Machine tool stability prediction method of side-hanging type machine tool body based on double-triangle structure
By using a double triangle structure and neural network model to predict stability on the hanging bed, the problem of insufficient stability of the hanging bed structure is solved, and higher prediction accuracy and stability are achieved, and the structure is adjusted according to the needs to meet the needs. Different working conditions.
Patent Information
- Application Number
- CN202410962049.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-08
Smart Images

Figure CN118916655A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of machine tools, in particular to a method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure. Background Art
[0002] With the continuous development of laser cutting technology, the size of laser cutting plates has gradually increased, and the size of laser cutting equipment has increased. Therefore, the requirements for the bed of the laser cutting machine to bear the laser cutting equipment have also increased. Therefore, the requirements for the stability of the bed are also increasing, especially for the bed of the hanging machine tool.
[0003] However, the existing hanging bed has the problem of poor stability. For example, patent CN219094300U discloses a side-hanging bed structure, including a bracket, a movable guide mechanism and a plurality of reinforcing ribs, wherein the bracket includes a middle rib, an upper rib and a lower rib, and the reinforcing ribs include a transverse reinforcing rib and a longitudinal reinforcing rib; the middle rib and the lower rib are connected by a plurality of transverse reinforcing ribs; the middle rib and the upper rib are connected by a plurality of longitudinal reinforcing ribs and middle reinforcing ribs are installed between each longitudinal reinforcing rib; each transverse reinforcing rib and longitudinal reinforcing rib are arranged in one-to-one correspondence with each other; the movable guide mechanism is composed of two groups of guide components, which are respectively installed on the upper rib and the middle reinforcing rib; an inclined frame is installed between each transverse reinforcing rib and the longitudinal reinforcing rib, and a support plate is arranged on the inclined frame.
[0004] The patent achieves the stability of the bed by setting reinforcing ribs, but the above patent cannot achieve the stability of the hanging bed connection, resulting in unstable installation. Secondly, the above patent also cannot predict the stability of the machine tool. Summary of the invention
[0005] In view of the shortcomings of the prior art, the present invention provides a method for predicting the stability of a side-mounted bed based on a double triangle structure. The present invention increases the stability of the hanging bed through the double triangle structure. At the same time, according to the structural weight and installation position of the slide and laser cutting equipment added to the lathe under different working conditions, the stability of the bed is predicted, and the corresponding number of double triangle structures are set according to the prediction results.
[0006] The technical solution of the present invention is: a method for predicting thermal deformation of a machine tool based on a side-mounted bed with a double triangle structure, comprising the following steps:
[0007] S1) Determine the number of double triangle structures according to the weight of the slide and laser cutting equipment installed on the machine tool;
[0008] S2), arranging a plurality of sensors on the side-mounted bed of the double triangle structure, and acquiring a machine tool data set through the sensors;
[0009] S3), preprocessing the machine tool data set collected in step S2); then performing feature extraction on the preprocessed data set;
[0010] S4), screening the extracted features and then fusing them to obtain a machine tool data set after feature fusion;
[0011] S5), dividing the machine tool data set after feature fusion into a training set and a test set;
[0012] S6) Train the constructed neural network model through the training set, and use the test set to predict and verify the trained neural network model.
[0013] Preferably, in step S2), the sensors include temperature sensors, acceleration sensors and thermal deformation sensors, and the machine tool data set includes motion state data and thermal deformation state data, wherein the motion state data includes machine tool spindle temperature data and vibration data, and the temperature sensor and acceleration sensor are used to collect the machine tool spindle temperature data and vibration data, and the thermal deformation sensor is used to collect the spindle thermal deformation state data.
[0014] Preferably, in step S3), the machine tool data set preprocessing comprises the following steps:
[0015] S31), performing time registration on the acquired data set, extracting state information according to the time axis, dividing the total operation time of the system into m segments of machine tool states, and using the data set in each state as an initial sample;
[0016] S32), using the average value filtering algorithm to clean the temperature and thermal deformation data in each sample data;
[0017] S33), the vibration signal data in each initial sample is processed by using the EMD decomposition and wavelet threshold joint denoising method.
[0018] Preferably, step S4) comprises the following steps:
[0019] S41), arranging the temperature features, vibration features and thermal deformation features extracted in step S3) according to the samples to obtain a temperature feature matrix, a vibration feature matrix and a thermal deformation feature matrix respectively;
[0020] S42), based on the correlation analysis, the vibration features are screened, and the vibration features with correlation coefficients less than a preset threshold are eliminated; then the vibration dimension reduction matrix after elimination is fused with the temperature feature matrix according to the time axis to obtain a fused information feature matrix.
[0021] Preferably, in step S6), the neural network model takes the motion state data and thermal deformation state data of the machine tool before the current moment t as input, and takes the thermal deformation of the spindle within a time period τ after the current moment t as output.
[0022] Preferably, in step S6), the neural network model includes a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a hidden layer and a fully connected layer which are arranged in sequence.
[0023] Preferably, in step S6), the training of the neural network model is specifically as follows: inputting the data in the training set into the constructed neural network model and calculating the error, updating the parameters in the neural network model according to the calculated error, and then continuing to input data according to the updated neural network model for error calculation, and repeating this cycle until the prediction accuracy of the neural network model for the data reaches the preset requirements, and the model training is completed.
[0024] Preferably, the side-mounted bed with a double-triangular structure includes a bed body, and a plurality of lower triangular connecting structures are arranged at the lower end of the bed body; and a plurality of upper triangular connecting structures are also arranged at the upper end of the bed body on the other side of the lower triangular connecting structure.
[0025] Preferably, the lower triangular connection structure and the upper triangular connection structure are both arranged along the length direction of the bed body.
[0026] Preferably, the lower triangular connection structure is the same as the upper triangular connection structure, and both include a first connection plate, a second connection plate, and a third connection plate, and the first connection plate, the second connection plate, and the third connection plate are connected by welding or riveting.
[0027] Preferably, the first connecting plate and the second connecting plate are both U-shaped steel plates, and a fixing piece is provided at the connection between the first connecting plate and the second connecting plate. The fixing piece is provided with a plurality of connecting screw holes, and the connection between the first connecting plate and the second connecting plate is achieved by cooperating with corresponding bolts through the connecting screw holes.
[0028] Preferably, each end of the lower triangular connecting structure and the upper triangular connecting structure is further provided with a corresponding connecting lug, and the connecting lug is provided with a corresponding connecting hole.
[0029] Preferably, the bed body is a three-layer steel plate structure, and the layers of the three-layer steel plate structure are connected by bolts.
[0030] Preferably, a side wall guide rail is provided on the side wall of the bed body at the upper end of the lower triangular connecting structure, and an upper guide rail is provided on the upper end of the bed body.
[0031] The beneficial effects of the present invention are:
[0032] 1. The present invention integrates the vibration information and temperature information of various parts of the machine tool, which can improve the reliability and stability of the model and effectively improve the prediction accuracy;
[0033] 2. The present invention increases the stability of the bed through the double triangle structure, and the bed structure is compact, which is convenient for workshop layout;
[0034] 3. The present invention can adjust the number of upper triangular connecting structures and lower triangular connecting structures according to demand, so as to arrange them accordingly according to stability requirements, thereby further improving the stability of the bed. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Figure 1 A schematic diagram of a process of an embodiment of the present invention;
[0036] Figure 2 It is a structural schematic diagram of the bed of the present invention;
[0037] Figure 3 It is a schematic diagram of the installation of the lower triangle connection structure of the present invention;
[0038] In the figure, 1-bed body; 2-lower triangular connection structure; 3-upper triangular connection structure; 4-first connection plate; 5-second connection plate; 6-third connection plate; 7-connection ear; 8-side wall guide rail; 9-upper guide rail. DETAILED DESCRIPTION
[0039] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings:
[0040] Example 1
[0041] This embodiment provides a method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure, comprising the following steps:
[0042] S1) Determine the number of double triangle structures according to the weight of the slide and laser cutting equipment installed on the machine tool;
[0043] S2), arranging a plurality of sensors on the side-mounted bed 1 of the double triangle structure, and acquiring a machine tool data set through the sensors;
[0044] S3), preprocessing the machine tool data set collected in step S2); then performing feature extraction on the preprocessed data set;
[0045] S4), screening the extracted features and then fusing them to obtain a machine tool data set after feature fusion;
[0046] S5), dividing the machine tool data set after feature fusion into a training set and a test set;
[0047] S6) Train the constructed neural network model through the training set, and use the test set to predict and verify the trained neural network model.
[0048] As a preferred embodiment of the present invention, in step S2), the sensors include a temperature sensor, an acceleration sensor and a thermal deformation sensor, and the machine tool data set includes motion state data and thermal deformation state data, wherein the motion state data includes machine tool spindle temperature data and vibration data, and the temperature sensor and acceleration sensor are used to collect the machine tool spindle temperature data and vibration data, and the thermal deformation sensor is used to collect the spindle thermal deformation state data.
[0049] As a preferred embodiment of this invention, in step S3), the machine tool data set preprocessing includes the following steps:
[0050] S31), performing time registration on the acquired data set, extracting state information according to the time axis, dividing the total operation time of the system into m segments of machine tool states, and using the data set in each state as an initial sample;
[0051] S32), using the average value filtering algorithm to clean the temperature and thermal deformation data in each sample data;
[0052] S33), the vibration signal data in each initial sample is processed by using the EMD decomposition and wavelet threshold joint denoising method.
[0053] As a preferred embodiment of this invention, step S4) includes the following steps:
[0054] S41), arranging the temperature features, vibration features and thermal deformation features extracted in step S3) according to the samples to obtain a temperature feature matrix, a vibration feature matrix and a thermal deformation feature matrix respectively;
[0055] S42), based on the correlation analysis, the vibration features are screened, and the vibration features with correlation coefficients less than a preset threshold are eliminated; then the vibration dimension reduction matrix after elimination is fused with the temperature feature matrix according to the time axis to obtain a fused information feature matrix.
[0056] As a preferred embodiment of the present invention, in step S6), the neural network model uses the motion state data and thermal deformation state data of the machine tool before the current moment t as input, and uses the thermal deformation of the spindle within a time period τ after the current moment t as output.
[0057] As a preferred embodiment of this embodiment, in step S6), the neural network model includes a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a hidden layer and a fully connected layer which are arranged in sequence.
[0058] As a preferred embodiment of the present invention, in step S6), the training of the neural network model is specifically as follows: the neural network model is trained using a gradient descent algorithm, the optimizer selects an Adam optimizer with an adaptive learning rate, and the first-order moment estimation and second-order moment estimation of the Adam optimizer gradient are used to dynamically adjust the learning rate of each parameter, the total number of training rounds is M, and during the training process, the back propagation algorithm is used to obtain the gradient to update the weights and biases of the model, and the update formula is:
[0059]
[0060] Among them, w represents weight, b represents bias, and α is the learning rate; It is the process of finding the partial derivative of the loss on the corresponding parameters.
[0061] Example 2
[0062] like Figure 2 and 3 As shown, this embodiment provides a side-mounted bed based on a double triangle structure, including a bed body 1, wherein a plurality of lower triangle connecting structures 2 are arranged at the lower end of the bed body 1; and a plurality of upper triangle connecting structures 3 are also arranged at the upper end of the bed body 1 on the other side relative to the lower triangle connecting structure 2.
[0063] As a preferred embodiment of the present invention, the lower triangular connection structure 2 and the upper triangular connection structure 3 are both arranged along the length direction of the bed body 1 .
[0064] As a preferred embodiment of the present invention, the lower triangular connection structure 2 is the same as the upper triangular connection structure 3, and both include a first connection plate 4, a second connection plate 5, and a third connection plate 6. The first connection plate 4, the second connection plate 5, and the third connection plate 6 are connected by welding or riveting.
[0065] As a preferred embodiment of the present invention, the first connecting plate 4 and the second connecting plate 5 are both U-shaped steel plates, and a fixing piece is further provided at the connection between the first connecting plate 4 and the second connecting plate 5. The fixing piece is provided with a plurality of connecting screw holes, and the connection between the first connecting plate 4 and the second connecting plate 5 is achieved by cooperating with corresponding bolts through the connecting screw holes.
[0066] As a preferred embodiment of the present invention, each end of the lower triangular connection structure 2 and the upper triangular connection structure 3 is also provided with a corresponding connection lug 7, and the connection lug 7 is provided with a corresponding connection hole. In the present embodiment, the lower triangular connection structure 2 is connected to the ground through the connection lug 7; the upper triangular connection structure 3 is connected to the upper end surface of the bed body 1 through the connection lug 7; and the third connection plate 6 of the upper triangular connection structure 3 is connected to the mounting processing equipment. In the present embodiment, a corresponding number of lower triangular connection structures 2 and upper triangular connection structures 3 are set according to stability requirements.
[0067] As a preferred embodiment of the present invention, the bed body 1 is a three-layer steel plate structure, and the layers of the three-layer steel plate structure are connected by bolts.
[0068] As a preferred embodiment of the present invention, a side wall guide rail 8 is further provided on the side wall of the bed body 1 at the upper end of the lower triangular connecting structure 2 .
[0069] As a preferred embodiment of the present invention, an upper guide rail 9 is further provided at the upper end of the bed body 1 .
[0070] The above embodiments and descriptions are only for illustrating the principles and best embodiments of the present invention. Without departing from the spirit and scope of the present invention, the present invention may be subject to various changes and improvements, all of which fall within the scope of the present invention to be protected.
Claims
1. A method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure, characterized in that: The steps include: S1) Determine the number of double triangle structures according to the weight of the slide and laser cutting equipment installed on the machine tool; S2), arranging a plurality of sensors on the side-mounted bed with a double triangle structure, and acquiring a machine tool data set through the plurality of sensors; S3), preprocessing the machine tool data set collected in step S2); then performing feature extraction on the preprocessed data set; S4), screening the extracted features and then fusing them to obtain a machine tool data set after feature fusion; S5), dividing the machine tool data set after feature fusion into a training set and a test set; S6) Train the constructed neural network model through the training set, and use the test set to predict and verify the trained neural network model.
2. The method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure according to claim 1, characterized in that: In step S2), the sensors include temperature sensors, acceleration sensors and thermal deformation sensors, and the machine tool data set includes motion state data and thermal deformation state data, wherein the motion state data includes machine tool spindle temperature data and vibration data. The temperature sensor and acceleration sensor are used to collect the machine tool spindle temperature data and vibration data, and the thermal deformation sensor is used to collect the spindle thermal deformation state data.
3. The method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure according to claim 1 is characterized in that: In step S3), the machine tool data set preprocessing includes the following steps: S31), performing time registration on the acquired data set, extracting state information according to the time axis, dividing the total operation time of the system into m segments of machine tool states, and using the data set in each state as an initial sample; S32), using the average value filtering algorithm to clean the temperature and thermal deformation data in each sample data; S33), the vibration signal data in each initial sample is processed by using the EMD decomposition and wavelet threshold joint denoising method.
4. The method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure according to claim 1, characterized in that: Step S4) specifically comprises the following steps: S41), arranging the temperature features, vibration features and thermal deformation features extracted in step S3) according to the samples to obtain a temperature feature matrix, a vibration feature matrix and a thermal deformation feature matrix respectively; S42), based on the correlation analysis, the vibration features are screened, and the vibration features with correlation coefficients less than a preset threshold are eliminated; then the vibration dimension reduction matrix after elimination is fused with the temperature feature matrix according to the time axis to obtain a fused information feature matrix.
5. The method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure according to claim 1, characterized in that: In step S6), the neural network model takes the motion state data and thermal deformation state data of the machine tool before the current moment t as input, and takes the thermal deformation of the spindle within a time period τ after the current moment t as output.
6. The method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure according to claim 5, characterized in that: In step S6), the neural network model includes a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a hidden layer and a fully connected layer arranged in sequence; The training of the neural network model is specifically as follows: the neural network model is trained using a gradient descent algorithm, the optimizer selects an Adam optimizer with an adaptive learning rate, and the first-order moment estimation and second-order moment estimation of the Adam optimizer gradient are used to dynamically adjust the learning rate of each parameter. The total number of training rounds is M. During the training process, the back propagation algorithm is used to obtain the gradient to update the weights and biases of the model. The update formula is: Among them, w represents weight, b represents bias, and α is the learning rate; It is the process of finding the partial derivative of the loss on the corresponding parameters.
7. The method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure according to claim 1, characterized in that: The side-mounted bed with a double-triangular structure comprises a bed body (1), wherein a plurality of lower triangular connecting structures (2) are arranged at the lower end of the bed body (1); and a plurality of upper triangular connecting structures (3) are also arranged at the upper end of the bed body (1) on the other side of the lower triangular connecting structures (2).
8. The method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure according to claim 7, characterized in that: The lower triangular connection structure (2) and the upper triangular connection structure (3) are both arranged along the length direction of the bed body (1); each end of the lower triangular connection structure (2) and the upper triangular connection structure (3) is also provided with a corresponding connection lug (7), and the connection lug (7) is provided with a corresponding connection hole.
9. The thermal prediction method for stability of a machine tool based on a double triangle structure side-mounted bed according to claim 8, characterized in that: The lower triangular connection structure (2) and the upper triangular connection structure (3) are the same and both include a first connection plate (4), a second connection plate (5) and a third connection plate (6); the first connection plate (4), the second connection plate (5) and the third connection plate (6) are connected by welding or riveting; The first connecting plate (4) and the second connecting plate (5) are both U-shaped steel plates, and a fixing piece is also provided at the connection between the first connecting plate (4) and the second connecting plate (5). The fixing piece is provided with a plurality of connecting screw holes, and the connection between the first connecting plate (4) and the second connecting plate (5) is achieved by matching the connecting screw holes with corresponding bolts.
10. The method for predicting the stability of a machine tool based on a side-mounted bed with a double triangle structure according to claim 9, characterized in that: A side wall guide rail (8) is also provided on the side wall of the bed body (1) at the upper end of the lower triangular connection structure (2); an upper guide rail (9) is also provided at the upper end of the bed body (1).
