Data-driven milling intelligent analysis device, method and equipment and storage medium

Abstract

The invention provides a data-driven milling intelligent analysis device, method and equipment and a storage medium, particularly relates to the technical field of intelligent manufacturing, and provides an intelligent analysis software architecture in a component visual arrangement form and a construction method for process data under various processes, workpieces, procedures, instructions and positions in milling. The overall framework provides an integrated milling intelligent analysis scheme covering processing data acquisition, processing, analysis and application in the form of application layer data set and algorithm component dragging, and the sensing layer obtains processing process data of internal and external systems of a machine tool; the data access layer is responsible for data storage, retrieval and access; the business logic layer is responsible for executing functions such as data set management, algorithm management and intelligent analysis of workflow; the application layer provides a user interface corresponding to a series of functions such as analysis process construction based on component dragging, so that the threshold of artificial intelligence enabling milling machining is lowered, and a full link of data acquisition, storage, analysis and application is opened.

Description

Technical Field

[0001] The present application relates to the field of intelligent manufacturing technology, and in particular to a data-driven intelligent analysis device, method, equipment and storage medium for milling processing. Background Art

[0002] With the continuous advancement of intelligent manufacturing, the manufacturing industry's demand for data analysis and the in-depth application of artificial intelligence algorithms are increasing. Artificial intelligence algorithms involve multiple disciplines. Due to factors such as the rapid iteration of algorithm updates, the transition from traditional manufacturing professionals to intelligent manufacturing professionals is difficult. Mechanical engineers need to master machine learning frameworks and programming tools, which prolongs the algorithm iteration cycle. Furthermore, in the traditional intelligent analysis of milling processing data, each step—from data collection, storage, and transmission to AI engineers reading stored files, data processing, visualization analysis, model building, algorithm training, and application—is relatively independent and heavily reliant on manual operations. A large number of processing data processing methods and intelligent analysis algorithms lack systematic management, process databases and algorithm libraries lack systematic packaging, and the reuse rate of key algorithm modules is low. These obstacles hinder the convenient application of algorithms by manufacturing personnel, hindering the rapid implementation of core scenarios such as surface roughness prediction and chatter warning.

[0003] Among them, milling processing and manufacturing is an important organizational part of intelligent manufacturing. The existing manual data processing and analysis methods can no longer meet the needs of convenient analysis of massive milling process data. It is urgent to build a base code intelligent analysis system for manufacturing engineers. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a data-driven intelligent analysis device, method, equipment and storage medium for milling processing. To solve the above problems, the present invention constructs a complete intelligent analysis process by visually arranging the data set components and algorithm components of the software application layer to achieve the connection of the overall process of milling processing data collection, processing, targeted analysis and application. This method provides a new idea for overcoming the technical barriers faced by technicians in the field of mechanical processing in terms of interdisciplinary data processing, algorithm development and industrial software integrated application, and can support their rapid application of artificial intelligence algorithms to achieve goals such as surface roughness prediction, anomaly monitoring, and process parameter optimization, thereby improving the work efficiency and quality control level in the processing process. This method uses industrial software as a carrier, integrates milling processing data collection, and encapsulates data processing methods and analysis algorithms as basic components. In a componentized and process-based form, it forms the accumulation, management and convenient reuse of intelligent analysis databases and algorithm libraries. This provides manufacturing personnel with a convenient thin-walled part milling processing data analysis platform, effectively improving their processing and analysis capabilities of processing data in an intelligent manufacturing environment.

[0005] In a first aspect, the present application provides a data-driven intelligent analysis device for milling processing, the device comprising: The perception layer is used to collect milling process data from the CNC system inside the CNC machine tool and milling process data from external sensors; Data access layer, used to store, retrieve and access structured and semi-structured data; The business logic layer is used to perform basic data management, preprocessing, hierarchical organization, data set management, algorithm management, and intelligent analysis workflows for the milling process. It processes, analyzes, and visualizes data at different data scales, including process level, operation level, workpiece level, instruction set, and position level. It is divided into a basic data management module, a data processing module, and an intelligent analysis module. The application layer is used to provide the data set management interface, algorithm management interface, intelligent analysis interface, data visualization interface and data foundation management interface.

[0006] Preferably, the milling process data of the CNC system inside the CNC machine tool includes acquisition time, spindle current, spindle load, feed axis current, feed axis load, processing workpiece number, process number, instruction line number, mechanical coordinates of the feed axis in the X-axis direction, mechanical coordinates of the feed axis in the Y-axis direction, mechanical coordinates of the feed axis in the Z-axis direction, rotation arc of the rotation axis in the A-axis direction, and rotation arc of the rotation axis in the C-axis direction; The external sensor milling process data includes acquisition time, spindle vibration signal in the X-axis direction, spindle vibration signal in the Y-axis direction, spindle vibration signal in the Z-axis direction, spindle force signal in the X-axis direction, spindle force signal in the Y-axis direction, and spindle force signal in the Z-axis direction.

[0007] Preferably, the structured data includes the CNC system milling process data of the CNC machine tool, the external sensor milling process data, processing arrangement information, data set component information, algorithm component information, and intelligent analysis process component information stored in the database; The processing arrangement information includes the processing arrangement information number, the processing workpiece number list, the processing number, the processed number, the corresponding processing workpiece NC program file storage path and the workpiece quality inspection file storage path; The dataset component information includes the dataset component number, a list of artifact numbers involved, the signal type contained, the sample sequence length, the training set, the test set division ratio, and the file storage path; The algorithm component information includes the algorithm component number, the algorithm type involved, the training rounds, the loss function used, the input parameters and the algorithm script storage file path; The intelligent analysis process component information includes node information and edge information for storing the roughness prediction intelligent analysis process; wherein, the node information includes: node number, data or algorithm component number, node type, execution order, and storage path of processed data. The execution order is from small to large, and nodes with smaller values ​​are executed first, and the data is uniformly stored in an array format. The edge information includes: edge number, source node number, and target node number.

[0008] Preferably, the semi-structured data includes the numerical control program files required for the operation of each workpiece, files stored in the data array processed by the data set component, surface storage degree detection result files, contour error detection result files and tool wear result detection files.

[0009] Preferably, the data basic management module is used to add, delete, modify or query the specific business logic of processing arrangement information, data set information and algorithm information functions, and also provides paging and sorting functions.

[0010] The data processing module includes a data preprocessing submodule and a data hierarchical organization submodule. The data processing module is used to organize the data into a form applicable to deep learning analysis algorithms and visualization; The data preprocessing submodule is used to call the data access layer data to obtain signal data according to the data set component information, and provide operations such as null value processing, outlier processing, signal denoising, and signal downsampling; The data hierarchical organization submodule is used to filter the data corresponding to the process, workpiece, process, instruction, and position segment required by the user based on the processing arrangement number, workpiece number, process number, instruction line number, and processing coordinates selected by the user, and divide the entire sequence into several subsequences according to the subsequence length set by the user; and store the data in a file format; The intelligent analysis module is used to call the data files and corresponding algorithm scripts required by the components in sequence according to the execution order of the components in the intelligent analysis process; wherein, the intelligent analysis process is constructed by the user by dragging and dropping data set components and algorithm components.

[0011] Preferably, the data set management interface is used to create, modify, delete and retrieve the data set component information; The algorithm management interface is used to create, modify, delete and retrieve the algorithm component information; The intelligent analysis interface is used to graphically arrange and monitor the milling processing data analysis process in real time; The data visualization interface is used to visualize the timing signals according to the user's specific process, workpiece, process, instruction and position visualization requirements; The user interface corresponding to the data basic management is used to create, modify, delete and retrieve the processing arrangement information.

[0012] The data-driven milling intelligent analysis device provided by this application brings the following beneficial effects: The present application provides a data-driven intelligent analysis device for milling processing, which is used to split the overall process of intelligent analysis of thin-walled parts milling processing data into several steps, and encapsulate the data operations required for each step into independent components, so that according to different analysis requirements, the required components can be called to build a specific analysis process. At the same time, the system can continuously encapsulate components according to new data and requirements to achieve the expansion of the component library. The device covers functions such as data acquisition, data organization of different scales and forms of expression, intelligent analysis and visualization. At the same time, compared with the traditional time domain signal analysis software architecture, it additionally expands the time domain analysis of different data scales.

[0013] In a second aspect, the present application further provides a data-driven intelligent analysis method for milling processing, the method comprising: The perception layer collects milling process data from the CNC system inside the CNC machine tool and milling process data from external sensors; The data access layer stores, retrieves and accesses structured and semi-structured data; The business logic layer performs basic data management, preprocessing, hierarchical organization, data set management, algorithm management, and intelligent analysis workflows for the milling process. It processes, analyzes, and visualizes data at different data scales, including process level, operation level, workpiece level, instruction set, and position level. It is divided into a basic data management module, a data processing module, and an intelligent analysis module. The application layer provides a data set management interface, an algorithm management interface, an intelligent analysis interface, a data visualization interface, and a data foundation management interface.

[0014] The data-driven intelligent analysis method for milling processing provided in the embodiment of the present application has the same technical features as the data-driven intelligent analysis device for milling processing provided in the above embodiment, so it can also solve the same technical problems and achieve the same technical effects.

[0015] In a third aspect, the present application provides a computing device, including a memory and a processor; One or more computer programs are stored in the memory, and the one or more computer programs include instructions; when the instructions are executed by the processor, the computing device executes the method as described in any one of the first aspects.

[0016] In a fourth aspect, the present application provides a computer-readable storage medium for storing a computer program for executing the method as described in any one of the first aspects.

[0017] In a fifth aspect, the present application provides a computer program product, which includes one or more computer instructions. When the computer instructions are executed by a computer, the computer executes the method as described in any one of the first aspects.

[0018] Other features and advantages of the present application will be described in the following description, and in part will become apparent from the description, or understood by practicing the present application. The purpose and other advantages of the present application are realized and obtained by the structures particularly pointed out in the description and drawings.

[0019] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0021] Figure 1 A schematic structural diagram of a data-driven milling intelligent analysis device provided in an embodiment of the present application; Figure 2 A schematic flow chart of a data-driven intelligent analysis method for milling provided in an embodiment of the present application; Figure 3 A flow chart of another data-driven intelligent analysis method for milling provided in an embodiment of the present application; Figure 4 A schematic diagram of milling processing data modeling and intelligent analysis provided in an embodiment of the present application; Figure 5 A schematic structural diagram of another data-driven milling intelligent analysis device provided in an embodiment of the present application; Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0023] To facilitate understanding of this embodiment, the embodiments of this application are described in detail below.

[0024] The embodiment of the present application provides a data-driven intelligent analysis device for milling processing, such as Figure 1 As shown, Figure 1 This is a schematic diagram of the structure of a data-driven milling intelligent analysis device provided in an embodiment of the present application. The device includes: The perception layer 200 is used to collect milling process data of the CNC system inside the CNC machine tool and milling process data of the external sensor; Data access layer 207, used to store, retrieve and access structured data and semi-structured data; Specifically, the collected structured data is organized in the form of a structure at the perception layer, where the general structure Data_Struct for various types of data can be defined as Data_Struct={id, property1, property2, …,propertyn, subData1, subData2, …, subDatam}, where id is the unique identifier of the data structure of each structured data, property1 is the data attribute of the first structured data, property2 is the data attribute of the second structured data, propertyn is the data attribute of the nth structured data, and subDatam is the sub-data structure of the data structure of the mth structured data.

[0025] The structured data is persistently stored in a relational database data table. The relational database data table is represented by DataTable = {PK_id, Field1, Field2, …, Fieldn, FK_id1, FK_id2, …, FK_idm}, where PK_id is the primary key of the relational database data table. Each PK_id corresponds to the unique identifier id of each structured data structure (having a first mapping relationship). Field1 is each field corresponding to the first relational database data table, Field2 is each field corresponding to the second relational database data table, Fieldn is each field corresponding to the nth relational database data table, and the nth field Fieldn in each relational database data table corresponds to the data attribute propertyyn of the nth structured data (having a second mapping relationship). FK_idm is the mth associated foreign key, and the mth associated foreign key FK_idm corresponds to a sub-data structure of the mth structured data structure (having a third mapping relationship). In addition, the above semi-structured data is stored in XML, CSV, JSON and other format files.

[0026] The business logic layer 217 is used to perform basic data management, preprocessing, hierarchical organization, data set management, algorithm management, and intelligent analysis workflows for the milling process. It processes, analyzes, and visualizes data at different data scales, including the process level, operation level, workpiece level, instruction set, and position level. It is divided into a basic data management module, a data processing module, and an intelligent analysis module. The application layer 221 is used to provide a data set management interface, an algorithm management interface, an intelligent analysis interface, a data visualization interface, and a data foundation management interface.

[0027] Preferably, the milling process data of the CNC system inside the CNC machine tool includes acquisition time, spindle current, spindle load, feed axis current, feed axis load, machining workpiece number, process number, instruction line number, mechanical coordinates of the feed axis in the X-axis direction, mechanical coordinates of the feed axis in the Y-axis direction, mechanical coordinates of the feed axis in the Z-axis direction, rotation arc of the rotary axis in the A-axis direction, and rotation arc of the rotary axis in the C-axis direction; The structure Machine_internal_data corresponding to the milling process data of the CNC system within the CNC machine tool can be defined as Struct Machine_internal_data={mid_id; acq_time; axis_info; program_info; pos_info}, where mid_id is the unique identifier of the data structure of the milling process data of the CNC system within the CNC machine tool, acq_time is the acquisition time of the milling process data of the CNC system within the CNC machine tool, axis_info is a substructure that stores data such as the spindle current, spindle load, feed axis current, and feed axis load, program_info is a substructure that stores CNC program running information such as the workpiece number, process number, and instruction line number, and loc_info is a substructure that stores processing position information such as the mechanical coordinates of the feed axis in the X-axis direction, the mechanical coordinates of the feed axis in the Y-axis direction, the mechanical coordinates of the feed axis in the Z-axis direction, the rotation arc of the rotary axis in the A-axis direction, and the rotation arc of the rotary axis in the C-axis direction.

[0028] The substructure axis_info that stores data such as the spindle current can be defined as Struct axis_info={axis_id; em; et; fem; fet}, where axis_id, em, et, fem, and fet correspond to the axis information number, spindle current, spindle load, feed axis current, and feed axis load, respectively.

[0029] The substructure program_info that stores the NC program running information can be defined as Struct program_info={c; w; o; i}, where c, w, o, and i correspond to the program information number, processing workpiece number, process number, and instruction line number, respectively.

[0030] The substructure l that stores the processing position information can be defined as Struct l={loc_id; x; y; z; angle_a; angle_c}, where loc_id, x, y, z, angle_a, and angle_c represent the axis position information number, the mechanical coordinates of the feed axis in the X-axis direction, the mechanical coordinates of the feed axis in the Y-axis direction, the mechanical coordinates of the feed axis in the Z-axis direction, the rotation arc of the rotation axis in the A-axis direction, and the rotation arc of the rotation axis in the C-axis direction, respectively.

[0031] The milling process data of the external sensor includes the acquisition time, the spindle vibration signal in the X-axis direction, the spindle vibration signal in the Y-axis direction, the spindle vibration signal in the Z-axis direction, the spindle force signal in the X-axis direction, the spindle force signal in the Y-axis direction, and the spindle force signal in the Z-axis direction.

[0032] The data structure Sensor corresponding to the above-mentioned external sensor milling process data can be defined as StructSensor={sensor_id; acq_time; vx; vy; vz; fx; fy; fz}, where sensor_id is the unique identifier of the data structure of the external sensor milling process data, acq_time is the acquisition time, and vx, vy, vz, fx, fy, and fz represent the spindle vibration signal in the X-axis direction, the spindle vibration signal in the Y-axis direction, the spindle vibration signal in the Z-axis direction, the spindle force signal in the X-axis direction, the spindle force signal in the Y-axis direction, and the spindle force signal in the Z-axis direction, respectively.

[0033] Preferably, the structured data includes CNC machine tool internal CNC system milling process data, external sensor milling process data, processing arrangement information, data set component information, algorithm component information, and intelligent analysis process component information stored in the database.

[0034] The processing arrangement information includes the processing arrangement information number, the processing workpiece number list, the processing number, the processed number, the corresponding processing workpiece NC program file storage path and the workpiece quality inspection file storage path.

[0035] The above-mentioned processing schedule information data structure process_schedule_info can be defined as Structprocess_schedule_info={ps_id; ps_list;process_number_total; process_number_done; nc_filepath; quality_filepath}, where ps_id is the processing schedule information number, that is, the unique identifier of the processing schedule information data structure, ps_list, process_number_total, process_number_done, nc_filepath, and quality_filepath respectively represent the processing workpiece number list, the number of processes, the number of processes completed, the corresponding processing workpiece NC program file storage path, and the workpiece quality inspection file storage path.

[0036] The dataset component information includes the dataset component number, the list of artifact numbers involved, the signal type contained, the sample sequence length, the training set and test set division ratio, and the file storage path; The dataset component information data structure dataset_element_info can be defined as Structdataset_element_info={data _id; ps_used_list; signal_type; seq_len; div_ratio; data_filepath}, where data _id is the dataset component information number, ps_used_list, signal_type, seq_len, div_ratio, and data_filepath represent the artifact number list involved, the signal type contained, the sample sequence length, the training set and test set division ratio, and the file storage path, respectively.

[0037] Algorithm component information includes the algorithm component number, algorithm type involved, training rounds, loss function used, input parameters, and algorithm script storage file path; The algorithm component information data structure algorithm_info can be defined as Struct algorithm_info={algorithm _id; algorithm_type; epoch; loss_type; in_param; algorithm_filepath}, where algorithm_id is the algorithm component number, algorithm_type, epoch, loss_type, in_param, and algorithm_filepath respectively represent the algorithm type involved, training round, loss function used, input parameters, and algorithm script storage file path.

[0038] The intelligent analysis process component information includes the node information and edge information of the roughness prediction intelligent analysis process. The node information includes: node number, data or algorithm component number, node type, execution order, and storage path of processed data. The execution order is from small to large, and the nodes with smaller values ​​are executed first. The data are uniformly stored in array format. The edge information includes: edge number, source node number, and target node number.

[0039] The above-mentioned node information data structure node can be defined as Struct node={node _id; use_id;node _type; exec_seq; done_filepath}, where node_id represents the node number, use_id, node_type, exec_seq, and done_filepath represent the data or algorithm component number, node type, execution order, and storage path of the processed data, respectively.

[0040] The edge information data structure edge can be defined as Struct edge={edge_id; source_node; target_node}, where edge_id represents the edge number, source_node and target_node represent the source node number and target node number respectively. Preferably, the semi-structured data includes the numerical control program files required for the operation of each workpiece, the files stored in the data array processed by the data set component, the surface storage degree detection result files, the contour error detection result files and the tool wear result detection files.

[0041] Preferably, the data basic management module is used to add, delete, modify or query the specific business logic of processing arrangement information, data set information and algorithm information functions, and also provides paging and sorting functions.

[0042] The data processing module includes a data preprocessing submodule and a data hierarchical organization submodule. The data processing module is used to organize data into a form that can be applied by deep learning analysis algorithms and visualization; The data preprocessing submodule is used to call the data access layer data to obtain signal data according to the dataset component information, and provides operations such as null value processing, outlier processing, signal denoising, and signal downsampling; The data hierarchical organization submodule is used to filter the data corresponding to the user's required process, workpiece, process, instruction, and position segment according to the processing arrangement number, workpiece number, process number, instruction line number, and processing coordinates selected by the user, and divide the entire sequence into several subsequences according to the subsequence length set by the user; the data is stored in a file format; The intelligent analysis module is used to call the data files and corresponding algorithm scripts required by the components in the execution order of the intelligent analysis process. Among them, the intelligent analysis process is constructed by the user by dragging and dropping dataset components and algorithm components.

[0043] Preferably, a data set component management interface is used to create, modify, delete and retrieve data set component information; Algorithm component management interface, used to create, modify, delete and retrieve algorithm component information; Drag-and-drop intelligent analysis process construction and adjustment interface, used for graphical arrangement and real-time execution monitoring of milling data analysis processes; Data visualization interface, used to visualize timing signals according to user-specific process, workpiece, process, instruction and position visualization requirements; The user interface corresponding to data basic management is used to create, modify, delete and retrieve processing arrangement information.

[0044] like Figure 4 As shown, Figure 4 This is a schematic diagram of milling data modeling and intelligent analysis provided by the embodiment of the present application. In this method, the spindle X-axis vibration v is obtained from the CNC system, external sensors, tool imagers, surface roughness detectors and other quality inspection equipment. x , spindle Y-axis vibration v y , spindle Z-axis vibration v z , spindle current e m , spindle torque e t , process number c, workpiece number w, operation number o, instruction line number i, machining position l, spindle speed s, feed speed f, cutting depth d, contour error Q(x, y, z), tool wear T(t), acquisition time t, and feed axis coordinates x, y, and z in the X, Y, and Z directions.

[0045] Machine tool response data RD such as vibration signals t With the machine tool processing task PT as input, the mapping relationship between the system response data RD, the machine tool processing task PT to the machine tool equipment and processing status Y is established from the five levels of process, workpiece, procedure, instruction and position: Y=g[RD, PT].

[0046] For data access, the collected data is organized into different types, among which the system response data RD t ={t, v x , v y , v z , e m , e t} indicates that the system response data involves the acquisition time t, the main axis X-axis vibration v x , spindle Y-axis vibration v y , spindle Z-axis vibration v z , spindle current e m , spindle torque e t Etc. A machine tool processing task PT = {t, s, f, d} represents the machine tool processing task information, including acquisition time t, spindle speed s, feed rate f, cutting depth d, etc. The processing task PT can be described from five levels: process, workpiece, procedure, instruction, and location, and in both time and space domains.

[0047] Data modeling includes time domain processing task data modeling PTk=fk[k(t), t], spatial domain processing task data modeling PT kʹ =f kʹ [k(t),x,y,z] and time domain system response data are converted into spatial domain system response data for RD modeling s =h(RD t ,x,y,z).

[0048] The time-domain processing task data modeling can be expressed as PTk=fk[k(t), t], k=c1, w1, o1, i1, l1, where c1 represents the process parameters such as cutting depth and feed rate that characterize the processing task from the process level, w1 represents the description from the workpiece level, o1 represents the description from the process level, i1 represents the description from the instruction level, and l1 represents the description from the position level. In addition, t represents the acquisition time, k(t) represents the change function of the individual data such as cutting depth and feed rate described at the process, workpiece, process, instruction and position levels with respect to t, and fk is used to represent the mapping relationship between a single type of processing task description information and the acquisition time t. k(t) and the acquisition time t are inputs, and are mapped to the function of the change of multiple processing task description information at the acquisition time t. The spatial processing task data modeling can be expressed as PT kʹ =f kʹ [k(t),x,y,z], k=c2, w2, o2, i2, l2, c2 represents the process parameters such as cutting depth and feed rate that characterize the machining task from the process level, w2 represents the description from the workpiece level, o2 represents the description from the process level, i2 represents the description from the instruction level, and l2 represents the description from the position level. In addition, x represents the mechanical coordinate of the X-axis at the acquisition time t, y represents the mechanical coordinate of the Y-axis at the acquisition time t, and z represents the mechanical coordinate of the Z-axis at the acquisition time t. kʹ It is used to characterize the mapping relationship between a single type of processing task description information data and the acquisition time t. k(t) and the acquisition coordinates x, y, z are inputs, and it is mapped to a function of the changes in the acquisition time t and the processing position coordinates x, y, z of multiple processing task description information. The time domain signal is combined with the processing position coordinates to transform into the spatial domain form, which can be expressed as RD s =h(RD t ,x,y,z), used for intelligent analysis applications in the spatial domain, and system response data RD in the time domain t With the coordinates of the feed axes in the three directions of x, y, and z as input, the time domain signal is mapped to the processing points with the same acquisition time through the mapping relationship h, and the spatial domain data matrix RD containing the position information is obtained by traversing the entire processing space. s .

[0049] Furthermore, intelligent data analysis involves both time domain and space domain analysis to predict the contour deviation at each position and the tool wear value at each moment.

[0050] The analysis in time domain is expressed as T(t)=g k [RD t ,PTk ], k=c3, w3, o3, i3, l3, c3 represents intelligent analysis from the process level, w3 represents intelligent analysis from the workpiece level, o3 represents intelligent analysis from the process level, i3 represents intelligent analysis from the instruction level, l3 represents intelligent analysis from the position level, T(t) is the change of the machine tool tool wear value with respect to time t, the specific analysis process T(t)=g k [RD t ,PT k ]Time domain system response data RD t and processing task description information PT k As input, through the neural network model g k Establish a mapping relationship with tool wear.

[0051] The spatial form of the analysis is expressed as Q(x,y,z)=g k [RD s ,PT kʹ ], k=c4, w4, o4, i4, l4, c4 represents intelligent analysis from the process level, w4 represents intelligent analysis from the workpiece level, o4 represents intelligent analysis from the process level, i4 represents intelligent analysis from the instruction level, and l4 represents intelligent analysis from the position level. Q(x,y,z) is the contour error function of each processing point of the workpiece. The specific analysis process Q(x,y,z)=g k [RD s ,PT kʹ ], with airspace system response data RD s and processing task description information PT kʹ As input, through the neural network model g k Establish a mapping relationship between the error and the workpiece contour.

[0052] An embodiment of the present application provides a data-driven intelligent analysis device for milling processing, which aims to predict the surface roughness of part plane milling processing and can effectively serve the intelligent analysis process of thin-walled part milling processing. It will realize the prediction of quality indicators such as process level, workpiece level, procedure level, instruction level and position level surface roughness in the form of component dragging as an overall demand, and divide the overall architecture into perception layer, data access layer, business logic layer and application layer according to data collection and persistence, data access, data processing and analysis, and data application.

[0053] Based on the above data-driven milling processing intelligent analysis device embodiment, the present application embodiment also provides a data-driven milling processing intelligent analysis method, such as Figure 2 As shown, Figure 2A flow chart of a data-driven intelligent analysis method for milling provided in an embodiment of the present application. The method comprises the following steps: S201, the perception layer collects milling process data of the internal CNC system of the CNC machine tool and the milling process data of the external sensor.

[0054] S202: The data access layer stores, retrieves and accesses structured data and semi-structured data.

[0055] S203, the business logic layer performs basic management, preprocessing, hierarchical organization, data set management, algorithm management and intelligent analysis workflow of the milling process data, processes and analyzes data at different data scales such as process level, process level, workpiece level, instruction set and position level, and performs visualization, and divides the data into basic management module, data processing module and intelligent analysis module.

[0056] S204, the application layer provides a data set management interface, an algorithm management interface, an intelligent analysis interface, a data visualization interface, and a data foundation management interface.

[0057] The data-driven intelligent analysis method for milling processing provided in the embodiment of the present application has the same technical features as the data-driven intelligent analysis device for milling processing provided in the above embodiment, so it can also solve the same technical problems and achieve the same technical effects.

[0058] The present application also provides another data-driven intelligent analysis method for milling processing, such as Figure 3 As shown, Figure 3 A flow chart of another data-driven intelligent analysis method for milling provided in an embodiment of the present application. The method comprises the following steps: S301, clarify the functional model, conceptual data model and behavioral model in the intelligent analysis system.

[0059] Specifically, the functional model is used to describe the functional modules that the system should have. Specifically, this method divides the software into the perception layer, data access layer, business logic layer and application layer. Figure 5 As shown, Figure 5A schematic diagram of the structure of another data-driven intelligent analysis device for milling provided in an embodiment of the present application. This method aims to realize the functions of processing arrangement management, data set management, algorithm management, construction of intelligent analysis process and data visualization in the application layer 221, and combines the business logic layer 217 and the data access layer 207 to complete the analysis of the surface roughness of the workpiece using the data collected by the application perception layer 200. Among them, the business logic layer 217 divides the data into different scales, including process level, workpiece level, procedure level, instruction level and position level, to provide targeted analysis of the processing data under different processes, workpieces, procedures, instructions and positions. The perception layer 200 is used to collect real-time data of the milling process from the CNC system 203 inside the machine tool and the external sensor 201. Milling process data from the CNC system 203 within the machine tool includes data collected during the milling process, including the acquisition time, spindle current, spindle load, feed axis current, feed axis load, workpiece number, process number, instruction line number, X-axis feed axis mechanical coordinates, Y-axis feed axis mechanical coordinates, Z-axis feed axis mechanical coordinates, and the rotational arc of the A-axis and C-axis rotation axes. Milling process data from the external sensor 201 includes data collected during the milling process, spindle vibration signals in three directions, and force signals.

[0060] Specifically, the milling process data from the CNC system 203 within the machine tool can be acquired through a data acquisition interface provided by the CNC system 203. The milling process data from the external sensor 201 can be measured using an acceleration sensor and a pressure sensor attached to the machine tool spindle. Both the acceleration sensor and the pressure sensor are external sensors 201. The milling process data from the CNC system 203 within the machine tool and the external sensor 201 are stored on the data acquisition card 202 of the perception layer 200.

[0061] After establishing communication with the external sensors 201 and the CNC system 203 within the machine tool, the data acquisition card 202, combined with the data acquisition interface, data acquisition module 204, and thread management module 205 provided by the CNC system 203, collects and persists the data required by the application layer 221. The data acquisition card 202 includes functions for configuring the IP address and communication port of the communication network, thread management, timer triggering, and API function interface calls. By configuring the IP address and communication port of the communication network for the external sensors 201, CNC system 203, and analysis system, these sensors 201, CNC system 203, and analysis system can be added to the same local area network to achieve interconnection.

[0062] The data access layer 207 provides data storage and access interfaces. Specifically, the data access layer 207 includes a structured data access module 2071 and a semi-structured data access module 2072. The structured data access module 2071 is responsible for internal CNC system data 208, external sensor data 209, processing schedule information 210, dataset component information 211, algorithm component information 212, and intelligent analysis process information 213. The semi-structured data access module 2072 includes CNC program files 214, dataset storage files 215, and quality inspection result files 216.

[0063] The business logic layer 217 is used to realize data processing and analysis, pre-process, clean and classify the collected raw data, and provide support for the upper-level intelligent analysis. It mainly includes the milling data basic management 218, the data processing module 219 and the intelligent analysis module 220. Among them, the data processing module 219 calls the data access interface of the data access layer 207 according to the user submission form of the application layer 221 through the data segment screening module 2191 to obtain the required data, and uses the data pre-processing module 2192 to process the outliers, smooth the noise and process the null values ​​of the raw data. Finally, the hierarchical organization module 2193 organizes the data into processes, workpieces, procedures, instructions and locations according to the data scale required for analysis.

[0064] The application layer 221 is used to support user interaction, allowing users to configure and manage the data base management interface 223, the data set component management interface 224, the algorithm component management interface 225, and the intelligent analysis interface 226, and display the analysis results through the data visualization interface 222.

[0065] The conceptual data model defines the types and relationships of data elements in the system, including external sensor data 209, internal CNC system data 208, processing schedule information 210, dataset component information 211, algorithm component information 212, and intelligent analysis process information 213. External sensor data 209 and internal CNC system data 208 are linked to processing schedule information 210 via foreign keys, ensuring that the system can accurately track the processing data for each processing schedule.

[0066] The behavioral model describes the dynamic behavior of each layer or each functional module in the software under different states, including the interaction between the data acquisition card 202 and the external sensor 201 and the numerical control system 203; data processing operations such as synchronization and fusion of multi-source data, division of data under five data scales, and data cleaning; the specific implementation of the intelligent analysis process in the form of dragging and dropping components; and the response to user operations such as adding, deleting, modifying, and checking, and the operation of the intelligent analysis process when the user interacts with the system through the application layer 221.

[0067] S302, clarifying the database data tables corresponding to the perception layer data collection and the data access layer data addition, deletion, modification and query operations.

[0068] Based on the entity attributes in the conceptual data model and the data characteristics of the milling process, this embodiment stores the data in a relational database and designs a corresponding data table structure. For each data table, fields are set according to actual needs, and appropriate data types and field lengths are selected.

[0069] The internal CNC system data 208 corresponds to the internal data table of the machine tool in the database, including the acquisition time, spindle current, spindle load, feed axis current, feed axis load, processing workpiece number, process number, instruction line number, mechanical coordinates of the feed axis in the X-axis direction, mechanical coordinates of the feed axis in the Y-axis direction, mechanical coordinates of the feed axis in the Z-axis direction, rotation arc of the rotary axis in the A-axis direction, rotation arc of the rotary axis in the C-axis direction, and the corresponding processing arrangement number. The unique identification primary key is the acquisition time, and the processing arrangement number is the foreign key.

[0070] The process data of the external sensor 201 corresponds to the external sensor data table in the database. The fields include the acquisition time, the spindle vibration signal and force signal in three directions, and the corresponding processing schedule number. The unique identification primary key is the acquisition time, and the processing schedule number is the foreign key.

[0071] The processing arrangement information 210 corresponds to the processing arrangement information table in the database, and the fields include: processing arrangement number, processing workpiece number list, processing number, processed number, NC program file path, quality inspection file path, and the unique identification primary key is the processing arrangement number.

[0072] The data set component information 211 corresponds to the data component information table in the database, and the fields include: data set component number, workpiece number list, signal type, sample sequence length, training set and test set division ratio, file storage path, and the unique identification primary key is the data set component number.

[0073] The algorithm component information 212 corresponds to the algorithm component information table in the database, and the fields include: algorithm component number, algorithm type, training round, loss function, input parameters, algorithm script storage file path, and the unique identification primary key is the algorithm component number.

[0074] The intelligent analysis process information 213 corresponds to the intelligent analysis process information table in the database, and the fields include: node number, data or algorithm component number, node type, execution order, and processed data storage path. The unique identification primary key is the node number.

[0075] The relationship between nodes in the analysis process corresponds to the analysis process node information table in the database. The fields include: edge number, source node number, target node number, and the unique identification primary key is the edge number.

[0076] S303, clarify the data collection, data access and storage logic, data processing and intelligent analysis process functions involved in the behavior model.

[0077] The main thread in data acquisition is used to call functions, control timers, and control the startup and termination of the acquisition and data storage threads. Once the data acquisition card 202, the machine tool's internal CNC system 203, and the analysis system are successfully connected, the analysis system's timer calls the API function interfaces in each acquisition thread at preset intervals, such as the interface for obtaining the line number of the running instruction. The processing data returned by all interfaces is encapsulated into the physical and logical structures corresponding to the aforementioned internal CNC system data 208 and external sensor data 209, and is added to different data queues. The data storage thread continuously retrieves the returned data from the non-empty data queue and stores it in the database (data storage module 206).

[0078] For example, the perception layer 200 uses the FOCAS data communication protocol developed over a TCP / IP interface to acquire internal electronic control data from the FANUC 0i-MF numerical control system of a Hardinge GX 710Plus milling machine. The data acquisition card 202 acquires signals from each external sensor 201 and transmits them to the card as UDP packets. The data structure within the perception layer 200 is primarily based on the physical and logical structure corresponding to the processing data from the numerical control system 203 and external sensors 201. Object-oriented programming is employed to encapsulate different types of data into classes. During system operation, the milling process data accurately describes the milling machine's processing tasks and operating status.

[0079] The data foundation management interface 223 primarily describes how to connect to the user interface via a RESTful API (Representational State Transfer) interface based on user interaction with the system interface. This interface receives user operation instructions for the following database tables, enabling operations such as adding, deleting, modifying, and querying information in the machining schedule table, data component information table, algorithm component information table, and intelligent analysis process information table. This interface also interacts with the data acquisition card 202, responsible for storing milling process data collected from the CNC system 203 and external sensor 201, as well as accessing this collected machining process data.

[0080] The data processing module 219 filters out the required processing data at the process level, workpiece level, operation level, instruction level and position level according to the data requirements of intelligent applications such as processing signal visualization and tool wear prediction, and provides basic data preprocessing, and finally converts the data into a form that can be used for upper-level data visualization or artificial intelligence applications.

[0081] Data screening is to select data at the process level, workpiece level, operation level, instruction level and position level according to the user requirement form submitted by the application layer 221 .

[0082] Data preprocessing involves denoising, standardizing, and normalizing the collected raw data to ensure data quality and avoid the impact of abnormal data on subsequent analysis.

[0083] Data format conversion is to convert the pre-processed data into a format suitable for upper-level data visualization tools and artificial intelligence models, such as time series data format, image data format, etc., to ensure that the data can be efficiently used by the intelligent analysis system.

[0084] Intelligent data analysis describes how to sequentially call various components and return the corresponding analysis results based on the user-created intelligent analysis process and the information in the corresponding intelligent analysis process information table. The information in the intelligent analysis process information table includes the order of each algorithm node, the storage path of the corresponding algorithm file, and the storage path of intermediate data. This ensures the orderly transmission of data flow and the analysis process, and ultimately ensures that the analysis results are returned to the user interface.

[0085] The intelligent analysis process runs to analyze the intelligent analysis process created by the user through the user interaction interface and to clarify the calling sequence of each analysis node and algorithm component.

[0086] Component calls are based on the user-defined intelligent analysis process information table, loading relevant algorithm components from the algorithm component library and calling them in sequence. The intermediate data generated after each algorithm node is executed is stored according to the defined path for use by subsequent algorithm nodes.

[0087] After all analysis nodes are completed, the final analysis results are fed back to the user through the interface. These results include tool wear predictions and machining process optimization suggestions. The system supports data visualization as needed and provides analysis reports to external systems through interfaces.

[0088] S304: Clarify the display format of the data involved in the functional interface of each application layer and the required data processing and analysis functions.

[0089] The dataset component management interface 224 provides functionality for creating and managing dataset components. Users fill out the corresponding form using drop-down boxes and text boxes based on parameters such as the required data scale, processing data type, sequence length, data representation, and the ratio between training and validation sets. This form then sends a request to the business logic layer 217 to create a specific dataset component. This interface also supports searching for existing dataset components based on representation and sequence length, as well as modifying and deleting component information.

[0090] The algorithm component management interface 225 provides buttons for uploading user-defined algorithm files and parameter description files. Users can enter the algorithm type and detailed description, and upload the algorithm execution file and parameter description file. The business logic layer 217 then integrates the customized algorithm. Default algorithm component parameters, including training rounds, learning rate, loss function, and optimization algorithm, can also be adjusted to meet different data analysis requirements. Furthermore, this interface supports searching created algorithm components by algorithm type and description, as well as modifying and deleting algorithm components.

[0091] The intelligent analysis interface 226 integrates a drag-and-drop dynamic analysis process building tool. This tool consists of a sidebar and a main canvas area. The sidebar displays the various dataset components and algorithm components within the system, while the canvas area is used to create and visualize intelligent analysis process diagrams. Users can pre-select the required dataset or algorithm component in the sidebar and drag it onto the canvas to configure the process. After the process configuration is complete, the user clicks the "Run" button to view the current process execution progress and the results of each node in real time through the intelligent analysis results 2222.

[0092] Furthermore, the data visualization interface 222 provides visualization of corresponding time series signals based on the user's specific process, workpiece, operation, instruction, and position visualization requirements. In the processing signal time domain visualization 2221, the user can select the desired signal type through a drop-down menu. The data visualization interface displays the corresponding time series data in the form of a line graph. At the same time, the processing signal time and space domain visualization 2223 also displays the actual processing coordinates in the form of a scatter plot, using the X-axis, Y-axis, and Z-axis coordinates of the processing point as the scatter point location. The selected signal value is mapped to the scatter point color, intuitively showing the spatial distribution of the signal during the processing process.

[0093] The data base management interface 223 provides comprehensive management of the system's processing schedule information 210, including data items such as processing schedule number, processing batch, NC program, processing status, and quality inspection documents. Users can create new processing records by filling in these data items and perform categorized search, editing, updating, and deletion operations on all processing schedule information, thereby achieving systematic management and dynamic maintenance of processing records.

[0094] The present application also provides a computing device. Figure 6 As shown, this figure is a schematic diagram of a computing device provided by an embodiment of the present application, wherein the computing device 400 includes a bus 401, a processor 402, a communication interface 403, and a memory 404. The processor 402, the memory 404, and the communication interface 403 communicate with each other via the bus 401.

[0095] The bus 401 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0096] The processor 402 may be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0097] Communication interface 403 is used for external communication. Memory 404 may include volatile memory, such as random access memory (RAM). Memory 404 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0098] The memory 404 stores executable codes, and the processor 402 executes the executable codes to perform the aforementioned data-driven intelligent analysis method for milling processing.

[0099] Embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium can be any available medium capable of being stored by a computing device, or a data storage device such as a data center that contains one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, hard disk, or magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to execute the aforementioned milling processing data analysis method.

[0100] The present application also provides a computer program product comprising one or more computer instructions that, when loaded and executed on a computing device, fully or partially generate the process or function described in the present application.

[0101] The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer or data center to another website, computer or data center via wired (e.g., coaxial cable, optical fiber) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0102] When the computer program product is executed by a computer, the computer performs any of the aforementioned data-driven intelligent milling analysis methods. The computer program product may be a software installation package, which can be downloaded and executed on a computer when any of the aforementioned data-driven intelligent milling analysis methods is needed.

[0103] The descriptions of the processes or structures corresponding to the above figures have different emphases. For parts that are not described in detail in a certain process or structure, please refer to the relevant descriptions of other processes or structures.

[0104] The above description is only a specific implementation method of the present application, but the protection scope of the present application is not limited thereto. Any changes or replacements within the technical scope disclosed in the present application should be included in the protection scope of the present application.

Claims

1. A data-driven intelligent analysis device for milling processing, characterized in that: The device comprises: The perception layer is used to collect milling process data from the CNC system inside the CNC machine tool and milling process data from external sensors; Data access layer, used to store, retrieve and access structured and semi-structured data; The business logic layer is used to perform basic data management, preprocessing, hierarchical organization, data set management, algorithm management, and intelligent analysis workflows for the milling process. It processes, analyzes, and visualizes data at different data scales, including process level, operation level, workpiece level, instruction set, and position level. It is divided into a basic data management module, a data processing module, and an intelligent analysis module. The application layer is used to provide the data set management interface, algorithm management interface, intelligent analysis interface, data visualization interface and data foundation management interface.

2. The data-driven milling processing intelligent analysis device according to claim 1 is characterized in that: The milling process data of the CNC system inside the CNC machine tool includes acquisition time, spindle current, spindle load, feed axis current, feed axis load, processing workpiece number, process number, instruction line number, mechanical coordinates of the feed axis in the X-axis direction, mechanical coordinates of the feed axis in the Y-axis direction, mechanical coordinates of the feed axis in the Z-axis direction, rotation arc of the rotation axis in the A-axis direction, and rotation arc of the rotation axis in the C-axis direction; The external sensor milling process data includes acquisition time, spindle vibration signal in the X-axis direction, spindle vibration signal in the Y-axis direction, spindle vibration signal in the Z-axis direction, spindle force signal in the X-axis direction, spindle force signal in the Y-axis direction, and spindle force signal in the Z-axis direction.

3. The data-driven milling processing intelligent analysis device according to claim 2, characterized in that: The structured data includes the CNC system milling process data of the CNC machine tool, the external sensor milling process data, processing arrangement information, data set component information, algorithm component information, and intelligent analysis process component information stored in the database; The processing arrangement information includes the processing arrangement information number, the processing workpiece number list, the processing number, the processed number, the corresponding processing workpiece NC program file storage path and the workpiece quality inspection file storage path; The dataset component information includes the dataset component number, a list of artifact numbers involved, the signal type contained, the sample sequence length, the training set, the test set division ratio, and the file storage path; The algorithm component information includes the algorithm component number, the algorithm type involved, the training rounds, the loss function used, the input parameters and the algorithm script storage file path; The intelligent analysis process component information includes node information and edge information for storing the roughness prediction intelligent analysis process; wherein, the node information includes: node number, data or algorithm component number, node type, execution order, and storage path of processed data. The execution order is from small to large, and nodes with smaller values ​​are executed first, and the data is uniformly stored in an array format. The edge information includes: edge number, source node number, and target node number.

4. The data-driven milling processing intelligent analysis device according to claim 3 is characterized in that: The semi-structured data includes the numerical control program files required for the operation of each workpiece, the files stored in the data array processed by the data set component, the surface storage degree detection result files, the contour error detection result files and the tool wear result detection files.

5. The data-driven milling processing intelligent analysis device according to claim 4, characterized in that: The data basic management module is used to add, delete, modify or query the specific business logic of processing arrangement information, data set information and algorithm information, and also provides paging and sorting functions; The data processing module includes a data preprocessing submodule and a data hierarchical organization submodule. The data processing module is used to organize the data into a form applicable to deep learning analysis algorithms and visualization; The data preprocessing submodule is used to call the data access layer data to obtain signal data according to the data set component information, and provide operations such as null value processing, outlier processing, signal denoising, and signal downsampling; The data hierarchical organization submodule is used to filter the data corresponding to the process, workpiece, process, instruction, and position segment required by the user based on the processing arrangement number, workpiece number, process number, instruction line number, and processing coordinates selected by the user, and divide the entire sequence into several subsequences according to the subsequence length set by the user; and store the data in a file format; The intelligent analysis module is used to call the data files and corresponding algorithm scripts required by the components in sequence according to the execution order of the components in the intelligent analysis process; wherein, the intelligent analysis process is constructed by the user by dragging and dropping data set components and algorithm components.

6. The data-driven milling processing intelligent analysis device according to claim 5, characterized in that: The data set management interface is used to create, modify, delete and retrieve the data set component information; The algorithm management interface is used to create, modify, delete and retrieve the algorithm component information; The intelligent analysis interface is used to graphically arrange and monitor the milling processing data analysis process in real time; The data visualization interface is used to visualize the timing signals according to the user's specific process, workpiece, process, instruction and position visualization requirements; The user interface corresponding to the data basic management is used to create, modify, delete and retrieve the processing arrangement information.

7. A data-driven intelligent analysis method for milling processing, characterized in that: The method comprises: The perception layer collects milling process data from the CNC system inside the CNC machine tool and milling process data from external sensors; The data access layer stores, retrieves and accesses structured and semi-structured data; The business logic layer performs basic data management, preprocessing, hierarchical organization, data set management, algorithm management, and intelligent analysis workflows for the milling process. It processes, analyzes, and visualizes data at different data scales, including process level, operation level, workpiece level, instruction set, and position level. It is divided into a basic data management module, a data processing module, and an intelligent analysis module. The application layer provides a data set management interface, an algorithm management interface, an intelligent analysis interface, a data visualization interface, and a data foundation management interface.

8. A computing device, characterized in that including memory and processor; One or more computer programs are stored in the memory, and the one or more computer programs include instructions; when the instructions are executed by the processor, the computing device executes the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium is used to store a computer program, and the computer program is used to execute the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The computer program product comprises one or more computer instructions. When the computer instructions are executed by a computer, the computer performs the method according to any one of claims 1 to 6.

Images