Motor quasi-stop control method and system applied to horizontal inverted combined tool magazine

Abstract

The present invention relates to the technical field of motor control, and particularly to a motor quasi-stop control method and system applied to a horizontal inverted combined tool magazine. The method includes: selecting the zero position of the horizontal inverted combined tool magazine to establish a coordinate system; determining the target position of the motor; collecting the operating state parameters and actual position of the motor; establishing a motor quasi-stop control prediction model, using the operating state parameters as the input of the motor quasi-stop control prediction model to obtain a predicted position; calculating the prediction error between the target position and the predicted position, and formulating a motor operation adjustment strategy to preliminarily adjust the motor control parameters; comparing the actual position with the target position through a feedback control algorithm to obtain an actual error signal and optimize the motor control parameters. Through the present invention, the complex control problems caused by the influence of inertia and gravity in the horizontal inverted combined tool magazine are effectively solved, the quasi-stop accuracy of the motor is improved, and at the same time, the operation efficiency and stability of the entire tool magazine system are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of motor control, and particularly to a motor quasi-stop control method and system applied to a horizontal inverted combined tool magazine. Background Art

[0002] Horizontal inverted combined tool magazines are widely used in various numerically controlled machine tools, especially in manufacturing environments that require rapid switching of multiple tools to complete complex machining tasks, such as the aerospace, automotive manufacturing, and mold processing industries. In a horizontal inverted combined tool magazine, due to its unique structural design, the spatial layout of the tool magazine and the tools is more complex. At the same time, the influence of inertia and gravity brought by the inverted structure is more obvious. This complexity places higher requirements on the precise control of the motor. Especially during the tool change process, it is required that the motor can accurately stop at a specified position during the tool change process to ensure the safe locking and rapid exchange of the tools.

[0003] In traditional tool magazine control systems, the docking position of the motor usually depends on a simple feedback control system or a control algorithm based on fixed parameters. This single error compensation mechanism has insufficient accuracy in complex scenarios. For example, the influence of gravity and load, especially when random errors and systematic errors coexist, it is easy to lead to poor control effects. In the face of complex working conditions, the motor must control the tool magazine to accurately stop in the inverted state. If a traditional motor control scheme is used, it may not be able to effectively counteract the gravity, resulting in problems such as inaccurate tool replacement and increased machining errors due to the inability of the tool to be accurately parked. Therefore, developing a motor quasi-stop control method applied to a horizontal inverted combined tool magazine has become one of the current research directions. Summary of the Invention

[0004] The present invention provides a motor quasi-stop control method and system applied to a horizontal inverted combined tool magazine, which can effectively solve the problems in the background art.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] A motor quasi-stop control method applied to a horizontal inverted combined tool magazine, the method comprising:

[0007] Select the zero position of the horizontal inverted combined tool magazine and establish a coordinate system based on the zero position;

[0008] Determine the target position of the motor according to the coordinate system;

[0009] Collect the operating state parameters and actual position of the motor;

[0010] Establish a motor quasi-stop control prediction model, and use the operating state parameters as the input of the motor quasi-stop control prediction model to predict the quasi-stop position of the motor to obtain a predicted position;

[0011] Calculate the prediction error between the target position and the predicted position, and formulate a motor operation adjustment strategy based on the prediction error to preliminarily adjust the motor control parameters;

[0012] Compare the actual position with the target position through a feedback control algorithm to obtain an actual error signal, and optimize the motor control parameters according to the actual error signal.

[0013] Further, determining the target position of the motor according to the coordinate system includes:

[0014] Define the working area coordinates of the horizontal inverted combined tool magazine according to the coordinate system, and establish a coordinate database;

[0015] Number the working area of the horizontal inverted combined tool magazine, and identify the target working area coordinates in the coordinate database according to the number;

[0016] Introduce the target position offset of the motor, and generate the target position of the motor in combination with the target position offset and the target working area coordinates.

[0017] Further, introducing the target position offset of the motor includes:

[0018] Based on the established coordinate system, define the direction of the target position offset through the direction of the target position of the motor relative to the working area;

[0019] Determine the space required by the motor, define the minimum safety distance, and calculate the magnitude of the target offset according to the space required by the motor and the minimum safety distance. The calculation formula is as follows:

[0020] ;

[0021] Wherein, is the target offset, is the direction coefficient, is the motor length, is the minimum safety distance, is the clearance distance;

[0022] Generate the target position of the motor in combination with the magnitude and direction of the target position offset and the target working area coordinates.

[0023] Further, it includes introducing the target position offset for each axis in the established coordinate system, and calculating the target position separately for each axis.

[0024] Further, establishing the motor quasi-stop control prediction model includes:

[0025] Collect the historical motion state data and position data of the motor, and perform timestamp alignment to construct a time series data set;

[0026] Preprocess the time series data set;

[0027] Select the structure of the motor quasi-stop control prediction model and compile the motor quasi-stop control prediction model;

[0028] Input the preprocessed time series data set and implement the training, evaluation, and optimization of the motor quasi-stop control prediction model;

[0029] Predict the quasi-stop position of the motor through the optimized motor quasi-stop control prediction model.

[0030] Further, the timestamp alignment includes:

[0031] Analyze the timestamp frequencies of each data and detect inconsistent timestamp intervals;

[0032] Select a suitable alignment method to align each data on the same time series;

[0033] Filter the data after timestamp alignment to obtain a time series data set.

[0034] Further, the motor quasi-stop control prediction model includes:

[0035] An input layer that accepts the processed time series data set;

[0036] A bidirectional LSTM layer, composed of a forward LSTM layer, a backward LSTM layer, and a merging layer, for processing time series data and capturing time dependencies and sequence information;

[0037] A dense layer, composed of several fully connected layers, that maps the output of the bidirectional LSTM layer to the required dimension through the fully connected layers;

[0038] A dropout layer that randomly drops a certain proportion of neurons during training;

[0039] An output layer that generates the final predicted result of the quasi-stop position of the motor.

[0040] Further, an error compensation layer is added after the bidirectional LSTM layer of the motor quasi-stop control prediction model to introduce system error compensation and dynamically adjust the predicted position. The error compensation formula is:

[0041] ;

[0042] Where, is the predicted position after error compensation, is the preliminary accurate stop position predicted by the bidirectional LSTM layer, is the mass of the load, is the acceleration due to gravity, is the angle between the motor movement direction and the vertical direction, is the moment of inertia of the motor, is the acceleration of the motor at time t, is the radius of rotation of the motor, is the coefficient of friction, is the normal force of the motor system, is the stiffness constant of the motor system.

[0043] Further, it includes: the merging layer splices the outputs generated by the forward LSTM layer and the backward LSTM layer in the feature dimension by means of connection.

[0044] Applied to the motor accurate stop control system of a horizontal inverted combined tool magazine, the system includes:

[0045] A coordinate system establishment module, selects the zero position of the horizontal inverted combined tool magazine, and establishes a coordinate system based on the zero position;

[0046] A target position determination module, determines the target position of the motor according to the coordinate system;

[0047] A data acquisition module, acquires the operating state parameters and the actual position of the motor;

[0048] A prediction model establishment module, establishes a motor accurate stop control prediction model, uses the operating state parameters as the input of the motor accurate stop control prediction model to predict the accurate stop position of the motor, and obtains a predicted position;

[0049] A prediction error adjustment module, calculates the prediction error between the target position and the predicted position, and formulates a motor operation adjustment strategy according to the prediction error to preliminarily adjust the motor control parameters;

[0050] An actual error adjustment module, compares the actual position with the target position through a feedback control algorithm to obtain an actual error signal, and optimizes the motor control parameters according to the actual error signal.

[0051] Through the technical solution of the present invention, the following technical effects can be achieved:

[0052] Effectively solves the complex control problems brought by inertia and gravity in the horizontal inverted combined tool magazine, making the tool change process more stable and efficient. At the same time, considering the historical state and future movement trend of the motor, the quasi-stop accuracy of the motor is improved, the adaptability of the system under complex working conditions is enhanced, and the operation efficiency and stability of the entire tool magazine system are improved.

[0053] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following specifically gives the specific implementation manners of this application. Brief Description of the Drawings

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to these drawings.

[0055] Figure 1 It is a schematic flow chart of a motor quasi-stop control method applied to a horizontal inverted combined tool magazine;

[0056] Figure 2 It is a schematic flow chart for timestamp alignment;

[0057] Figure 3 It is a schematic structural diagram of a motor quasi-stop control prediction model;

[0058] Figure 4 It is a schematic structural diagram of a motor quasi-stop control system applied to a horizontal inverted combined tool magazine. Detailed Description of the Embodiments

[0059] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this invention belongs. The terms used in the specification of this invention are only for the purpose of describing specific embodiments, and are not intended to limit this invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0061] Embodiment 1

[0062] Such as Figure 1As shown, a motor quasi-stop control method applied to a horizontal inverted combined tool magazine, the method comprising:

[0063] S1: Select the zero position of the horizontal inverted combined tool magazine and establish a coordinate system based on the zero position;

[0064] Specifically, in order to accurately position and control the movement of the motor in the system, the zero position can be determined to establish a coordinate system to uniformly describe the positions of the motor and all other components. By clarifying the zero point and establishing the coordinate system, the control system can accurately control the mechanical movement according to the position data in the coordinate system. For the established coordinate system, it can be a Cartesian coordinate system, a polar coordinate system, etc. Selecting a suitable coordinate system can better achieve the precise positioning of the tool magazine motor control.

[0065] S2: Determine the target position of the motor according to the coordinate system;

[0066] S3: Collect the operating state parameters and actual position of the motor;

[0067] Specifically, in order to perform precise motor control, the operating state parameters and actual position of the motor can be collected, including parameters such as position, speed, voltage, current, etc. To collect these parameters, appropriate sensors or detection devices can be selected, installed and calibrated, and a data acquisition system can be prepared. The data acquisition system needs to have sufficient input channels and be able to support sensors with a high sampling rate; sensor signals may be affected by noise such as electromagnetic interference and mechanical vibration. Therefore, when collecting data, a low-pass filter or other signal processing techniques can be used to eliminate noise and signal interference to ensure the accuracy of the collected data, obtain high-precision motor operating parameters, and provide strong support for the control system.

[0068] S4: Establish a motor quasi-stop control prediction model, use the operating state parameters as the input of the motor quasi-stop control prediction model to predict the quasi-stop position of the motor, and obtain the predicted position;

[0069] Based on the above embodiments, in order to predict the quasi-stop position of the motor in advance and avoid overshoot or insufficiency, a suitable model can be selected as the motor quasi-stop control prediction model, including long short-term memory network, support vector regression, multi-layer perceptron, etc. The quasi-stop position of the motor is predicted through the operating state parameters of the motor; and step S3 continuously collects the real-time operating state parameters of the motor, and these parameters can reflect the current operating state of the motor. If the collected operating state parameters are to be used as the input of the prediction model, the collected operating state parameters can be preprocessed, including operations such as denoising and normalization, to ensure the quality and consistency of the data and meet the input requirements of the prediction model.

[0070] S5: Calculate the prediction error between the target position and the predicted position, and formulate a motor operation adjustment strategy based on the prediction error to preliminarily adjust the motor control parameters;

[0071] S6: Compare the actual position with the target position through a feedback control algorithm to obtain an actual error signal, and optimize the motor control parameters according to the actual error signal.

[0072] In this embodiment, the motor control parameters can be adjusted in stages. The prediction error adjustment is carried out in the initial and middle stages of the motor movement, mainly to reduce large-range errors, while the actual error adjustment is carried out when the motor is approaching the target position, mainly for precise correction; in step S5, the target position of the motor is compared with the predicted stop position predicted by the prediction model, and the error between the two is calculated, that is, the prediction error, which is mainly used to formulate a preliminary adjustment strategy to adjust the operation state of the motor in advance to reduce the deviation when the motor actually stops. For example, if the prediction error is large and the motor is approaching the target position, the motor speed can be gradually reduced to make it slowly approach the target, etc.

[0073] In step S6, based on the difference between the current position obtained by actual measurement and the target position, and based on the actual error signal, the control system will adjust the operation parameters of the motor to further optimize the operation of the motor. Usually, classical feedback control algorithms such as PID control are used in this stage to ensure that the motor can be quickly and stably adjusted to the target position; during the motor movement process, the system usually gradually transitions from a rough adjustment based on the prediction error to a fine adjustment based on the actual error, and the two complement each other to jointly ensure that the motor can accurately stop at the target position.

[0074] The present invention effectively solves the complex control problems brought by the influence of inertia and gravity in the horizontal inverted combined tool magazine, makes the tool replacement process more stable and efficient, and at the same time considers the historical state and future movement trend of the motor, thereby improving the quasi-stop accuracy of the motor, enhancing the adaptability of the system under complex working conditions, and at the same time improving the operation efficiency and stability of the entire tool magazine system.

[0075] On the basis of the above embodiment, determining the target position of the motor according to the coordinate system includes:

[0076] A1: Define the working area coordinates of the horizontal inverted combined tool magazine according to the coordinate system, and establish a coordinate database;

[0077] A2: Number the working area of the horizontal inverted combined tool magazine, and identify the target working area coordinates in the coordinate database according to the number;

[0078] Specifically, to ensure that the motor can access each tool precisely and systematically, a spatial position can be defined for each tool or station in the horizontal inverted combined tool magazine. The coordinates of each tool position should be defined relatively based on the zero position, and should be accurately calibrated according to the geometric structure of the tool magazine design, and managed through the establishment of a database; each working area will be assigned a unique number, and this number is associated with the corresponding coordinates in the coordinate database. When a specific tool needs to be found, the system can query the coordinate database through the number to find the coordinates of the corresponding working area.

[0079] A3: Introduce the target position offset of the motor, and generate the target position of the motor by combining the target position offset and the target working area coordinates.

[0080] In this embodiment, the introduction of the offset is to meet different operation requirements and ensure that the motor can adapt to factors such as tool type, station characteristics, or mechanical structure when stopping at the precise position. For example, for larger or longer tools, the motor may need to dock at a slightly farther position to avoid collisions, while for smaller tools, the motor can be closer to the center; by introducing the offset, the system can flexibly adjust the position of the motor to accurately dock it to a specific tool position, ensuring that the motor has a more appropriate and optimized final target position before the movement starts.

[0081] To obtain the optimal target position of the motor, introduce the target position offset of the motor, including:

[0082] A31: Based on the established coordinate system, define the direction of the target position offset through the direction of the target position of the motor relative to the working area;

[0083] A32: Determine the space required by the motor, define the minimum safety distance, and calculate the magnitude of the target offset according to the space required by the motor and the minimum safety distance. The calculation formula is as follows:

[0084] ;

[0085] Where, is the target offset, is the direction coefficient, is the length of the motor, is the minimum safety distance, is the clearance distance;

[0086] Specifically, to ensure that the motor and the tool do not collide, a certain safety distance should be left when the motor approaches the tool. To align the center of the motor with the tool, the target position of the motor needs to consider the length of the motor itself, and it should be ensured that the motor does not touch other obstacles or equipment during movement, especially in a complex environment, so a clearance distance can be left. To ensure that the final movement path of the motor is reasonable and the motor can move in the correct direction, a direction coefficient can be introduced. Through the direction coefficient, it can be determined whether the motor moves in the positive or negative direction on each coordinate axis, so as to ensure that the motor can dock with the tool in the correct direction.

[0087] A33: Generate the target position of the motor by combining the magnitude and direction of the target position offset and the target working area coordinates.

[0088] In this step, by combining the magnitude and direction of the offset, the standard target working area coordinates are adjusted to generate the final target position of the motor. The calculation of the target position will be based on the standard coordinates plus the offset. The calculation formula is: Motor target position = Target working area coordinates + Offset magnitude × Offset direction.

[0089] As a preference of this embodiment, in the established coordinate system, a target position offset needs to be introduced for each axis, and the calculation of the target position is performed separately for each axis.

[0090] Specifically, since the motor operates in multiple axes and the target position of each axis needs to be calculated separately, the offset of each axis can be added to the standard target position to finally obtain the actual target position of the motor on each axis. This method of calculating the target position axis by axis can ensure precise control of the motor in each direction in a complex three-dimensional space, guarantee the best movement effect of the motor on each axis in the three-dimensional space, and ensure that the offsets in different axes do not affect each other.

[0091] Based on the above embodiment, a motor quasi-stop control prediction model is established, including:

[0092] B1: Collect the historical motion state data and position data of the motor, and perform timestamp alignment to construct a time series data set;

[0093] B2: Preprocess the time series data set;

[0094] The data collected in this step comes from multiple sensors or devices and may be acquired at different times. To ensure that all status and position data can accurately reflect the state of the motor at the same moment, timestamp alignment can be performed to ensure that each data point can be matched with other relevant data points; and the collected data is used as the training data for the prediction model. To ensure the quality and consistency of the input data, data preprocessing can be performed, including operations such as data cleaning, normalization or standardization, smoothing or noise reduction, etc., so that the model can effectively learn on a cleaner and more normalized dataset.

[0095] B3: Select the structure of the motor quasi-stop control prediction model and compile the motor quasi-stop control prediction model;

[0096] Specifically, to enable the model to be efficiently trained and optimized, an appropriate model structure can be selected, and each component of the model can be configured for compilation. Compiling the model is to transform the selected model structure into an executable training framework, defining parameters such as the optimization objective, loss function, optimizer, etc., so that the model can be correctly trained and evaluated.

[0097] B4: Input the preprocessed time series dataset and implement the training, evaluation, and optimization of the motor quasi-stop control prediction model;

[0098] In the time series dataset, the original time series data is divided into multiple subsequences to construct sample-label pairs. Each subsequence is the input of the model, and the label is the quasi-stop position of the motor. After construction, the time series dataset can be divided into a training set, a validation set, and a test set to train the model; during the training process of the model, the model parameters can be adjusted through the backpropagation algorithm to reduce the loss value and improve the prediction ability of the model. After each round of training, the validation set data can be used to evaluate the model, calculate the loss and metrics on the validation set, monitor the generalization ability of the model, and prevent overfitting or underfitting. After the model training is completed, the test set can be used to evaluate the final performance of the model. The test set should be data that the model has never seen during the training process; after the prediction model is trained, validated, and evaluated, it can gradually learn the relationship between the motor motion state and the quasi-stop position and provide accurate quasi-stop position predictions in the final application.

[0099] B5: Predict the quasi-stop position of the motor through the optimized motor quasi-stop control prediction model.

[0100] Based on the above embodiments, as Figure 2 shown, performing timestamp alignment includes:

[0101] B11: Analyze the timestamp frequencies of each data and detect inconsistent timestamp intervals;

[0102] B12: Select an appropriate alignment method to align each data on the same time series;

[0103] B13: Filter the data with aligned timestamps to obtain a time series dataset.

[0104] As a preference of this embodiment, in order for all data to be within the same time series framework, the timestamp frequencies of each data source can be analyzed to determine whether there are inconsistent time intervals, find the inconsistent time intervals in the dataset, determine which data points need to be adjusted, and mark those points with overly large or small intervals, and prepare to take corresponding alignment measures, including:

[0105] (1) Upsampling: If the interval between data points is too large, more data points can be generated through interpolation to adjust the data to a higher-frequency sampling interval.

[0106] (2) Downsampling: If the interval between data points is too small or the data volume is too large, the data points can be reduced by means such as averaging or taking the median value to adjust the data to a lower-frequency sampling interval.

[0107] After upsampling or downsampling all datasets, all data points can be aligned to a unified time point, such as every second, every minute, etc., to ensure that all datasets are aligned with this time series.

[0108] As a preference of this embodiment, as Figure 3 shown, the motor accurate stop control prediction model includes:

[0109] An input layer, which accepts the processed time series dataset;

[0110] The input layer is the first layer of the neural network, responsible for accepting the processed time series dataset and inputting the motion state and position data of the motor into the subsequent layers of the model for processing and prediction.

[0111] A bidirectional LSTM layer, which consists of a forward LSTM layer, a backward LSTM layer and a merging layer, and is used to process time series data and capture time-dependent relationships and sequence information;

[0112] The bidirectional LSTM layer can use two LSTM layers to process time series data, obtain information from both past and future time steps simultaneously to improve the prediction accuracy. The LSTM can capture long-term dependencies, and the bidirectional structure can extract useful features from the bidirectional time series.

[0113] The bidirectional LSTM layer consists of a forward LSTM layer, a backward LSTM layer, and a merging layer. The forward LSTM layer processes the input time series data, gradually processing the data from the start point to the end point of the time series, capturing the dependencies from the past to the future in the time series; the backward LSTM layer corresponds to the forward LSTM layer, and it processes the data from the end point to the start point of the time series, capturing the dependencies from the future to the past; the merging layer combines the outputs of the forward LSTM layer and the backward LSTM layer to form a comprehensive representation based on forward and backward information.

[0114] The dense layer consists of several fully connected layers, mapping the output of the bidirectional LSTM layer to the required dimension through the fully connected layers;

[0115] The dense layer consists of several fully connected layers. Each neuron in a layer is connected to all neurons in the previous layer. The output of the bidirectional LSTM layer is usually a high-dimensional time series representation. The dense layer can map the output of the bidirectional LSTM layer to the specified dimension through full connection, further processing the data output by the LSTM layer; in order to enhance the expressive power of the model, multiple fully connected layers can be stacked so that the model can gradually extract high-order features in the data. Each fully connected layer further processes the output of the previous layer, gradually compressing the high-dimensional data into the required dimension.

[0116] The dropout layer randomly drops a certain proportion of neurons during the training process;

[0117] The dropout layer can randomly drop a certain proportion of neurons during the training process to prevent the model from overfitting. And during each training process, the dropped neurons are different. Therefore, different parts of the model are forced to learn independent feature representations, which can effectively improve the generalization ability of the model and enhance the robustness and adaptability of the model.

[0118] The output layer generates the final predicted result of the accurate stop position of the motor.

[0119] To further improve the accuracy of the motor accurate stop control prediction model, an error compensation layer is added after the bidirectional LSTM layer of the motor accurate stop control prediction model, introducing system error compensation to dynamically adjust the predicted position. The error compensation formula is:

[0120] ;

[0121] where, is the predicted position after error compensation, is the preliminary accurate stop position predicted by the bidirectional LSTM layer, is the mass of the load, is the acceleration due to gravity, is the angle between the motor movement direction and the vertical direction, is the moment of inertia of the motor, is the acceleration of the motor at time t, is the radius of rotation of the motor, is the coefficient of friction, is the normal force of the motor system, is the stiffness constant of the motor system.

[0122] In this preferred solution, the error compensation layer is a functional layer added after the bidirectional LSTM layer. Its main function is to correct the quasi-stop position predicted by the bidirectional LSTM model through the compensation of system errors. Due to the influence of gravity, inertia, and friction on the quasi-stop position of the motor, the prediction result of the bidirectional LSTM may have deviations. The error compensation layer detects these errors and makes dynamic adjustments to ensure that the finally output quasi-stop position is more accurate.

[0123] As a preference of this embodiment, it includes: The merging layer splices the outputs generated by the forward LSTM layer and the backward LSTM layer in the feature dimension by connection.

[0124] Specifically, in order to completely retain all the information generated by the forward LSTM and the backward LSTM layers, while enhancing the expression ability of the model and maintaining the independence of the forward and backward features, the output feature vectors of the forward LSTM layer and the backward LSTM layer at the same time step can be spliced together in the feature dimension by connection to form a new feature vector. The model can obtain the feature representations of the past and the future at each time step, avoiding information loss. If splicing is performed by means such as summation or averaging, the forward and backward features will be compressed together, which may lead to information loss.

[0125] Embodiment 2:

[0126] As shown in Figure 4 , the motor quasi-stop control system applied to the horizontal inverted combined tool magazine, the system includes:

[0127] A coordinate system establishment module, which selects the zero position of the horizontal inverted combined tool magazine and establishes a coordinate system based on the zero position;

[0128] A target position determination module, which determines the target position of the motor according to the coordinate system;

[0129] A data acquisition module, which acquires the operating state parameters and the actual position of the motor;

[0130] A prediction model establishment module, which establishes a motor quasi-stop control prediction model, uses the operating state parameters as the input of the motor quasi-stop control prediction model to predict the quasi-stop position of the motor, and obtains the predicted position;

[0131] The prediction error adjustment module calculates the prediction error between the target position and the predicted position, and formulates a motor operation adjustment strategy according to the prediction error to preliminarily adjust the motor control parameters;

[0132] The actual error adjustment module compares the actual position with the target position through a feedback control algorithm to obtain an actual error signal, and optimizes the motor control parameters according to the actual error signal.

[0133] The above adjustment system in the present invention can effectively implement the motor accurate stop control method of the horizontal inverted combined tool magazine, and the technical effects that can be achieved are as described in the above embodiments, which will not be elaborated here.

[0134] Although the present application has been described in conjunction with specific features and their embodiments, it is obvious that various modifications and combinations can be made without departing from the spirit and scope of the present application. Accordingly, the present specification and the drawings are merely exemplary illustrations of the present application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A motor accurate stop control method applied to a horizontal inverted combined tool magazine, characterized in that: The method comprises: Selecting a zero point position of the horizontal inverted combined tool magazine, and establishing a coordinate system based on the zero point position; determining a target position of the motor according to the coordinate system; Collect the running status parameters and actual position of the motor; Establishing a motor quasi-stop control prediction model, using the operating state parameters as inputs of the motor quasi-stop control prediction model to predict the quasi-stop position of the motor to obtain a predicted position; Calculating a prediction error between the target position and the predicted position, and formulating a motor operation adjustment strategy according to the prediction error to preliminarily adjust the motor control parameters; Comparing the actual position with the target position through a feedback control algorithm to obtain an actual error signal, and optimizing the motor control parameters according to the actual error signal; The motor control parameters are adjusted in stages. The predicted error adjustment is performed in the early and middle stages of the motor movement, while the actual error adjustment is performed when the motor approaches the target position. During the motor movement, there is a gradual transition from coarse adjustment based on the predicted error to fine adjustment based on the actual error.

2. The motor accurate stop control method applied to the horizontal inverted combined tool magazine according to claim 1 is characterized in that: Determining the target position of the motor according to the coordinate system includes: Defining the coordinates of the working area of ​​the horizontal inverted combined tool magazine according to the coordinate system, and establishing a coordinate database; Numbering the working areas of the horizontal inverted combined tool magazine, and identifying the coordinates of the target working area in the coordinate database according to the numbers; The target position offset of the motor is introduced, and the target position of the motor is generated by combining the target position offset and the target working area coordinates.

3. The motor accurate stop control method applied to the horizontal inverted combined tool magazine according to claim 2 is characterized in that: The target position offset of the motor is introduced, including: Based on the established coordinate system, defining the direction of the target position offset by the direction of the target position of the motor relative to the working area; Determine the space required by the motor, define the minimum safety distance, and calculate the size of the target position offset according to the space required by the motor and the minimum safety distance. The calculation formula is as follows: ； in, is the target position offset, is the direction coefficient, is the motor length, is the minimum safe distance, is the clearance distance; The target position of the motor is generated by combining the size and direction of the target position offset and the target working area coordinates.

4. The motor accurate stop control method applied to the horizontal inverted combined tool magazine according to claim 3 is characterized in that: Including in the established coordinate system, the target position offset needs to be introduced for each axial direction, and the target position is calculated separately in each axial direction.

5. The motor accurate stop control method applied to the horizontal inverted combined tool magazine according to claim 1 is characterized in that: The step of establishing a motor accurate stop control prediction model comprises: Collect historical motion state data and position data of the motor, align timestamps, and construct a time series data set; Preprocessing the time series data set; Selecting the structure of the motor quasi-stop control prediction model and compiling the motor quasi-stop control prediction model; Input the preprocessed time series data set to implement the training, evaluation and optimization of the motor accurate stop control prediction model; The quasi-stop position of the motor is predicted by using the optimized motor quasi-stop control prediction model.

6. The motor accurate stop control method applied to the horizontal inverted combined tool magazine according to claim 5 is characterized in that: The timestamp alignment comprises: Analyze the timestamp frequency of each data and detect inconsistent timestamp intervals; Select an appropriate alignment method to align the data to the same time series; The data after the timestamp alignment is filtered to obtain a time series data set.

7. The motor accurate stop control method applied to the horizontal inverted combined tool magazine according to claim 5 is characterized in that: The motor accurate stop control prediction model includes: An input layer, receiving the processed time series dataset; The bidirectional LSTM layer, which consists of a forward LSTM layer, a backward LSTM layer, and a merging layer, is used to process time series data and capture time dependencies and sequence information; A dense layer, composed of several fully connected layers, maps the output of the bidirectional LSTM layer to the required dimension through the fully connected layers; Dropout layer, randomly discarding a certain proportion of neurons during training; The output layer generates the final prediction result of the accurate stop position of the motor.

8. The motor accurate stop control method applied to the horizontal inverted combined tool magazine according to claim 7 is characterized in that: An error compensation layer is added after the bidirectional LSTM layer of the motor accurate stop control prediction model, system error compensation is introduced, and the predicted position is dynamically adjusted. The error compensation formula is: ； in, is the predicted position after error compensation, is the preliminary accurate stop position predicted by the bidirectional LSTM layer, is the mass of the load, is the acceleration due to gravity, is the angle between the motor movement direction and the vertical direction, is the moment of inertia of the motor, is the acceleration of the motor at time t, is the rotation radius of the motor, is the friction coefficient, is the normal force of the motor system, is the stiffness constant of the motor system.

9. The motor accurate stop control method applied to the horizontal inverted combined tool magazine according to claim 7, characterized in that: include: The merging layer concatenates the outputs generated by the forward LSTM layer and the backward LSTM layer in the feature dimension by means of connection.

10. The motor accurate stop control system applied to the horizontal inverted combined tool magazine is characterized by: The system comprises: A coordinate system establishment module selects a zero point position of the horizontal inverted combined tool magazine and establishes a coordinate system based on the zero point position; A target position determination module, which determines the target position of the motor according to the coordinate system; Data acquisition module, collecting the running status parameters and actual position of the motor; A prediction model building module is used to build a motor quasi-stop control prediction model, and use the operating state parameters as inputs of the motor quasi-stop control prediction model to predict the quasi-stop position of the motor to obtain a predicted position; A prediction error adjustment module is used to calculate the prediction error between the target position and the predicted position, and formulate a motor operation adjustment strategy according to the prediction error to perform preliminary adjustment on the motor control parameters; An actual error adjustment module compares the actual position with the target position through a feedback control algorithm to obtain an actual error signal, and optimizes the motor control parameters according to the actual error signal; The motor control parameters are adjusted in stages. The predicted error adjustment is performed in the early and middle stages of the motor movement, while the actual error adjustment is performed when the motor approaches the target position. During the motor movement, there is a gradual transition from coarse adjustment based on the predicted error to fine adjustment based on the actual error.

Images