed on these data. Denoising processing can remove outliers and interference information in the data, making the data more pure and stable; normalization processing is to scale the data to a unified range, usually between 0 and 1, which can accelerate the convergence speed of the model and improve the training efficiency. The processed data is divided into a training set and a test set. The training set is used for model training, and the test set is used to verify the performance of the model.

[0029] Subsequently, a forward neural network model is constructed. The model includes an input layer, one or two hidden layers, and an output layer. The input layer receives five key parameters: pressure, target distance, cutting speed, abrasive flow rate, and cutting angle; the output layer outputs the effective cutting depth. The design of this structure enables the model to learn the complex non-linear relationship between the input parameters and the output cutting depth. During the construction of the model, the ReLU activation function is selected, which can effectively solve the problem of gradient disappearance and enable the model to converge faster during the training process. At the same time, the mean square error (MSE) is used as the loss function, which quantifies the difference between the model's predicted value and the actual value and provides a clear goal for model optimization. The Adam optimizer is selected as the optimizer, which combines the advantages of multiple optimization algorithms, can adaptively adjust the learning rate, and further improve the training efficiency and stability of the model.

[0030] Furthermore, based on the training set, the forward neural network model is trained. During the training process, the model calculates the output result through forward propagation, and then calculates the error between the predicted value and the actual value according to the loss function. Through the backpropagation algorithm, the weights are updated using gradient descent or the Adam optimization algorithm to gradually reduce the error. During the training process, the number of training epochs is set and the change in loss is monitored. When the loss no longer decreases significantly, it is considered that the model has converged. Then, the trained model is used to predict using the test set to verify the performance of the model. If the output result reaches the preset value, that is, the prediction accuracy of the model meets the requirements, a trained neural network inverse prediction model is obtained.

[0031] Briefly speaking, by denoising and normalizing the experimental data, the quality of the data and the training efficiency of the model are improved. The constructed forward neural network model can accurately learn the relationship between the input parameters and the output cutting depth, providing a reliable basis for subsequent parameter inversion. By using the ReLU activation function, the MSE loss function, and the Adam optimizer, the model can converge quickly and achieve a high prediction accuracy. The finally obtained trained neural network inverse prediction model can not only accurately predict the cutting depth, but also invert the optimal processing parameters according to the target cutting depth, greatly improving the quality and efficiency of stone inlaid water jet cutting.

[0032] Please refer to Figure 5 , S3, use the trained nozzle wear prediction model to perform wear prediction processing on the optimized processing parameters to obtain the predicted nozzle wear amount; Preferably, before using the trained nozzle wear prediction model to perform wear prediction processing on the optimized machining parameters, it further includes: obtaining second abrasive water jet experimental data, performing normalization processing on the second abrasive water jet experimental data, and dividing it into a training set, a validation set, and a test set. The second abrasive water jet experimental data includes abrasive flow rate, injection pressure, cutting angle, usage time, and the corresponding nozzle wear amount; Construct a nozzle wear prediction model. Set the input layer of the nozzle wear prediction model to include 4 neurons, corresponding to abrasive flow rate, pressure, cutting angle, and usage time respectively. Set the output layer of the nozzle wear prediction model to include 1 neuron, corresponding to the nozzle wear amount. Among them, the activation function of the hidden layer of the nozzle wear prediction model is the ReLU function, and the activation function of the output layer of the nozzle wear prediction model uses a linear activation function; Based on the training set and the validation set, use random values to initialize the weights of the nozzle wear prediction model, calculate the loss function according to the optimized machining parameters, and use the mean square error MSE as the optimization objective to perform backpropagation, and update the weights of the nozzle wear prediction model using the gradient descent or Adam optimization algorithm; Set the number of training epochs of the nozzle wear prediction model and monitor the loss change during the training process. Calculate the error index of the test set according to the nozzle wear prediction model. When it is judged that the error index reaches the threshold, end the training.

[0033] Specifically, in this embodiment, before using the nozzle wear prediction model to perform wear prediction, a series of data processing and model construction work needs to be carried out. First, obtain the second abrasive water jet experimental data, which includes abrasive flow rate, injection pressure, cutting angle, usage time, and the corresponding nozzle wear amount. These data reflect the wear situation of the nozzle under different processing conditions and are the basis for constructing the nozzle wear prediction model. Perform normalization processing on these data to eliminate the influence of different dimensions and numerical ranges on model training and improve the stability and convergence speed of the model. Then, divide the processed data into a training set, a validation set, and a test set, where the training set is used for model training, the validation set is used to adjust the hyperparameters of the model, and the test set is used to evaluate the performance of the model.

[0034] Secondly, a nozzle wear prediction model is constructed. The input layer of this model is set to contain 4 neurons, corresponding to abrasive flow rate, pressure, cutting angle, and usage time respectively. These input parameters can comprehensively reflect the working state of the nozzle during the machining process. The output layer is set to contain 1 neuron, corresponding to the nozzle wear amount, directly giving the wear prediction value of the nozzle under the current machining conditions. In the model structure, the ReLU function is selected as the activation function for the hidden layer, which can effectively solve the problem of gradient disappearance and enable the model to converge faster during the training process. The linear activation function is adopted for the output layer to ensure that the output value can accurately reflect the actual value of the nozzle wear amount.

[0035] Based on the training set and the validation set, initialize the weights of the nozzle wear prediction model with random values. Calculate the loss function according to the optimized machining parameters, and adopt the mean square error (MSE) as the optimization objective. MSE can quantify the difference between the model prediction value and the actual value, providing a clear direction for model optimization. Through backpropagation, use the gradient descent or Adam optimization algorithm to update the weights of the nozzle wear prediction model. During the training process, set the number of training epochs of the nozzle wear prediction model and closely monitor the change of loss during the training process. When the loss no longer decreases significantly, it is considered that the model has converged. Then, calculate the error metrics of the test set according to the nozzle wear prediction model, such as mean square error (MSE), mean absolute error (MAE), and coefficient of determination (R 2 ). When it is judged that the error metrics reach the preset threshold, end the training to obtain the trained nozzle wear prediction model.

[0036] This step provides high-quality data support for model training by normalizing the second abrasive water jet experimental data and reasonably dividing the data set, improving the accuracy and reliability of the model. The constructed nozzle wear prediction model can accurately predict the wear situation of the nozzle under different machining conditions, providing a scientific basis for subsequent machining compensation. By using the ReLU activation function and the linear activation function, and adopting MSE as the optimization objective and the Adam optimization algorithm, the model can converge quickly and reach a high prediction accuracy. The finally obtained trained nozzle wear prediction model can not only provide accurate wear prediction for the machining process, but also help optimize the nozzle maintenance plan, extend the service life of the nozzle, reduce production costs, and improve production efficiency.

[0037] S4. Perform radius compensation processing on the corresponding machining trajectory according to the predicted nozzle wear amount to obtain the final machining profile, and cut the slate according to the final machining profile.

[0038] Specifically, in this embodiment, the core of this step lies in performing radius compensation processing on the machining trajectory according to the predicted nozzle wear amount, so as to obtain the final machining profile, and cutting the slate accordingly. Perform radius compensation processing on each machining profile according to the predicted nozzle wear amount obtained from the nozzle wear prediction model. Nozzle wear will cause a deviation between the actual cutting trajectory and the designed trajectory. By predicting the nozzle wear amount, the machining trajectory can be adjusted in advance to compensate for the deviation caused by nozzle wear. This compensation processing can ensure that the size and shape of the cut stone components meet the design requirements, thereby improving the overall quality of stone inlay.

[0039] Immediately afterwards, when performing radius compensation, it is necessary to comprehensively consider the nozzle wear amount, machining parameters, and geometric characteristics of the machining trajectory. By accurately calculating the compensation radius, the compensated contour line is used as the final machining profile. This process requires not only an accurate nozzle wear prediction model but also an efficient compensation algorithm to ensure that the compensated machining profile can meet the requirements of high-precision machining. After obtaining the final machining profile, cut the slate according to this profile. During the cutting process, using the optimized machining parameters and the compensated machining profile, the water jet cutting machine can accurately complete the cutting task. Since the machining trajectory has been compensated according to nozzle wear, the size and shape accuracy of the cut stone components are higher, and they can better meet the requirements of stone inlay.

[0040] It can effectively solve the influence of nozzle wear on machining accuracy and ensure that the size and shape of the cut stone components meet the design requirements; it can improve the overall quality of stone inlay, reduce the splicing error caused by nozzle wear, and thus enhance the aesthetics and practicality of the product.

[0041] In summary, the water jet cutting method for stone inlay aims to significantly improve the cutting accuracy and machining efficiency of stone inlay through intelligent machining parameter optimization and nozzle wear prediction, while reducing production costs and material waste. First, the method obtains the stone inlay pattern to be processed, preprocesses it, splits the pattern into multiple different parts, determines the basic information of each component, such as size, shape, and color, and classifies the components according to color. This process not only improves the flexibility of machining but also ensures high-quality machining by accurately matching the slate. The preprocessed pattern is converted into a closed and complete machining trajectory, providing an accurate path for subsequent cutting.

[0042] Secondly, in the processing parameter optimization stage, a pre-trained neural network inverse prediction model is used to inversely deduce the optimal processing parameters according to the target cutting depth, including pressure, target distance, cutting speed, abrasive flow rate, and cutting angle. Through the denoising and normalization of experimental data, and the training of the forward neural network model, the model can quickly and accurately predict the impact of processing parameters on cutting quality. This process not only improves the efficiency of parameter optimization but also reduces the time cost of relying on experience and repeated trials in traditional methods.

[0043] Furthermore, through the nozzle wear prediction model, the wear of the optimized processing parameters is predicted. By obtaining the second abrasive water jet experimental data, normalizing it, and training the model, the nozzle wear prediction model can accurately predict the wear of the nozzle under different processing conditions. This prediction function not only provides a scientific basis for processing compensation but also helps optimize the nozzle maintenance plan, extend the service life of the nozzle, and reduce production costs. Finally, according to the predicted nozzle wear amount, radius compensation is performed on the processing trajectory to obtain the final processing profile, and the slate is cut accordingly. This compensation process can effectively solve the impact of nozzle wear on processing accuracy, ensure that the size and shape of the cut stone components meet the design requirements, and thus improve the overall quality of stone inlay.

[0044] Briefly speaking, the stone inlay water jet cutting method achieves high precision and high efficiency in stone inlay water jet cutting through an intelligent neural network model and an accurate compensation algorithm. It not only improves the processing quality, reduces the splicing error caused by nozzle wear, but also reduces production costs and material waste, with significant economic benefits and broad application prospects.

[0045] Please refer to Figure 6 , the second embodiment of the present invention provides a stone inlay water jet cutting device, which includes: A pretreatment unit 101 for obtaining the stone inlay pattern to be processed, preprocessing the stone inlay pattern, and obtaining processing trajectories of multiple different parts; An inverse deduction unit 102 for calling a pre-trained neural network inverse prediction model to perform inverse deduction processing on the processing trajectory and screening out the optimized processing parameters corresponding to the processing trajectory; A prediction unit 103 for using the trained nozzle wear prediction model to perform wear prediction processing on the optimized processing parameters to obtain the predicted nozzle wear amount; A compensation unit 104 for performing radius compensation processing on the corresponding processing trajectory according to the predicted nozzle wear amount to obtain the final processing profile, and cutting the slate according to the final processing profile.

[0046] The third embodiment of the present invention provides a stone inlay water jet cutting device, which includes: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the stone inlay water jet cutting method described in any one of the above is implemented.

[0047] The fourth embodiment of the present invention provides a readable storage medium, which includes: a computer program stored therein, and the computer program can be executed by the processor of the device where the storage medium is located to implement the stone inlay water jet cutting method described in any one of the above.

[0048] Exemplarily, each of the above devices and each process step can be implemented by a computer program. The computer program can be divided into one or more units. The one or more units are stored in the memory and executed by the processor to complete the present invention.

[0049] The so-called processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0050] The memory can be used to store the computer program and / or module. The processor realizes various functions of the present invention by running or executing the computer program and / or module stored in the memory, and by calling the data stored in the memory. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one magnetic disk storage device, a flash device, or other volatile solid-state storage devices.

[0051] Among them, if the unit integrated in the electronic device or printer is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-described embodiment methods of the present invention, it can also be completed by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-described various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice within the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0052] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the attached drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that they have a communication connection, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0053] The above is the preferred implementation manner of the present invention. It should be pointed out that for those of ordinary skill in the art in the technical field, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A stone inlay water jet cutting method, characterized in that, Including: Obtain a stone inlay pattern to be processed, preprocess the stone inlay pattern to obtain processing trajectories of multiple different parts; Call a pre-trained neural network inverse prediction model to perform inverse processing on the processing trajectory, and screen out optimized processing parameters corresponding to the processing trajectory; Use a trained nozzle wear prediction model to perform wear prediction processing on the optimized processing parameters to obtain a predicted nozzle wear amount; Perform radius compensation processing on the corresponding processing trajectory according to the predicted nozzle wear amount to obtain a final processing profile, and cut the stone slab according to the final processing profile.

2. The stone inlay water jet cutting method according to claim 1, characterized in that Obtain a stone inlay pattern to be processed, preprocess the stone inlay pattern to obtain processing trajectories of multiple different parts, specifically: Obtain a stone inlay pattern to be processed, split the stone inlay pattern into multiple different component patterns, determine the basic information of each component, classify each component pattern according to the color in the basic information, and at the same time, screen out matching stone slabs according to the basic information; Convert each component pattern into a processing trajectory, where each part of the processing trajectory is a closed and complete pattern.

3. The stone inlay water jet cutting method according to claim 2, characterized in that The basic information includes the size, shape and color of the component.

4. The stone inlay water jet cutting method according to claim 1, characterized in that Call a pre-trained neural network inverse prediction model to perform inverse processing on the processing trajectory, and screen out optimized processing parameters corresponding to the processing trajectory, specifically: Obtain a preset target effective cutting depth as the expected output of the neural network inverse prediction model, and construct a loss function according to the error between the current predicted output result and the target effective cutting depth; Keep the structure and weights of the neural network inverse prediction model unchanged, take the input pressure, target distance, cutting speed, abrasive flow rate and cutting angle as variables, and initialize them to random values or empirical values; Based on the backpropagation algorithm and minimizing the loss function, update the input parameters in reverse until the gap between the cutting depth obtained by iteration and the expected output is less than a preset value, and take the input parameter combination corresponding to this cutting depth as the optimized processing parameters.

5. The stone inlay water jet cutting method according to claim 4, characterized in that, The optimized processing parameters include the pressure, cutting speed, abrasive flow rate, target distance, and cutting angle of the abrasive water jet.

6. The stone inlay water jet cutting method according to claim 4, characterized in that Before calling a pre-trained neural network inverse prediction model to perform inverse processing on the processing trajectory, it also includes: Obtain abrasive water jet experimental data, perform denoising processing and normalization processing on the abrasive water jet experimental data, and divide it into a training set and a test set, where the abrasive water jet experimental data includes effective cutting depth, pressure, target distance, cutting speed, abrasive flow rate, cutting angle; Construct a forward neural network model, where the forward neural network model includes an input layer, one or two hidden layers and an output layer, the input of the forward neural network model is pressure, target distance, cutting speed, abrasive flow rate, cutting angle, and the output of the forward neural network model is the effective cutting depth; According to the training set, the forward neural network model is trained using the ReLU activation function, loss function, and optimizer, and the forward neural network model is predicted according to the test set until the output result reaches the preset value, obtaining a trained neural network inverse prediction model, where the loss function is the mean square error MSE, and the optimizer is the Adam optimizer.

7. The stone inlay water jet cutting method according to claim 1, characterized in that, Before using the trained nozzle wear prediction model to perform wear prediction processing on the optimized processing parameters, it further includes: Obtain the second abrasive water jet experimental data, perform normalization processing on the second abrasive water jet experimental data, and divide it into a training set, a validation set, and a test set. Among them, the second abrasive water jet experimental data includes abrasive flow rate, jet pressure, cutting angle, usage time, and the corresponding nozzle wear amount; Construct a nozzle wear prediction model. Set the input layer of the nozzle wear prediction model to include 4 neurons, corresponding to the abrasive flow rate, pressure, cutting angle, and usage time respectively. Set the output layer of the nozzle wear prediction model to include 1 neuron, corresponding to the nozzle wear amount. Among them, the activation function of the hidden layer of the nozzle wear prediction model is the ReLU function, and the activation function of the output layer of the nozzle wear prediction model uses a linear activation function; Based on the training set and the validation set, use random values to initialize the weights of the nozzle wear prediction model, calculate the loss function according to the optimized processing parameters, and use the mean square error MSE as the optimization objective to perform backpropagation, and update the weights of the nozzle wear prediction model using the gradient descent or Adam optimization algorithm; Set the number of training epochs of the nozzle wear prediction model and monitor the loss change during training. Calculate the error index of the test set according to the nozzle wear prediction model. When it is judged that the error index reaches the threshold, end the training.

8. A stone inlaid water jet cutting device, characterized in that, It includes: A preprocessing unit for obtaining the stone inlay pattern to be processed, preprocessing the stone inlay pattern, and obtaining the processing trajectories of multiple different parts; An inverse inference unit for calling the pre-trained neural network inverse prediction model to perform inverse inference processing on the processing trajectory and screening out the optimized processing parameters corresponding to the processing trajectory; A prediction unit for using the trained nozzle wear prediction model to perform wear prediction processing on the optimized processing parameters to obtain the predicted nozzle wear amount; A compensation unit for performing radius compensation processing on the corresponding processing trajectory according to the predicted nozzle wear amount to obtain the final processing contour, and cutting the stone slab according to the final processing contour.

9. A stone inlay water jet cutting device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the stone inlay water jet cutting method according to any one of claims 1 to 7.

10. A readable storage medium, characterized in that, A computer program is stored, and the computer program can be executed by the processor of the device where the storage medium is located to implement the stone inlay water jet cutting method according to any one of claims 1 to 7.

Images