Automatic operation control method and system for spiral flat lathe

Abstract

The present application relates to the field of spiral flat car machine control technology, and discloses a spiral flat car machine automatic operation control method and system. The method comprises: according to the spiral motion characteristic parameter set and the working condition prediction matrix, the transmission area of ​​the spiral flat car machine is divided into n equal parts according to the spiral direction to obtain n adaptive data acquisition control units and a multi-dimensional state monitoring matrix; the multi-dimensional state monitoring matrix is ​​subjected to feature extraction processing to obtain a working condition evaluation vector and a dynamic deviation feature set; the working condition evaluation vector and the dynamic deviation feature set are subjected to iterative optimization processing to obtain a real-time feedback control instruction set and a dynamic compensation parameter matrix; the spiral flat car machine is subjected to intelligent closed-loop control according to the real-time feedback control instruction set and the dynamic compensation parameter matrix to obtain real-time control data and a process feature mapping set. The present application realizes high-precision automatic control of the spiral flat car machine during continuous loading operations, ensuring the uniformity of material loading and operating efficiency.

Description

Technical Field

[0001] The present application relates to the field of spiral lathe control, and in particular to an automated operation control method and system for a spiral lathe. Background Art

[0002] With the rapid development of railway transportation business, the automation and intelligence of train loading operations are constantly improving. The spiral flat car machine is an important loading equipment in railway freight, mainly used for loading operations of bulk materials such as coal and ore. The current spiral flat car machine loads the goods through a spiral guide rail conveyor and is equipped with basic control systems and monitoring devices, such as position sensors, weight sensors, etc., to monitor the basic parameters of the loading process. With the increase in railway freight volume, some railway transportation companies have also begun to introduce digital control systems on spiral flat car machines to perform basic data collection and control of the loading process.

[0003] However, the existing spiral flat car control method still has many shortcomings: the material distribution during loading is uneven, which can easily cause the car body to tilt and cause safety hazards; the lack of accurate real-time monitoring and dynamic compensation mechanism makes it difficult to cope with various abnormal situations during loading; the control accuracy is not high, and the control error of material loading amount is large; the coordination between the control units is poor, which affects the overall loading efficiency; the data acquisition and processing capabilities are limited, and cannot meet the precise control requirements of large-scale freight operations. Summary of the invention

[0004] The present application provides a method and system for controlling the automated operation of a spiral flat car machine, which are used to achieve high-precision automated control of the spiral flat car machine during continuous loading operations, thereby ensuring uniformity of material loading and operating efficiency.

[0005] In the first aspect, the present application provides an automated operation control method for a spiral flat car machine, the automated operation control method for a spiral flat car machine comprising: collecting working state parameters from the spiral flat car machine to generate an operating state parameter set and a multidimensional feature trajectory vector; calculating the operating state parameter set and the multidimensional feature trajectory vector, and outputting a spiral motion feature parameter set and a working condition prediction matrix; dividing the spiral flat car machine transmission area into n adaptive data acquisition control units according to the spiral direction according to the spiral motion feature parameter set and the working condition prediction matrix, and generating a multidimensional state monitoring matrix, wherein n is a positive integer greater than or equal to 5; extracting feature data of the multidimensional state monitoring matrix, and calculating a working condition evaluation vector and a dynamic deviation feature set; iteratively optimizing the working condition evaluation vector and the dynamic deviation feature set through a recursive neural network, and generating a real-time feedback control instruction set and a dynamic compensation parameter matrix; and having n adaptive data acquisition control units perform intelligent closed-loop control according to the real-time feedback control instruction set and the dynamic compensation parameter matrix, and output real-time control data and a process feature mapping set.

[0006] In a second aspect, the present application provides an automatic operation control system for a spiral flat lathe, the automatic operation control system for a spiral flat lathe comprising:

[0007] The acquisition module is used to collect working state parameters from the spiral flat car machine and generate an operating state parameter set and a multi-dimensional feature trajectory vector;

[0008] A calculation module, used for calculating the operating state parameter set and the multi-dimensional characteristic trajectory vector, and outputting a spiral motion characteristic parameter set and an operating condition prediction matrix;

[0009] A processing module, for dividing the spiral flat car machine transmission area into n adaptive data acquisition control units according to the spiral motion characteristic parameter set and the working condition prediction matrix, and generating a multi-dimensional state monitoring matrix, wherein n is a positive integer greater than or equal to 5;

[0010] An extraction module is used to extract characteristic data of the multidimensional condition monitoring matrix and calculate a working condition evaluation vector and a dynamic deviation characteristic set;

[0011] An optimization module, used for iteratively optimizing the operating condition evaluation vector and the dynamic deviation feature set through a recursive neural network to generate a real-time feedback control instruction set and a dynamic compensation parameter matrix;

[0012] The control module is used for executing intelligent closed-loop control according to the real-time feedback control instruction set and the dynamic compensation parameter matrix by n adaptive data acquisition control units, and outputting real-time control data and process characteristic mapping sets.

[0013] In the technical solution provided by the present application, by performing multi-dimensional data acquisition and processing on the working state, the operating state parameter set and the multi-dimensional characteristic trajectory vector are obtained, and the comprehensive monitoring of the operating state of the spiral flat car machine is realized, providing a reliable data basis for subsequent control, and by performing numerical calculation and processing on the operating state parameter set and the multi-dimensional characteristic trajectory vector, the spiral motion characteristic parameter set and the working condition prediction matrix are obtained, and the motion characteristics of the equipment and the future working condition change trend are accurately grasped. According to the spiral motion characteristic parameter set and the working condition prediction matrix, the transmission area is divided into n equal parts, n adaptive data acquisition control units are set and a multi-dimensional state monitoring matrix is ​​obtained, and the refined partition control of the transmission area of ​​the spiral flat car machine is realized. Feature extraction processing is performed on the multi-dimensional state monitoring matrix to obtain the working condition evaluation vector and the dynamic deviation feature set, which provides a basis for the identification and processing of abnormal working conditions. The working condition evaluation vector and the dynamic deviation feature set are iteratively optimized by a recursive neural network to obtain a real-time feedback control instruction set and a dynamic compensation parameter matrix, which improves the generation efficiency and accuracy of the control instructions. Finally, based on n adaptive data acquisition control units, intelligent closed-loop control is performed according to the real-time feedback control instruction set and dynamic compensation parameter matrix to obtain real-time control data and process feature mapping sets, thus realizing precise control and dynamic compensation of the spiral flat car machine. The entire control process forms a complete closed loop, from data acquisition, feature extraction, instruction generation to control execution, and each link is closely connected, ensuring the stability and reliability of the loading process, and effectively solving the problems of response lag and low control accuracy in traditional control methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0015] Figure 1 A schematic diagram of an embodiment of the automatic operation control method of the spiral flat lathe in the embodiment of the present application;

[0016] Figure 2 This is a schematic diagram of an embodiment of the automatic operation control system of the spiral flat car machine in the embodiment of the present application. DETAILED DESCRIPTION

[0017] The embodiments of the present application provide a method and system for controlling the automated operation of a spiral flat lathe. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0018] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 , an embodiment of the spiral flat lathe automatic operation control method in the embodiment of the present application includes:

[0019] Step S101, collecting working state parameters from the spiral flat car machine, generating an operating state parameter set and a multi-dimensional feature trajectory vector;

[0020] Step S102, calculating the operating state parameter set and the multi-dimensional characteristic trajectory vector, and outputting the spiral motion characteristic parameter set and the operating condition prediction matrix;

[0021] Step S103, according to the spiral motion characteristic parameter set and the working condition prediction matrix, the spiral flat car machine transmission area is divided into n adaptive data acquisition control units according to the spiral direction, and a multi-dimensional state monitoring matrix is ​​generated, wherein n is a positive integer greater than or equal to 5;

[0022] Step S104, extracting characteristic data of the multidimensional condition monitoring matrix, and calculating a working condition assessment vector and a dynamic deviation feature set;

[0023] Step S105, iteratively optimizing the operating condition evaluation vector and the dynamic deviation feature set through a recursive neural network to generate a real-time feedback control instruction set and a dynamic compensation parameter matrix;

[0024] Step S106: n adaptive data acquisition control units execute intelligent closed-loop control according to the real-time feedback control instruction set and the dynamic compensation parameter matrix, and output real-time control data and a process characteristic mapping set.

[0025] It is understandable that the execution subject of the present application can be the automatic operation control system of the spiral flat car machine, or a terminal or a server, which is not limited here. The present application embodiment is described by taking the server as the execution subject as an example.

[0026] Specifically, multi-dimensional data acquisition and processing is performed on the working state. The specific acquisition process includes three-dimensional coordinate acquisition of the spatial position of the spiral guide rail, involving position data in the three directions of X, Y, and Z, and collecting pressure data of the presser foot height sensor array. The multi-point tension sensing network is arranged at the key position of the spiral guide rail to collect tension distribution data and edge profile features. The high-precision encoder is responsible for collecting speed signals, and the mechanical vibration sensor collects vibration data during equipment operation. After preprocessing, these raw data are obtained to obtain an operating state parameter set, including spiral trajectory benchmark data, pressure distribution vectors, etc., and a multi-dimensional feature trajectory vector is obtained to reflect the characteristics of the equipment motion trajectory. Then, the operating state parameter set and the multi-dimensional feature trajectory vector are numerically calculated. The key state features are extracted by data dimensionality reduction technology, and the feature data are numerically differentiated to obtain velocity and acceleration information. The reference trajectory set is used to extract spiral parameters to obtain pitch distribution and angle change data. After correlation analysis and data fusion, these processing results finally form a spiral motion feature parameter set to describe the equipment motion law and a working condition prediction matrix for predicting the equipment operation state.

[0027] According to the obtained parameter set and prediction matrix, the transmission area is divided into n equal parts. First, the complete motion trajectory of the spiral guide is obtained by curve fitting, and then equal arc length segmentation is performed to obtain n candidate control areas. Each area is equipped with an independent data acquisition control unit, which is responsible for data acquisition and control execution in the area. Through regional mapping and configuration processing, the coordinated work of each control unit is ensured to form a multi-dimensional state monitoring matrix to achieve full-area data monitoring. Feature extraction is performed on the multi-dimensional state monitoring matrix, and the monitoring data sequence is obtained by matrix decomposition. Time series analysis is performed to obtain state change information. Principal element analysis extracts key parameters, and data clustering forms parameter groupings. These processing results are finally integrated into a working condition evaluation vector, which reflects the operating status of each control unit, as well as a dynamic deviation feature set to characterize the operating deviation.

[0028] Recursive neural network is used for deep optimization. First, the data is standardized and reconstructed to form an evaluation data matrix and a deviation component vector. A time series feature network is established through time series coding and feature embedding. The hidden state sequence is obtained by recursive calculation, the long short-term memory unit is used for state transfer, and the gated recurrent unit is used for data fusion. Finally, a real-time feedback control instruction set and a dynamic compensation parameter matrix are generated.

[0029] The control unit based on the configuration performs intelligent closed-loop control. After the instruction is parsed, it is distributed to each control unit, and the compensation parameters are combined for regional matching and timing analysis. The consistency of control execution is ensured through synchronous coordination, and finally the real-time control data and process feature mapping set are obtained.

[0030] For example, when a spiral flat car machine is loading ore, the data of each sensor is collected first. The three-dimensional position sensor collects the coordinate data of 20 key points of the spiral guide rail, and the pressure sensor array collects the data of 100 pressure sampling points to form an operating status parameter set. These data are processed to extract the key characteristic parameters of the spiral motion, including the pitch variation range between 800-1200mm and the spiral angle variation between 30-45 degrees. The transmission area is divided into 8 control units, each of which is equipped with an independent data acquisition module. Through feature extraction, it is found that the pressure distribution in a certain area is uneven and the dynamic deviation exceeds the set threshold. The recursive neural network generates compensation instructions to adjust the loading parameters of the area, and finally reduces the pressure distribution deviation to the normal range to ensure a smooth loading process.

[0031] In the embodiment of the present application, by performing multi-dimensional data acquisition and processing on the working state, the operating state parameter set and the multi-dimensional characteristic trajectory vector are obtained, and the comprehensive monitoring of the operating state of the spiral flat car machine is realized, providing a reliable data basis for subsequent control, and by performing numerical calculation and processing on the operating state parameter set and the multi-dimensional characteristic trajectory vector, the spiral motion characteristic parameter set and the working condition prediction matrix are obtained, and the motion characteristics of the equipment and the future working condition change trend are accurately grasped. According to the spiral motion characteristic parameter set and the working condition prediction matrix, the transmission area is divided into n equal parts, n adaptive data acquisition control units are set and a multi-dimensional state monitoring matrix is ​​obtained, and the refined zoning control of the transmission area of ​​the spiral flat car machine is realized. Feature extraction processing is performed on the multi-dimensional state monitoring matrix to obtain the working condition evaluation vector and the dynamic deviation feature set, which provides a basis for the identification and processing of abnormal working conditions. The working condition evaluation vector and the dynamic deviation feature set are iteratively optimized by a recursive neural network to obtain a real-time feedback control instruction set and a dynamic compensation parameter matrix, which improves the generation efficiency and accuracy of the control instructions. Finally, based on n adaptive data acquisition control units, intelligent closed-loop control is performed according to the real-time feedback control instruction set and dynamic compensation parameter matrix to obtain real-time control data and process feature mapping sets, thus realizing precise control and dynamic compensation of the spiral flat car machine. The entire control process forms a complete closed loop, from data acquisition, feature extraction, instruction generation to control execution, and each link is closely connected, ensuring the stability and reliability of the loading process, and effectively solving the problems of response lag and low control accuracy in traditional control methods.

[0032] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0033] (1) Collect the three-dimensional coordinates of the spiral guide rail's spatial position, calculate the spiral trajectory reference data and position deviation matrix, synchronously read the real-time data of the presser foot height sensor array, and output the pressure distribution vector;

[0034] (2) reading multi-point tension sensor network data, generating a tension distribution curve and edge profile data, analyzing the edge profile data using Fourier transform, and obtaining a characteristic frequency set;

[0035] (3) Perform filtering operations on the high-precision encoder speed signal, derive the spiral motion angular velocity matrix and phase difference data, use wavelet decomposition to process the mechanical vibration sensor signal, and calculate the multi-scale vibration feature sequence;

[0036] (4) fusing the spiral trajectory reference data with the position deviation matrix to generate a spatial motion trajectory set, processing the pressure distribution vector and the tension distribution curve according to tensor decomposition, and outputting a force characteristic matrix;

[0037] (5) analyzing the cross-correlation between the characteristic frequency set and the spiral motion angular velocity matrix to determine a synchronization control sequence, and deriving a state characteristic vector based on the time-frequency analysis of the multi-scale vibration characteristic sequence and the phase difference data;

[0038] (6) Aggregating the spatial motion trajectory set, the force characteristic matrix, the synchronization control sequence, and the state characteristic vector to calculate the operating state parameter set and the multi-dimensional characteristic trajectory vector.

[0039] Specifically, in railway freight, the spatial position of the spiral guide rail is continuously sampled by installing a laser ranging sensor array on the guide rail. The sampled data includes the X, Y, and Z coordinate values ​​of each point on the guide rail. These original coordinate data are processed by coordinate transformation and calibration to form the spiral track benchmark data. At the same time, by comparing and analyzing the real-time collected data with the benchmark data, the position deviation of each sampling point is calculated, and the position deviation matrix is ​​constructed. The pressure foot height sensor array is arranged at the key position of the guide rail to collect pressure data in real time. The pressure values ​​between the sampling points are supplemented by the linear interpolation algorithm, and finally a pressure distribution vector representing the pressure distribution of the entire guide rail is generated. The multi-point tension sensing network is arranged at multiple node positions of the spiral guide rail to monitor the tension changes in real time during the material transportation process. After the tension data collected by each sensor node is standardized, a continuous tension distribution curve is generated by the interpolation algorithm. At the same time, the edge sensor collects the edge contour data of the material. These data are processed by Fourier transform, the time domain signal is converted into a frequency domain signal, the main frequency component is extracted, and a characteristic frequency set is formed. Fourier transform can effectively identify the periodic change characteristics during material transportation.

[0040] The high-precision encoder is installed on the driving device of the spiral guide to collect the speed signal. The original speed signal is processed by low-pass filtering to remove high-frequency noise, obtain smooth angular velocity data, and obtain the spiral motion angular velocity matrix through matrix operation. By analyzing the signal phase difference of adjacent sampling points, the phase difference data is obtained to evaluate the synchronization of motion. The vibration signal collected by the mechanical vibration sensor is processed by wavelet decomposition to decompose the signal into sub-signals in different frequency bands to obtain a vibration feature sequence containing multiple scale features. The spiral trajectory benchmark data and the position deviation matrix are fused, and the weighted average method is used to comprehensively consider the benchmark trajectory and real-time deviation to generate a set of spatial motion trajectories reflecting the actual motion state. The pressure distribution vector and the tension distribution curve are processed by tensor decomposition to decompose the multidimensional data into core tensors and factor matrices, extract the main force characteristics, and form a force feature matrix.

[0041] The characteristic frequency set and the spiral motion angular velocity matrix are cross-correlatedly analyzed to calculate the correlation between different frequency components and angular velocity changes, and generate a synchronous control sequence. The multi-scale vibration characteristic sequence and phase difference data are processed by time-frequency analysis, and the short-time Fourier transform method is used to obtain the distribution characteristics of the vibration signal in the time and frequency domains to form a state characteristic vector. Finally, the spatial motion trajectory set, force characteristic matrix, synchronous control sequence and state characteristic vector are subjected to data aggregation processing. Data standardization and feature combination methods are used to integrate various feature data into a unified format to generate an operating state parameter set and a multi-dimensional feature trajectory vector. The operating state parameter set contains various key parameters for equipment operation, and the multi-dimensional feature trajectory vector reflects the characteristic information of the motion trajectory.

[0042] For example, in a certain ore loading operation, when the spiral flat car machine loads ore, the laser ranging sensor collects the spatial coordinate data of 50 key points on the spiral guide rail at a sampling frequency of 100Hz. Through coordinate calibration and data filtering, the reference trajectory data is obtained, and the position deviation matrix with a maximum position deviation of 5mm is calculated. The data collected by the pressure sensor array shows that the pressure values ​​at different positions of the guide rail are distributed between 200-800N. After Fourier transformation of the data collected by the multi-point tension sensor network, the main frequency components of 2Hz, 4Hz and 6Hz are extracted, reflecting the periodic characteristics of material transportation. After filtering, the speed signal collected by the high-precision encoder shows that the angular velocity of the spiral guide rail fluctuates in the range of 15-20rad / s, and the phase difference between adjacent measuring points does not exceed 5 degrees. The signal collected by the vibration sensor is decomposed by 4 layers of wavelets to obtain the vibration characteristics of different frequency bands.

[0043] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0044] (1) dimensionality reduction processing of the operating state parameter set, generating a key state feature sequence and a motion reference matrix, feature decomposing the multi-dimensional feature trajectory vector, and outputting a reference trajectory set;

[0045] (2) Calculating the key state feature sequence by numerical differentiation to obtain the velocity change vector and acceleration matrix, performing singular value decomposition on the motion reference matrix, and determining the principal component feature set;

[0046] (3) extracting the spiral parameters of the reference trajectory set, calculating the pitch distribution sequence and the angle change matrix, analyzing the velocity change vector and the acceleration matrix according to dynamics, and deriving the motion characteristic vector;

[0047] (4) analyzing the correlation between the principal component feature set and the pitch distribution sequence, generating a feature correlation matrix and a time series mapping set, fusing the angle change matrix and the motion feature vector, and outputting a dynamic state sequence;

[0048] (5) Reconstructing the characteristic association matrix and the timing mapping set, calculating the spiral motion characteristic parameter set, predicting the dynamic state sequence, and obtaining the working condition prediction matrix.

[0049] Specifically, the operating state parameter set is subjected to data dimension reduction processing. The operating state parameter set contains various parameter data of equipment operation, such as guide rail position, pressure distribution, material tension and other information. The key variables in these parameters are extracted by principal component analysis method to form a key state feature sequence that characterizes the main characteristics of the system. At the same time, according to the motion law of the equipment, a motion reference matrix describing the standard motion state is established. The multidimensional feature trajectory vector is subjected to feature decomposition, and the complex motion trajectory data is decomposed into basic motion features to obtain a reference trajectory set containing standard motion trajectory information. When the key state feature sequence is subjected to numerical differentiation processing, the state change rate of adjacent time points is calculated by the difference method to obtain the speed change vector that describes the change of motion speed, and the acceleration matrix that reflects the motion acceleration characteristics is derived again. The motion reference matrix is ​​subjected to singular value decomposition processing to decompose the matrix into a left singular matrix, a singular value diagonal matrix and a right singular matrix, and the principal components with significant eigenvalues ​​are extracted to form a principal component feature set.

[0050] The reference trajectory set contains the basic characteristics of spiral motion. By extracting the geometric parameters of the spiral line and calculating the pitch values ​​at different positions, a pitch distribution sequence is formed. At the same time, the inclination changes of the spiral line at different positions are analyzed to obtain the angle change matrix. The velocity change vector and the acceleration matrix reflect the dynamic characteristics of the motion. Through force analysis and kinematic calculation, the motion feature vector that characterizes the motion characteristics is obtained. The principal component feature set and the pitch distribution sequence are subjected to correlation analysis, the correlation coefficients between the feature components are calculated, the feature association matrix is ​​constructed, and a time series mapping set is established according to the time series relationship. The angle change matrix and the motion feature vector are combined by weighted fusion method to comprehensively consider the geometric characteristics and dynamic characteristics to generate a dynamic state sequence that reflects the dynamic operation state of the system.

[0051] After data reconstruction, the feature association matrix and the time series mapping set are processed to reassemble the scattered feature information to form a complete set of spiral motion feature parameters. The time series prediction method is used for the dynamic state sequence to predict the future operating state based on the historical operating data and generate the working condition prediction matrix.

[0052] For example, when a spiral flat car machine is loading ore, the operating state parameter set contains 200 state parameters. Through principal component analysis, 20 main parameters with a contribution rate of 85% are selected to form a key state feature sequence. The feature sequence is numerically differentiated, and the speed change range is calculated to be between 0-2m / s, and the acceleration change range is between 0-0.5m / s². The parameters of the reference trajectory are extracted, and the pitch varies between 1000mm-1500mm and the spiral angle varies between 30-45 degrees. Through correlation analysis, the correlation coefficient matrix between the features is obtained, and the feature combination with an absolute value of the correlation coefficient greater than 0.8 indicates a significant correlation. Based on these data, the equipment state in the next working cycle is predicted to provide a basis for control decision-making. In the above process, the principal component analysis method is used for data dimensionality reduction. By calculating the eigenvalues ​​and eigenvectors, the principal components with a cumulative contribution rate reaching the preset threshold are selected. The central difference format is used for numerical differentiation to improve the calculation accuracy. Singular value decomposition is used to extract the main feature directions of the data. The Pearson correlation coefficient calculation method is used for correlation analysis. Time series forecasting uses an autoregressive moving average model to predict future states based on historical data. The combined use of these algorithms ensures the accuracy and reliability of data processing results.

[0053] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0054] (1) performing curve fitting on the spiral motion characteristic parameter set to generate a spiral guide rail motion trajectory and a regional segmentation sequence, processing the working condition prediction matrix according to a feature segmentation method, and outputting a partitioned working condition vector;

[0055] (2) dividing the spiral guide rail motion trajectory into equal arc lengths to obtain n candidate control regions and region boundary sets, processing the partition operating condition vectors based on data reorganization, and calculating the region state sequence;

[0056] (3) Processing the n candidate control regions using the region mapping principle to generate a region identification matrix and a control unit parameter set, and deriving a boundary feature vector by data matching the region boundary set and the region state sequence;

[0057] (4) outputting n adaptive data acquisition control units according to the area identification matrix and the control unit parameter set configuration parameters, processing the boundary feature vectors by using a state analysis method, and determining a partition monitoring sequence;

[0058] (5) Processing the partition monitoring sequence in a state combination manner to calculate a multi-dimensional state monitoring matrix, wherein the n adaptive data acquisition control units are sequentially distributed in the conveying area of ​​the spiral flat car machine along the spiral direction.

[0059] Specifically, the spiral motion characteristic parameter set is subjected to curve fitting. This parameter set contains the basic motion characteristics of the spiral guide rail, such as pitch, inclination and other parameters. These discrete feature points are fitted by the spline interpolation method to obtain a continuous spiral guide rail motion trajectory. At the same time, according to the change law of the motion characteristics, the trajectory is divided into multiple feature segments to form a regional segmentation sequence. The working condition prediction matrix contains the predicted data of the future operating state of the equipment, and the feature segmentation processing is performed on it. The predicted data is reorganized into partitioned working condition vectors according to the spatial position relationship. The equal arc length segmentation of the spiral guide rail motion trajectory is a key process. First, the total arc length of the spiral line is calculated, and then it is evenly divided into n equal parts, each corresponding to a candidate control area. The coordinate values ​​of the segmentation points constitute the regional boundary set, which is used to determine the specific range of each control area. The partitioned working condition vectors are reorganized according to the spatial correspondence, and the predicted working condition data are corresponded to each area to form a regional state sequence reflecting the operating status of each area.

[0060] The n candidate control regions are processed by region mapping to establish the correspondence between spatial positions and control units. Each region has a unique identification code, which constitutes a region identification matrix. At the same time, a control unit parameter set is generated based on the characteristic parameters of each region, such as position, length, curvature, etc. The region boundary set and the region state sequence are matched by data, and the boundary features are associated with the state data to form a boundary feature vector.

[0061] The regional identification matrix and control unit parameter set are used for regional configuration processing, and independent data acquisition control units are assigned to each region. These control units are responsible for collecting and processing the operating data in their respective regions. At the same time, the boundary feature vector is analyzed in state, the characteristic change law at the boundary is extracted, and the partition monitoring sequence is generated. Finally, the partition monitoring sequence is processed in state combination, and the monitoring data of each region is integrated into a multi-dimensional state monitoring matrix. The n adaptive data acquisition control units are distributed in sequence along the direction of the spiral guide rail, covering the entire transmission area, ensuring comprehensive monitoring of the equipment operation status.

[0062] For example, the total length of the guide rail of a spiral flat car machine is 20 meters, and 8 control areas are set. The spiral trajectory obtained by curve fitting contains 500 data points, which continuously records the spatial coordinate information of the guide rail. After equal arc length segmentation, the length of each control area is about 2.5 meters, and the coordinate values ​​of the area boundary points are recorded in the area boundary set. The working condition prediction matrix contains the operation prediction data for the next hour. After segmentation by area, each area obtains the prediction value of 120 time points. Through area mapping, a unique 4-digit identification code from 0001 to 0008 is assigned to each control area. Each control unit is equipped with data acquisition devices such as position sensors, pressure sensors, and vibration sensors. The multidimensional state monitoring matrix obtained after processing contains complete monitoring data of 8 areas at different time points. The key algorithms in the data processing process include: spline interpolation is used for curve fitting, and feature points are connected by cubic spline functions; equal arc length segmentation uses numerical integration method to calculate arc length; data matching uses the nearest neighbor algorithm to correspond boundary features to state data; state combination uses matrix splicing method to integrate data from each area into a unified format.

[0063] The partition control method decomposes the complex spiral guide rail system into several relatively independent but interrelated control areas. Each area has a dedicated control unit responsible for data collection and processing. Through the analysis of boundary features and the integration of status data, accurate monitoring and control of the operating status of the entire system is achieved. The coordinated work of each control unit ensures the smooth operation and precise control of the spiral flat car machine during the loading process.

[0064] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0065] (1) Processing the multidimensional state monitoring matrix by a matrix decomposition algorithm to obtain a monitoring data sequence and a state component matrix, analyzing the monitoring data sequence according to a time series, and calculating a state change vector;

[0066] (2) Execute principal element analysis on the state component matrix to generate a key parameter set and a characteristic weight vector, process the state change vector according to a trend analysis method, and output a change trend sequence;

[0067] (3) performing data clustering on the key parameter set and the feature weight vector, outputting a parameter grouping matrix and a feature label set, analyzing the change trend sequence with the aid of a threshold value, and determining an abnormal identification vector;

[0068] (4) processing the parameter grouping matrix and the feature marker set according to the data synthesis rule to obtain a working condition evaluation vector, processing the abnormal identification vector through a feature mapping algorithm, and deriving a deviation mapping matrix;

[0069] (5) Processing the deviation mapping matrix according to the data organization method and calculating the dynamic deviation feature set, wherein the operating condition evaluation vector includes the operating status data of each control unit of the spiral flat car machine.

[0070] Specifically, the multidimensional state monitoring matrix is ​​first subjected to matrix decomposition. The matrix contains the operating data collected by each monitoring point, such as position, pressure, vibration and other information. The non-negative matrix decomposition method is used to decompose the original matrix into the product of two low-dimensional matrices to obtain the monitoring data sequence and the state component matrix. The monitoring data sequence records the data changes of each monitoring point in chronological order. Through time series analysis methods such as moving average and exponential smoothing, the data change trend is extracted to form a state change vector. The state component matrix contains the data of different state components, and the key state parameters are extracted through principal element analysis. First, the variance contribution rate of each component is calculated, and the components with larger contribution rates are selected to form a key parameter set. At the same time, according to the importance of each parameter, the corresponding weight coefficient is calculated to form a characteristic weight vector. The state change vector is subjected to trend analysis, and linear regression and other methods are used to identify the change law of the data and generate a change trend sequence.

[0071] The key parameter set and feature weight vector are processed by data clustering. The K-means clustering algorithm is used to group similar parameters into a group to form a parameter grouping matrix. At the same time, an identifier is assigned to each parameter group to generate a feature tag set. The change trend sequence is analyzed by threshold value, and the upper and lower thresholds of the normal operating range are set. The data exceeding the threshold is marked to obtain the abnormal identification vector. The parameter grouping matrix and feature tag set are processed by data synthesis, and the characteristics and weights of each parameter group are comprehensively considered to generate a working condition evaluation vector that characterizes the overall operating status of the system. The abnormal identification vector is subjected to feature mapping to associate the abnormal state with the corresponding system deviation to form a deviation mapping matrix.

[0072] After data organization and processing, the deviation mapping matrix reorganizes various deviation features according to time sequence and spatial position to form a dynamic deviation feature set. The working condition evaluation vector contains the operating status data of each control unit of the spiral flat car machine, reflecting the operating conditions of the equipment at different locations and times.

[0073] For example, during the loading operation of a spiral flat car machine, the multidimensional state monitoring matrix contains data from 8 control areas, 10 monitoring parameters, and 600 time points. Through matrix decomposition, a 600×10 monitoring data sequence and a 10×8 state component matrix are obtained. Time series analysis shows that the pressure parameter fluctuates significantly at the 300-400 time point, and the vibration parameter exceeds the normal range at the 450-500 time point. The results of principal element analysis show that the variance contribution rates of the three parameters of pressure, vibration, and speed are 45%, 30%, and 15%, respectively, constituting the key parameter set, and the corresponding weight coefficients are 0.45, 0.3, and 0.15. Data clustering divides the 10 parameters into 3 groups: dynamic parameter group, kinematic parameter group, and state parameter group. Trend analysis found that the pressure parameter in the fourth area showed an upward trend with a slope of 0.05MPa / min, exceeding the preset threshold of 0.03MPa / min.

[0074] The comprehensive analysis results show that the pressure parameter anomaly mainly appears in the fourth area of ​​the 300-400 time period, resulting in a 20% decrease in the working condition assessment score. Through feature mapping, it was determined that the anomaly was related to the uneven distribution of loading materials, and the deviation value was 1.5 times the standard value. This information was integrated into the dynamic deviation feature set to provide a basis for subsequent control adjustments.

[0075] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0076] (1) Standardizing the operating condition evaluation vector to generate an evaluation data matrix and a feature distribution sequence, processing the dynamic deviation feature set through a data reconstruction algorithm, and outputting a deviation component vector group;

[0077] (2) processing the evaluation data matrix according to the time series coding rule to obtain the time series feature network and the state coding matrix, calculating the deviation component vector group based on feature embedding, and generating the deviation mapping sequence and the weight parameter set;

[0078] (3) recursively calculating the time series feature network and the state encoding matrix, deriving the hidden layer state sequence and the feature transfer matrix, processing the deviation mapping sequence and the weight parameter set in combination with a parameter combination method, and obtaining a compensation parameter sequence;

[0079] (4) transferring the state of the hidden layer state sequence in the long short-term memory unit, calculating the candidate instruction sequence and the control reference matrix, calculating the gradient of the feature transfer matrix and the compensation parameter sequence according to the back propagation principle, and outputting the parameter update vector;

[0080] (5) fusing the candidate instruction sequence and the parameter update vector by a gated loop unit to generate a control instruction matrix, processing the control reference matrix and the control instruction matrix according to rule mapping, and determining a real-time feedback control instruction set;

[0081] (6) Analyze the response characteristics of the control instruction matrix, calculate the response feature sequence and the compensation vector group, compensate the response feature sequence and the compensation vector group through the attention mechanism, and output the dynamic compensation parameter matrix.

[0082] Specifically, the working condition evaluation vector is subjected to data standardization. The working condition evaluation vector contains parameters of various dimensions, such as pressure (MPa), speed (m / s), vibration (Hz), etc. Through standardization, all parameters are mapped to the same numerical range for easy subsequent processing. The standardized data forms an evaluation data matrix, and the distribution characteristics of each parameter are extracted to generate a feature distribution sequence. After data reconstruction, the dynamic deviation feature set reorganizes different types of deviation data according to the feature type to form multiple groups of deviation component vectors. The evaluation data matrix contains time series information, and the information of the time dimension is encoded into a feature representation through time series coding. The encoding process uses a sliding window method to convert continuous time series data into a feature network structure to obtain a time series feature network and a state coding matrix. After feature embedding, the deviation component vector group maps the high-dimensional deviation data to a low-dimensional feature space to obtain a deviation mapping sequence, and generates a corresponding weight parameter set to adjust the importance of different deviation features. The time series feature network and the state coding matrix are recursively calculated through a recursive neural network to extract the dynamic characteristics of the time series data, form a hidden state sequence, and record the law of state transition. The relationship between different features is described by the feature transfer matrix. The deviation mapping sequence and the weight parameter set are combined through parameters, and the deviation features are weighted and combined according to the weight relationship to form a compensation parameter sequence.

[0083] The long short-term memory unit (LSTM) is used to process the hidden state sequence. The LSTM unit contains an input gate, a forget gate, and an output gate, and can learn long-term dependencies. Through the state transfer of the LSTM, the candidate instruction sequence and the control reference matrix are generated. The back propagation algorithm calculates the gradient of the feature transfer matrix and the compensation parameter sequence, updates the network parameters, and obtains the parameter update vector.

[0084] The gated recurrent unit (GRU) performs data fusion on the candidate instruction sequence and parameter update vector. GRU simplifies the structure of LSTM and only contains update gates and reset gates. It controls the flow of information through the gating mechanism and obtains the control instruction matrix after fusion. The control reference matrix and the control instruction matrix are mapped by rules to convert abstract control instructions into specific operation instructions to form a real-time feedback control instruction set.

[0085] After response analysis, the control instruction matrix evaluates the execution effect of the control instruction and obtains the response feature sequence. The compensation vector group contains the parameter information to be compensated, and the response feature sequence and the compensation vector group are dynamically compensated through the attention mechanism. The attention mechanism calculates the attention weight, highlights the influence of important features, and finally generates a dynamic compensation parameter matrix.

[0086] For example, when a spiral flat car is loading ore, the working condition evaluation vector contains 10 parameters such as pressure, speed, and vibration in 8 control areas. Through standardization, all parameter values ​​are mapped to the interval [-1,1]. The time series coding uses a sliding window of length 20 to encode the data of 600 time points. The recursive neural network contains 128 hidden layer nodes, and the hidden layer dimensions of the LSTM unit and the GRU unit are set to 64. During the actual operation, the pressure in the third area is abnormal, and the standardized pressure value increases from 0.3 to 0.8. The time series coding captures the characteristics of rapid pressure increase, and the LSTM unit generates a candidate instruction to reduce the feeding speed. The GRU unit fuses the speed adjustment instruction with the pressure compensation parameter to generate a control instruction to reduce the feeding amount by 20%. The attention mechanism recognizes that the pressure abnormality is the key issue, assigns an attention weight of 0.6, and generates the corresponding compensation parameters.

[0087] After the control command was executed, the pressure value gradually decreased to the normal range of 0.3-0.4 within 50 time points. The response analysis showed that the control effect was good, and the compensation parameter matrix recorded the parameter adjustment process of this abnormal processing. The whole process demonstrated the complete control chain from data processing, feature extraction, command generation to dynamic compensation, realizing the intelligent control of the spiral flat car machine.

[0088] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0089] (1) parsing the real-time feedback control instruction set, generating a control instruction sequence and an instruction identification matrix, processing the dynamic compensation parameter matrix according to a parameter decomposition method, and outputting a compensation parameter vector group;

[0090] (2) assigning the control instruction sequence and the instruction identification matrix, calculating n control instruction subsets, processing the compensation parameter vector group according to the area mapping, and obtaining a compensation parameter distribution matrix;

[0091] (3) matching and mapping the n control instruction subsets with the compensation parameter distribution matrix to generate a regional control sequence and a compensation instruction vector, analyzing the regional control sequence based on timing, and deriving a control timing matrix;

[0092] (4) Coordinating the synchronization relationship between the compensation instruction vector and the control timing matrix, calculating the real-time control sequence and the parameter correction matrix, detecting the state of the real-time control sequence, and outputting the real-time control data;

[0093] (5) Extracting the characteristic correlation of the parameter correction matrix to generate a process characteristic mapping set, wherein the n adaptive data acquisition control units respectively receive corresponding control instruction subsets for control processing.

[0094] Specifically, each control instruction is decoded and classified to form a control instruction sequence containing specific operation content, and a unique identification code is generated for each instruction to construct an instruction identification matrix. At the same time, the dynamic compensation parameter matrix is ​​parameter decomposed to separate different types of compensation parameters to form multiple groups of compensation parameter vectors. The control instruction sequence is allocated according to spatial position and function type, and the instructions are allocated to n control areas, and each area obtains a control instruction subset. The compensation parameter vector group is mapped by area, and the corresponding relationship between the parameters and the control area is established to form a compensation parameter distribution matrix.

[0095] n control instruction subsets and compensation parameter distribution matrix are processed for regional matching. The matching process uses the following formula: in: represents the jth matching result of the i-th region, represents the kth instruction component of the i-th control instruction subset, represents the kth compensation parameter of the jth region, represents the control weight of the ith region, represents the priority of the jth compensation parameter, , is the balance coefficient, p is the parameter dimension, Represents a tensor product operation.

[0096] The matching results form the regional control sequence and compensation instruction vector, and the control timing matrix is ​​obtained through timing analysis. The compensation instruction vector and the control timing matrix are synchronized and coordinated to achieve the timing synchronization of instruction execution, and generate the real-time control sequence and parameter correction matrix. The real-time control sequence is tested for status, the control execution is monitored, and real-time control data is formed.

[0097] The parameter correction matrix performs feature association processing, extracts feature information in the control process, and forms a process feature mapping set. The n adaptive data acquisition control units receive the corresponding control instruction subsets and execute specific control operations.

[0098] For example, a spiral flat car machine has 8 control areas and receives a real-time feedback control instruction set containing 100 instructions. Instruction parsing converts these instructions into specific operating parameters, such as speed adjustment value, pressure correction value, etc. Each instruction is equipped with a 4-digit digital identification code, forming an 8×100 instruction identification matrix. The dynamic compensation parameter matrix is ​​decomposed to obtain three groups of parameter vectors for speed compensation, pressure compensation, and position compensation. The 100 control instructions are assigned to 8 areas according to function and position, and each area receives 12-13 instructions. The compensation parameters are mapped by area to form an 8×3 compensation parameter distribution matrix.

[0099] Calculated by the area matching formula, when =0.6, =0.4, the matching degree between the speed control instruction and the speed compensation parameter of the third area reached 0.85, and the speed adjustment amount was determined to be -0.2m / s. Timing analysis showed that the execution interval of the control instruction was 100ms, forming an 8×100 control timing matrix. Synchronous coordination ensured that the control actions of the eight areas were carried out in coordination, and the control status every 50ms was recorded in the real-time control sequence. The parameter correction matrix reflects the parameter adjustment during the control process, such as the speed of the third area was smoothly adjusted within 5s.

[0100] The above describes the automatic operation control method of the spiral flat lathe in the embodiment of the present application. The following describes the automatic operation control system of the spiral flat lathe in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the automatic operation control system of the spiral flat car machine includes:

[0101] The acquisition module is used to collect working state parameters from the spiral flat car machine and generate an operating state parameter set and a multi-dimensional feature trajectory vector;

[0102] A calculation module, used for calculating the operating state parameter set and the multi-dimensional characteristic trajectory vector, and outputting a spiral motion characteristic parameter set and an operating condition prediction matrix;

[0103] A processing module, for dividing the spiral flat car machine transmission area into n adaptive data acquisition control units according to the spiral motion characteristic parameter set and the working condition prediction matrix, and generating a multi-dimensional state monitoring matrix, wherein n is a positive integer greater than or equal to 5;

[0104] An extraction module is used to extract characteristic data of the multidimensional condition monitoring matrix and calculate a working condition evaluation vector and a dynamic deviation characteristic set;

[0105] An optimization module, used for iteratively optimizing the operating condition evaluation vector and the dynamic deviation feature set through a recursive neural network to generate a real-time feedback control instruction set and a dynamic compensation parameter matrix;

[0106] The control module is used for executing intelligent closed-loop control according to the real-time feedback control instruction set and the dynamic compensation parameter matrix by n adaptive data acquisition control units, and outputting real-time control data and process characteristic mapping sets.

[0107] Through the cooperation of the above components, the operating state parameter set and multi-dimensional characteristic trajectory vector are obtained by multi-dimensional data acquisition and processing of the working state, and the comprehensive monitoring of the operating state of the spiral flat car machine is realized, providing a reliable data basis for subsequent control. In addition, the spiral motion characteristic parameter set and the working condition prediction matrix are obtained by numerical calculation and processing of the operating state parameter set and the multi-dimensional characteristic trajectory vector, and the motion characteristics of the equipment and the future working condition change trend are accurately grasped. According to the spiral motion characteristic parameter set and the working condition prediction matrix, the transmission area is divided into n equal parts, n adaptive data acquisition control units are set, and a multi-dimensional state monitoring matrix is ​​obtained, realizing the refined zoning control of the transmission area of ​​the spiral flat car machine. Feature extraction processing is performed on the multi-dimensional state monitoring matrix to obtain the working condition evaluation vector and the dynamic deviation feature set, which provides a basis for the identification and processing of abnormal working conditions. The working condition evaluation vector and the dynamic deviation feature set are iteratively optimized by the recursive neural network to obtain the real-time feedback control instruction set and the dynamic compensation parameter matrix, which improves the generation efficiency and accuracy of the control instruction. Finally, based on n adaptive data acquisition control units, intelligent closed-loop control is performed according to the real-time feedback control instruction set and dynamic compensation parameter matrix to obtain real-time control data and process feature mapping sets, thus realizing precise control and dynamic compensation of the spiral flat car machine. The entire control process forms a complete closed loop, from data acquisition, feature extraction, instruction generation to control execution, and each link is closely connected, ensuring the stability and reliability of the loading process, and effectively solving the problems of response lag and low control accuracy in traditional control methods.

[0108] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for controlling the automatic operation of a spiral flat lathe, characterized in that: The automatic operation control method of the spiral flat lathe comprises: Collect working state parameters from the spiral flat car machine to generate operating state parameter set and multi-dimensional feature trajectory vector; Calculating the operating state parameter set and the multi-dimensional characteristic trajectory vector, and outputting the spiral motion characteristic parameter set and the operating condition prediction matrix; According to the spiral motion characteristic parameter set and the working condition prediction matrix, the spiral flat car machine transmission area is divided into n adaptive data acquisition control units according to the spiral direction, and a multi-dimensional state monitoring matrix is ​​generated, wherein n is a positive integer greater than or equal to 5; Extracting characteristic data of the multidimensional condition monitoring matrix, and calculating a working condition evaluation vector and a dynamic deviation characteristic set; Iteratively optimizing the operating condition evaluation vector and the dynamic deviation feature set through a recursive neural network to generate a real-time feedback control instruction set and a dynamic compensation parameter matrix; N adaptive data acquisition control units execute intelligent closed-loop control according to the real-time feedback control instruction set and the dynamic compensation parameter matrix, and output real-time control data and a process characteristic mapping set.

2. The automatic operation control method of spiral flat lathe according to claim 1 is characterized in that: The method of collecting working state parameters from the spiral flat car machine and generating an operating state parameter set and a multi-dimensional feature trajectory vector includes: Collect the three-dimensional coordinates of the spiral guide rail's spatial position, calculate the spiral track reference data and position deviation matrix, synchronously read the real-time data of the presser foot height sensor array, and output the pressure distribution vector; Reading multi-point tension sensor network data, generating a tension distribution curve and edge profile data, analyzing the edge profile data using Fourier transform, and obtaining a characteristic frequency set; Perform filtering operations on high-precision encoder speed signals, derive spiral motion angular velocity matrix and phase difference data, process mechanical vibration sensor signals using wavelet decomposition, and calculate multi-scale vibration feature sequences; The spiral trajectory reference data and the position deviation matrix are integrated to generate a spatial motion trajectory set, the pressure distribution vector and the tension distribution curve are processed by tensor decomposition, and a force characteristic matrix is ​​output; Analyze the cross-correlation between the characteristic frequency set and the spiral motion angular velocity matrix to determine a synchronization control sequence, and derive a state characteristic vector based on the time-frequency analysis of the multi-scale vibration characteristic sequence and the phase difference data; The spatial motion trajectory set, the force characteristic matrix, the synchronous control sequence and the state characteristic vector are aggregated to calculate an operation state parameter set and a multi-dimensional characteristic trajectory vector.

3. The automatic operation control method of spiral flat lathe according to claim 1 is characterized in that: The step of calculating the operating state parameter set and the multi-dimensional characteristic trajectory vector and outputting the spiral motion characteristic parameter set and the operating condition prediction matrix includes: Dimensionality reduction processing is performed on the operating state parameter set to generate a key state feature sequence and a motion reference matrix, feature decompose the multi-dimensional feature trajectory vector, and output a reference trajectory set; Calculate the key state feature sequence by numerical differentiation to obtain the velocity change vector and acceleration matrix, perform singular value decomposition on the motion reference matrix, and determine the principal component feature set; Extracting the spiral parameters of the reference trajectory set, calculating the pitch distribution sequence and the angle change matrix, analyzing the velocity change vector and the acceleration matrix according to dynamics, and deriving the motion characteristic vector; Analyze the correlation between the principal component feature set and the pitch distribution sequence, generate a feature association matrix and a time series mapping set, fuse the angle change matrix and the motion feature vector, and output a dynamic state sequence; The characteristic association matrix and the timing mapping set are reconstructed, the characteristic parameter set of the spiral motion is calculated, the dynamic state sequence is predicted, and the working condition prediction matrix is ​​obtained.

4. The automatic operation control method of spiral flat lathe according to claim 1 is characterized in that: According to the spiral motion characteristic parameter set and the working condition prediction matrix, the spiral flat car machine transmission area is divided into n adaptive data acquisition control units according to the spiral direction, and a multi-dimensional state monitoring matrix is ​​generated, wherein n is a positive integer greater than or equal to 5, including: Performing curve fitting on the spiral motion characteristic parameter set to generate a spiral guide rail motion trajectory and a regional segmentation sequence, processing the working condition prediction matrix according to a feature segmentation method, and outputting a partitioned working condition vector; Divide the spiral guide rail motion trajectory into equal arc lengths to obtain n candidate control regions and region boundary sets, process the partition operating condition vectors based on data reorganization, and calculate the region state sequence; Processing the n candidate control regions using the region mapping principle to generate a region identification matrix and a control unit parameter set, and deriving a boundary feature vector by data matching the region boundary set and the region state sequence; Outputting n adaptive data acquisition control units according to the area identification matrix and the control unit parameter set configuration parameters, processing the boundary feature vectors using a state analysis method, and determining a partition monitoring sequence; The partition monitoring sequence is processed in a state combination manner to calculate a multi-dimensional state monitoring matrix, wherein the n adaptive data acquisition control units are sequentially distributed in the conveying area of ​​the spiral flat car machine along the spiral direction.

5. The automatic operation control method of spiral flat lathe according to claim 4 is characterized in that: The step of extracting characteristic data of the multidimensional condition monitoring matrix and calculating a working condition evaluation vector and a dynamic deviation feature set includes: Processing the multidimensional state monitoring matrix by a matrix decomposition algorithm to obtain a monitoring data sequence and a state component matrix, analyzing the monitoring data sequence according to a time series, and calculating a state change vector; Execute principal element analysis on the state component matrix to generate a key parameter set and a characteristic weight vector, process the state change vector according to a trend analysis method, and output a change trend sequence; Implementing data clustering on the key parameter set and the feature weight vector, outputting a parameter grouping matrix and a feature tag set, analyzing the change trend sequence with the help of a threshold value, and determining an abnormal identification vector; Process the parameter grouping matrix and the feature marker set according to the data synthesis rule to obtain a working condition evaluation vector, process the abnormal identification vector through a feature mapping algorithm, and derive a deviation mapping matrix; The deviation mapping matrix is ​​processed in a data organization manner to calculate a dynamic deviation feature set, wherein the operating condition evaluation vector includes operating status data of each control unit of the spiral flat car machine.

6. The automatic operation control method of spiral flat lathe according to claim 5 is characterized in that: The iterative optimization of the operating condition evaluation vector and the dynamic deviation feature set by a recursive neural network to generate a real-time feedback control instruction set and a dynamic compensation parameter matrix includes: The working condition evaluation vector is standardized to generate an evaluation data matrix and a feature distribution sequence, the dynamic deviation feature set is processed by a data reconstruction algorithm, and a deviation component vector group is output; Process the evaluation data matrix according to the time series coding rule to obtain a time series feature network and a state coding matrix, calculate the deviation component vector group based on feature embedding, and generate a deviation mapping sequence and a weight parameter set; Recursively calculating the time series feature network and the state encoding matrix, deriving the hidden layer state sequence and the feature transfer matrix, processing the deviation mapping sequence and the weight parameter set in combination with a parameter combination method, and obtaining a compensation parameter sequence; Transferring the state of the hidden layer state sequence in the long short-term memory unit, calculating the candidate instruction sequence and the control reference matrix, calculating the gradient of the feature transfer matrix and the compensation parameter sequence according to the back propagation principle, and outputting the parameter update vector; The candidate instruction sequence and the parameter update vector are fused by a gated loop unit to generate a control instruction matrix, and the control reference matrix and the control instruction matrix are processed according to rule mapping to determine a real-time feedback control instruction set; Analyze the response characteristics of the control instruction matrix, calculate the response feature sequence and the compensation vector group, compensate the response feature sequence and the compensation vector group through the attention mechanism, and output the dynamic compensation parameter matrix.

7. The automatic operation control method of spiral flat lathe according to claim 6 is characterized in that: The n adaptive data acquisition control units perform intelligent closed-loop control according to the real-time feedback control instruction set and the dynamic compensation parameter matrix, and output real-time control data and a process feature mapping set, including: Parsing the real-time feedback control instruction set, generating a control instruction sequence and an instruction identification matrix, processing the dynamic compensation parameter matrix according to a parameter decomposition method, and outputting a compensation parameter vector group; Assigning the control instruction sequence and the instruction identification matrix tasks, calculating n control instruction subsets, processing the compensation parameter vector group according to the area mapping, and obtaining a compensation parameter distribution matrix; Match and map the n control instruction subsets with the compensation parameter distribution matrix to generate a regional control sequence and a compensation instruction vector, analyze the regional control sequence based on the timing sequence, and derive a control timing matrix; Coordinate the synchronization relationship between the compensation instruction vector and the control timing matrix, calculate the real-time control sequence and the parameter correction matrix, detect the state of the real-time control sequence, and output real-time control data; The characteristic relevance of the parameter correction matrix is ​​extracted to generate a process characteristic mapping set, wherein the n adaptive data acquisition control units respectively receive corresponding control instruction subsets for control processing.

8. A spiral flat lathe automatic operation control system, used to implement the spiral flat lathe automatic operation control method according to any one of claims 1 to 7, characterized in that: The automatic operation control system of the spiral flat lathe comprises: The acquisition module is used to collect working state parameters from the spiral flat lathe and generate an operating state parameter set and a multi-dimensional feature trajectory vector; A calculation module, used for calculating the operating state parameter set and the multi-dimensional characteristic trajectory vector, and outputting a spiral motion characteristic parameter set and an operating condition prediction matrix; A processing module, for dividing the spiral flat car machine transmission area into n adaptive data acquisition control units according to the spiral motion characteristic parameter set and the working condition prediction matrix, and generating a multi-dimensional state monitoring matrix, wherein n is a positive integer greater than or equal to 5; An extraction module, used to extract characteristic data of the multidimensional condition monitoring matrix, and calculate a working condition evaluation vector and a dynamic deviation characteristic set; An optimization module, used for iteratively optimizing the operating condition evaluation vector and the dynamic deviation feature set through a recursive neural network to generate a real-time feedback control instruction set and a dynamic compensation parameter matrix; The control module is used for executing intelligent closed-loop control according to the real-time feedback control instruction set and the dynamic compensation parameter matrix by n adaptive data acquisition control units, and outputting real-time control data and process characteristic mapping sets.

Images