Real-time monitoring method and system for mineral material machine tool

Abstract

The invention discloses a real-time monitoring method and system for a mineral material machine tool, and relates to the technical field of mineral material machine tools, and the method comprises the steps: building a machine tool execution task of the mineral material machine tool, carrying out the task decomposition, and constructing N subtasks; establishing association mapping; establishing a motion data set of the subtasks; cutter image acquisition after cooling liquid cleaning is carried out at the start-stop node, and a verification image set of the subtasks is established; establishing execution exception of the subtasks; and inputting the motion data set, the verification image set and execution abnormity into an abnormity fusion authentication model, executing machine tool abnormity authentication, establishing machine tool abnormity early warning, and completing real-time monitoring of the mineral material machine tool according to the machine tool abnormity early warning. The technical problems that in the prior art, the monitoring refinement degree of the mineral material machine tool is low, and the monitoring result is unreliable are solved, and the technical effects that execution tasks of the mineral material machine tool are decomposed, multi-dimensional analysis is conducted based on the decomposition result, and the machine tool abnormity early warning reliability is improved are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of mineral material machine tools, and in particular to a real-time monitoring method and system for mineral material machine tools. Background Art

[0002] Mineral casting machine tools are widely used in machining due to their excellent wear resistance, stability, and sound absorption properties, which can meet specific processing needs and working environment requirements. To ensure the optimal operation of mineral casting machine tools, real-time monitoring of their operating status is necessary. Currently, this mainly involves monitoring mineral casting machine tools as a whole, but this monitoring and analysis often results in false alarms or missed alarms, which can prevent abnormal operation of mineral casting machine tools from being discovered in a timely manner.

[0003] The existing technology has technical problems such as low level of monitoring refinement of mineral material machine tools and unreliable monitoring results. Summary of the Invention

[0004] The present application provides a real-time monitoring method and system for a mineral material machine tool, which is used to solve the technical problems in the prior art of low level of monitoring refinement and unreliable monitoring results of mineral material machine tools.

[0005] In view of the above problems, the present application provides a real-time monitoring method and system for mineral material machine tools.

[0006] A first aspect of the present application provides a real-time monitoring method for a mineral material machine tool, the method comprising: Establishing a machine tool execution task for a mineral material machine tool, decomposing the machine tool execution task into N subtasks, wherein the task decomposition includes tool task decomposition and feed task decomposition; Obtain the quality inspection parameters of the workpiece and establish an association mapping between the quality inspection parameters and N subtasks, where one quality inspection parameter corresponds to one or more subtasks; Use displacement sensors to monitor the motion of N subtasks within the machine tool range within the task scope and establish a motion data set for the subtasks; Obtaining the start and stop nodes of N subtasks, collecting images of the tool after coolant cleaning using a visual acquisition sensor at the start and stop nodes, and establishing a verification image set for the subtasks; Obtaining a quality inspection result of the workpiece, and establishing an execution exception of the subtask based on the association mapping; The motion data set, the verification image set and the execution anomaly are input into an anomaly fusion authentication model, machine tool anomaly authentication is performed, a machine tool anomaly warning is established, and real-time monitoring of mineral material machine tools is completed based on the machine tool anomaly warning.

[0007] A second aspect of the present application provides a real-time monitoring system for a mineral material machine tool, the system comprising: A subtask construction module is used to establish a machine tool execution task for a mineral material machine tool, decompose the machine tool execution task into N subtasks, wherein the task decomposition includes tool task decomposition and feed task decomposition; An association mapping establishment module is used to obtain the quality inspection parameters of the workpiece and establish an association mapping between the quality inspection parameters and N subtasks, wherein one quality inspection parameter corresponds to one or more subtasks; A motion data set establishment module is used to monitor the motion of N subtasks within the machine tool range within the task scope through a displacement sensor and establish a motion data set for the subtasks; A verification image set establishment module is used to obtain the start and stop nodes of N subtasks, and to acquire images of the tool after coolant cleaning at the start and stop nodes using a visual acquisition sensor to establish a verification image set for the subtasks; an execution exception establishment module, configured to obtain a quality inspection result of a workpiece and establish an execution exception of a subtask based on the association mapping; The real-time monitoring module is used to input the motion data set, the verification image set and the execution anomaly into the anomaly fusion authentication model, perform machine tool anomaly authentication, establish a machine tool anomaly warning, and complete real-time monitoring of the mineral material machine tool based on the machine tool anomaly warning.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application establishes a machine tool execution task for a mineral material machine tool, decomposes the machine tool execution task into N subtasks, wherein the task decomposition includes tool task decomposition and feed task decomposition. The application then obtains the quality-checkable parameters of the workpiece and establishes an association mapping between the quality-checkable parameters and the N subtasks. Each quality-checkable parameter corresponds to one or more subtasks. A displacement sensor is then used to monitor the motion of the N subtasks within the machine tool range within the task scope, establishing a motion dataset for the subtasks. The application then obtains the start and stop nodes of the N subtasks. At the start and stop nodes, a visual acquisition sensor is used to capture images of the tool after coolant cleaning, establish a verification image set for the subtask, obtain the quality inspection results of the workpiece, and establish execution anomalies for the subtasks based on the association mapping. The motion dataset, verification image set, and execution anomalies are input into an anomaly fusion authentication model, machine tool anomaly authentication is performed, a machine tool anomaly warning is established, and real-time monitoring of the mineral material machine tool is completed based on the machine tool anomaly warning. This achieves the technical effect of decomposing the execution tasks of the mineral material machine tool, conducting multi-dimensional analysis based on the decomposition results, and improving the reliability of the machine tool anomaly warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A flowchart of a real-time monitoring method for a mineral material machine tool provided in an embodiment of the present application; Figure 2 A schematic structural diagram of a real-time monitoring system for a mineral material machine tool provided in an embodiment of the present application.

[0011] Explanation of the reference numerals: subtask construction module 11 , association mapping establishment module 12 , motion data set establishment module 13 , verification image set establishment module 14 , execution anomaly establishment module 15 , real-time monitoring module 16 . DETAILED DESCRIPTION

[0012] The present application provides a real-time monitoring method and system for mineral material machine tools, which is used to solve the technical problems in the prior art of low level of monitoring refinement and unreliable monitoring results of mineral material machine tools.

[0013] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only 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.

[0014] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0015] Example 1 like Figure 1 As shown, the present application provides a real-time monitoring method for a mineral material machine tool, wherein the method comprises: S100: establishing a machine tool execution task for a mineral material machine tool, decomposing the machine tool execution task into N subtasks, wherein the task decomposition includes tool task decomposition and feed task decomposition; Furthermore, the machine tool execution task is decomposed into N subtasks. In the embodiment of the present application, step S100 further includes: Establish a task decomposition model, input the machine tool execution task into the task decomposition model, and perform task decomposition, specifically including: S1: Perform tool call decomposition of the machine tool execution task through the tool decomposition unit and establish the initial decomposition result; S2: Synchronizing the initial decomposition result to the feed decomposition unit, decomposing the feed amount task of the same tool through the feed decomposition unit, and establishing an updated decomposition result; S3: Synchronize the update decomposition result to the continuous decomposition unit, perform continuous analysis of the action of the update decomposition result through the continuous decomposition unit, further decompose the discontinuous action task, and generate N subtasks.

[0016] In an embodiment of the present application, the mineral material machine tool is a machine tool constructed from mineral castings (such as artificial marble or artificial granite) and is used for machining. Based on order requirements, the machine tool execution tasks are established. These tasks are tasks performed by the mineral material machine tool on workpieces, including turning, milling, and grinding, as well as requirements that must be met after processing the workpieces, such as roughness, processing dimensions, and workpiece surface condition.

[0017] After obtaining the machine tool's task, the task is decomposed into two dimensions: tool and feed, resulting in N subtasks. This task decomposition includes tool and feed task decomposition. By decomposing the machine tool's task, a reliable basis for subsequent, refined, real-time monitoring of the machining process is achieved.

[0018] Optionally, the task decomposition model is a functional model for performing multi-dimensional decomposition of the machine tool execution task, including a tool decomposition unit, a feed decomposition unit, and a continuous decomposition unit. The tool decomposition unit is used to decompose the movement of the tool in the machine tool execution task, and classify tasks belonging to the same tool into one category, thereby obtaining the initial decomposition result. The feed decomposition unit is used to classify the different feed amounts of the same tool at different processes in the initial decomposition result. The continuous decomposition unit is used to continuously analyze multiple feed amount tasks of the same tool in the updated decomposition result.

[0019] Optionally, decomposition is first performed at the tool type level. In the tool decomposition unit, the machine tool execution tasks are searched using tool type as an index, and tasks belonging to the same tool are classified into one category, thereby obtaining the initial decomposition result. Tool types include turning tools, milling cutters, boring tools, etc.

[0020] Secondly, decomposition is performed from the feed level, and the task decomposition results of a tool type are extracted from the initial decomposition results respectively, and synchronized to the feed decomposition unit. In the feed decomposition unit, the feed executed at different stages of the tool is decomposed according to the task execution order to obtain the feed task set corresponding to the tool type. Multiple feed task sets corresponding to multiple tool types are used as updated decomposition results. Among them, the feed tasks in the multiple feed task sets have execution start and end time nodes. Finally, decomposition is performed from the action continuity level of the feed task, and the updated decomposition results are synchronized to the continuous decomposition unit. According to the execution start and end time nodes of the feed tasks in each feed task set, it is determined whether the feed task is a continuous action. If it is a continuous action, the multiple continuous feed tasks are regarded as a subtask, and the earliest execution start time and the latest execution stop time of the multiple feed tasks are used as the start and stop nodes of the subtask. If the action is not continuous, the discontinuous feed tasks are broken down into multiple subtasks, and the execution start and end time nodes of the discontinuous feed tasks are used as the starting nodes of the multiple subtasks. After the action continuity analysis, N subtasks are obtained.

[0021] S200: Acquire quality-inspectable parameters of a workpiece, and establish an association mapping between the quality-inspectable parameters and N subtasks, wherein one quality-inspectable parameter corresponds to one or more subtasks; In one possible embodiment, quality-checkable parameters are determined based on the type of workpiece (e.g., bearings, gears, hydraulic cylinders, etc.). These quality-checkable parameters are measurable characteristic parameters used to control the machining quality of the workpiece and assess its quality, including dimensional parameters, surface roughness, surface defect characteristic parameters (crack volume), and so on. Because the workpiece quality reflected by different quality-checkable parameters is related to different machining subtasks, such as the workpiece dimensional parameters being related to the feed rate of the turning tool and the rotational speed and feed rate of the grinding wheel, one quality-checkable parameter corresponds to one or more subtasks. Based on the type of quality-checkable parameters of the workpiece, a person skilled in the art will establish an association mapping between the quality-checkable parameters and the N subtasks based on the actual conditions of the workpiece.

[0022] S300: Using a displacement sensor to monitor the motion of N subtasks within a machine tool range within the task scope, and establish a motion data set for the subtasks; In one embodiment, a displacement sensor is used to monitor the motion of the machine tool range within the task range during the execution of N subtasks, and obtain a motion data set that can reflect the motion conditions during the execution of the subtask. The machine tool range is the time interval for monitoring during the execution of the subtask. For example, if the execution time of the subtask is 5 minutes, the machine tool range is 20 seconds. The execution process of the subtask is not monitored throughout, so as to avoid excessive monitoring data volume. In other words, the displacement sensor is used to record the movement distance of the tool in the N subtasks within the machine tool range when performing the processing task, and the recorded data is used as the motion data set.

[0023] For example, a turning tool is used to cut the surface of the workpiece with a feed rate of 5 mm. After the turning tool contacts the surface of the workpiece, the displacement sensor is activated to monitor the cutting movement of the turning tool until the machine tool range is met and the displacement monitoring is stopped. The data recorded by the displacement sensor during the monitoring process is used as the motion data set.

[0024] S400: Obtaining start and stop nodes of N subtasks, collecting images of the tool after coolant cleaning using a visual acquisition sensor at the start and stop nodes, and establishing a verification image set for the subtasks; Furthermore, the image of the tool after coolant cleaning is collected by the visual acquisition sensor at the start-stop node to establish a verification image set for the subtask. In the embodiment of the present application, step S400 further includes: Establishing a start-stop node association, and capturing images of the tool after coolant cleaning at the start node and the stop node, respectively, to create a start image and a stop image, wherein the start image and the stop image are images captured at the same angle; Perform image alignment of the start image and the stop image based on the start-stop node association, perform perspective recognition based on the alignment result, and establish an abnormal feature identification; A verification image set is established according to the abnormal feature identifier and the start image and the stop image.

[0025] Furthermore, the image alignment of the start image and the stop image is performed based on the association of the start and stop nodes, and perspective recognition is performed based on the alignment result to establish an abnormal feature identification. In this embodiment of the application, step S400 further includes: Acquire tool data, perform feature classification of the tool based on the tool data, and establish an important index for the tool partition; Obtain the recognition position of perspective recognition, perform recognition position mapping based on the important index of the tool partition, and obtain the important index mapping result; The position abnormality degree of perspective recognition is obtained, an abnormal value is calculated based on the position abnormality degree and the important index mapping result, and abnormal feature identification is completed based on the abnormal value calculation result.

[0026] In one possible embodiment, the start and stop nodes of the N subtasks are the start and stop nodes of the N subtasks, where the start node is the start time of the subtask and the stop node is the stop time of the subtask. A visual acquisition sensor is used at each of the start and stop nodes to capture images of the tool after coolant cleaning, thereby obtaining a verification image set that reflects the surface condition of the tool. The visual acquisition sensor may include a CCD sensor, a CMOS sensor, or the like.

[0027] In one possible embodiment, the start node and stop node of the same subtask are mapped to establish a start-stop node association. The visual acquisition sensor is then used to capture images of the tool after coolant cleaning at the start node and the stop node, respectively, to obtain a start image and a stop image. To minimize the impact of angular errors on the tool image, the start image and the stop image are captured at the same angle.

[0028] Optionally, based on the mapping relationship in the start node association, the start image and the stop image are aligned, that is, the start image and the stop image are overlapped, and perspective recognition is performed based on the alignment result to establish the abnormal feature identifier. The abnormal feature identifier is used to describe the abnormal characteristics of the tool state after the tool executes the subtask. Then, the abnormal feature identifier, the start image, and the stop image are used as the verification image set.

[0029] In one possible embodiment, the tool data describes the basic characteristics of the tool, including tool dimensions, tool tip geometry (tip shape, blade angle, blade length, etc.), and tool thickness. Based on the different tool positions reflected in the tool data, the tool is partitioned, and an importance index is constructed for each tool partition. Optionally, the tool dimensions, tool geometry, and tool thickness are input into a feature partitioning unit for partition identification, generating a partition identification result. The partition identification result includes the tool partitions and the importance index for each tool partition.

[0030] Preferably, a plurality of sample tool dimensions, a plurality of sample tool geometric parameters, a plurality of sample tool thicknesses, and a plurality of sample partition identification results are obtained as training data, wherein the plurality of sample partition identification results are obtained by a person skilled in the art after manually partitioning and marking the plurality of sample tools. A neural network model is constructed based on a convolutional neural network, and then key parameters such as the number of network layers, the number of neurons in each layer, and the activation function are determined by the staff. The convolutional neural network is then supervised trained using the training data until the model reaches convergence, thereby obtaining the trained feature partitioning unit.

[0031] The recognition position of perspective recognition is obtained, and the recognition position is used as an index to search based on the tool partition to obtain the corresponding tool partition, and the importance index of the corresponding tool partition is mapped to the recognition position to obtain an importance index mapping result. The importance index mapping result reflects the importance of the recognition position.

[0032] Perspective recognition is performed on the alignment results. Specifically, the similarity between the start image and the stop image at the perspective recognition position is analyzed to obtain a positional anomaly degree for perspective recognition. Optionally, image preprocessing is performed on the start image and the stop image at the perspective recognition position. Feature extraction is then performed on the start image and the stop image using a color histogram feature extraction algorithm to obtain a start feature vector and a stop feature vector. Similarity is calculated between the start feature vector and the stop feature vector using a cosine similarity calculation formula, and the inverse of the calculated result is used as the positional anomaly degree for perspective recognition.

[0033] Multiple identified locations, multiple positional anomaly degrees corresponding to the identified locations, and multiple important index mapping results corresponding to the identified locations are obtained. Weighted calculations are performed on the multiple positional anomaly degrees according to the multiple important index mapping results to obtain an anomaly value calculation result. The anomaly calculation result is used as an anomaly feature identifier. This achieves the technical effect of identifying abnormal wear of a tool after executing a task and improving monitoring reliability.

[0034] S500: Obtaining a quality inspection result of a workpiece, and establishing an execution exception of a subtask based on the association mapping; Furthermore, the quality inspection result of the workpiece is obtained, and the execution exception of the subtask is established based on the association mapping. In the embodiment of the present application, step S500 further includes: Establishing a quality abnormality level based on the quality test results; Performing an abnormal matching probability evaluation of the corresponding subtasks and quality inspection results according to the association mapping; An execution exception of the subtask is established based on the quality exception level and the exception matching probability evaluation.

[0035] In one possible embodiment, the workpiece is subjected to quality inspection for corresponding items based on the quality-inspectable parameters to generate the quality inspection results. The quality inspection results reflect the quality of the processing effect of the mineral material machine tool on the workpiece. Based on the quality inspection results and the association map, subtask execution anomalies are determined. Because different inspection items in the quality inspection results correspond to different quality-inspectable parameters, and each quality-inspectable parameter corresponds to one or more subtasks, the subtask with the execution anomaly can be determined based on the quality inspection results and the association map.

[0036] Optionally, the quality anomaly level is obtained by evaluating the quality inspection results according to a preset quality evaluation table. The preset quality evaluation table is a quality evaluation table set by a person skilled in the art based on the actual conditions of the workpiece. The association map is searched using the quality-inspectable parameters in the quality inspection results as an index to obtain multiple matching subtasks. Subtask matching is performed based on the quality-inspectable parameters of quality inspection results with higher quality anomaly levels, based on the association map, to obtain multiple matching subtask sets.

[0037] Next, using the subtask type as an index, the multiple matching subtask sets are searched to obtain the number of matches for each subtask type, thereby obtaining multiple matching subtasks and multiple matching numbers. The multiple matching numbers are then divided by the sum of the multiple matching numbers to obtain multiple abnormal match probabilities for the multiple matching subtasks. These multiple abnormal match probabilities are used as abnormal match probability evaluations. Based on the quality abnormality level (reflecting the degree of abnormality) and the abnormal match probability evaluation (the likelihood of abnormality), execution abnormality construction is performed on the subtasks.

[0038] S600: Input the motion data set, the verification image set and the execution anomaly into the anomaly fusion authentication model, perform machine tool anomaly authentication, establish a machine tool anomaly warning, and complete real-time monitoring of the mineral material machine tool based on the machine tool anomaly warning.

[0039] In one possible embodiment, the abnormal fusion authentication model is used to comprehensively analyze the motion data set, the verification image set and the execution abnormality, and comprehensively determine whether there is an abnormality in the machine tool processing from three dimensions. When the authentication result is a machine tool abnormality authentication, a machine tool abnormality warning is constructed, and the machine tool abnormality warning is used as a real-time monitoring result of the mineral material machine tool.

[0040] Preferably, multiple sample motion data sets, multiple sample verification image sets, multiple sample execution anomalies, and multiple sample authentication results are obtained as model training data. The authentication results include machine tool anomaly authentication (indicating an abnormality in the machine tool operation) and machine tool normal authentication (indicating normal operation). The convolutional neural network-based model is trained in batches using the model training data. Model parameters are updated based on the output accuracy of the previous batch of model training data, and then training is performed on the next batch of model training data until the model converges, resulting in the completed anomaly fusion authentication model. This achieves the technical effect of improving the reliability and efficiency of anomaly authentication.

[0041] Furthermore, step S600 in the embodiment of the present application further includes: Configuring a vibration monitoring sensor to perform full-process vibration monitoring of the machine tool while performing a task, and establishing a vibration monitoring data set; Performing a global vibration evaluation on the vibration monitoring data set and establishing a global abnormal node; Calling the implicit cutting force monitoring data of the node window through the global abnormal node, performing abnormal authentication of the global abnormal node with the implicit cutting force monitoring data, and generating an abnormal authentication result; The abnormality authentication result is input into the abnormality fusion authentication model after incremental learning to perform machine tool abnormality authentication update.

[0042] In one possible embodiment, a vibration monitoring sensor is installed on the mineral material machine tool, and the vibration monitoring sensor is used to monitor the vibration of the machine tool throughout the entire process of executing a task to obtain the vibration monitoring data set. The vibration monitoring data set reflects the vibration fluctuation of the mineral material machine tool.

[0043] Optionally, the vibration monitoring data is analyzed for abnormal nodes using a vibration tolerance interval (the range of data allowed for machine tool vibration) set by a person skilled in the art to obtain a global abnormal node. The global abnormal node is the time point at which the vibration monitoring data exceeds the vibration tolerance interval. During operation of the mineral material machine tool, a force sensor is used to monitor cutting force, and the monitoring data is stored in a storage unit of the force sensor as implicit cutting force monitoring data.

[0044] After obtaining the global abnormal node, the system retrieves the implicit cutting force monitoring data for the node window (the time point at which the global abnormal node is located) from the force sensor's storage unit based on the global abnormal node. A determination is then made as to whether the implicit cutting force monitoring data at the global abnormal node exceeds the cutting force data threshold. If so, an abnormality authentication result is obtained. This abnormality authentication result is then input into the abnormality fusion authentication model after incremental learning to update the machine tool abnormality authentication.

[0045] Furthermore, step S600 in the embodiment of the present application further includes: Collect mineral casting data of mineral material machine tools based on maintenance cycles and establish mineral casting data sets; Performing abnormal mineral composition collection on the mineral casting data set, performing component abnormality analysis on the abnormal collection results through an early warning network, and establishing a mineral casting monitoring result; The mineral casting monitoring results and the machine tool abnormality warning are used to complete the real-time monitoring of the mineral material machine tool.

[0046] In an embodiment of the present application, the maintenance cycle is the time interval between two consecutive maintenance and inspections of a mineral material machine tool. Mineral casting data of the mineral material machine tool is collected based on the maintenance cycle to create a mineral casting dataset. The mineral casting dataset is used to describe the mineral composition of the mineral material machine tool.

[0047] The mineral casting data set is used to identify abnormal mineral components based on the mineral casting component content warning threshold, and the mineral components that meet the mineral casting component content threshold are used as abnormal collection results. The abnormal collection results are input into the early warning network for component abnormality analysis to establish a mineral casting monitoring result. The mineral casting monitoring result reflects the degree of abnormal mineral composition of the mineral material machine tool. Optionally, multiple sample abnormal collection results and multiple sample mineral casting monitoring results are obtained as network training data, and the convolutional neural network layer is trained using the network training data, so that the convolutional neural network layer performs supervised learning on the mapping relationship between the abnormal collection results and the mineral casting monitoring results until the model converges, and the trained early warning network is obtained. Furthermore, the mineral casting monitoring results and the machine tool abnormality warning are used as monitoring results for real-time monitoring of the mineral material machine tool. The technical effect of real-time monitoring of the mineral material machine tool from two dimensions, the machine tool's own status and the machine tool's operating status, is achieved.

[0048] In summary, the embodiments of the present application have at least the following technical effects: This application establishes a machine tool execution task for a mineral material machine tool, decomposes the task into N subtasks, then obtains the quality-checkable parameters of the workpiece and establishes a correlation mapping between the quality-checkable parameters and the N subtasks, paving the way for subsequent refined anomaly authentication. Furthermore, displacement sensors are used to monitor the motion of the N subtasks within the machine tool range within the task scope, establishing a motion dataset for the subtasks and providing data support for machine tool anomaly warnings from the dimension of displacement distance. The start and stop nodes of the N subtasks are then obtained. At the start and stop nodes, visual acquisition sensors are used to capture images of the tool after coolant cleaning, establishing a verification image set for the subtasks, obtaining the quality inspection results of the workpiece, and establishing execution anomalies for the subtasks based on the correlation mapping. The motion dataset, verification image set, and execution anomalies are then input into an anomaly fusion authentication model, machine tool anomaly authentication is performed, machine tool anomaly warnings are established, and real-time monitoring of the mineral material machine tool is completed based on the machine tool anomaly warnings. This achieves the technical effect of multi-dimensionally collecting and analyzing the operation status of the mineral material machine tool, fusion authentication of the multi-dimensional analysis results, and improving the reliability of machine tool anomaly warnings.

[0049] Example 2 Based on the same inventive concept as the real-time monitoring method for a mineral material machine tool in the aforementioned embodiment, Figure 2 As shown, the present application provides a real-time monitoring system for mineral material machine tools. The system and method embodiments in the present application are based on the same inventive concept. The system includes: A subtask construction module 11 is used to establish a machine tool execution task for a mineral material machine tool, decompose the machine tool execution task into N subtasks, wherein the task decomposition includes tool task decomposition and feed task decomposition; An association mapping establishment module 12 is used to obtain the quality inspection parameters of the workpiece and establish an association mapping between the quality inspection parameters and N subtasks, wherein one quality inspection parameter corresponds to one or more subtasks; A motion data set establishment module 13 is used to monitor the motion of the N subtasks within the machine tool range within the task range through a displacement sensor, and establish a motion data set of the subtasks; A verification image set establishment module 14 is configured to obtain the start and stop nodes of N subtasks, and to acquire images of the tool after coolant cleaning at the start and stop nodes using a visual acquisition sensor to establish a verification image set for the subtasks; An execution exception establishment module 15 is used to obtain the quality inspection result of the workpiece and establish the execution exception of the subtask based on the association map; The real-time monitoring module 16 is used to input the motion data set, the verification image set and the execution anomaly into the anomaly fusion authentication model, perform machine tool anomaly authentication, establish a machine tool anomaly warning, and complete real-time monitoring of the mineral material machine tool based on the machine tool anomaly warning.

[0050] Furthermore, the real-time monitoring module 16 is configured to perform the following steps: Configuring a vibration monitoring sensor to perform full-process vibration monitoring of the machine tool while performing a task, and establishing a vibration monitoring data set; Performing a global vibration evaluation on the vibration monitoring data set and establishing a global abnormal node; Calling the implicit cutting force monitoring data of the node window through the global abnormal node, performing abnormal authentication of the global abnormal node with the implicit cutting force monitoring data, and generating an abnormal authentication result; The abnormality authentication result is input into the abnormality fusion authentication model after incremental learning to perform machine tool abnormality authentication update.

[0051] Furthermore, the execution exception establishment module 15 is used to perform the following steps: Performing an abnormal matching probability evaluation of the corresponding subtasks and quality inspection results according to the association mapping; An execution exception of the subtask is established based on the quality exception level and the exception matching probability evaluation.

[0052] Furthermore, the verification image set establishing module 14 is configured to perform the following steps: Establishing a start-stop node association, and capturing images of the tool after coolant cleaning at the start node and the stop node, respectively, to create a start image and a stop image, wherein the start image and the stop image are images captured at the same angle; Perform image alignment of the start image and the stop image based on the start-stop node association, perform perspective recognition based on the alignment result, and establish an abnormal feature identification; A verification image set is established according to the abnormal feature identifier and the start image and the stop image.

[0053] Furthermore, the verification image set establishing module 14 is configured to perform the following steps: Acquire tool data, perform feature classification of the tool based on the tool data, and establish an important index for the tool partition; Obtain the recognition position of perspective recognition, perform recognition position mapping based on the important index of the tool partition, and obtain the important index mapping result; The position abnormality degree of perspective recognition is obtained, an abnormal value is calculated based on the position abnormality degree and the important index mapping result, and abnormal feature identification is completed based on the abnormal value calculation result.

[0054] Furthermore, the subtask construction module 11 is used to perform the following steps: Establish a task decomposition model, input the machine tool execution task into the task decomposition model, and perform task decomposition, specifically including: S1: Perform tool call decomposition of the machine tool execution task through the tool decomposition unit and establish the initial decomposition result; S2: Synchronizing the initial decomposition result to the feed decomposition unit, decomposing the feed amount task of the same tool through the feed decomposition unit, and establishing an updated decomposition result; S3: Synchronize the update decomposition result to the continuous decomposition unit, perform continuous analysis of the action of the update decomposition result through the continuous decomposition unit, further decompose the discontinuous action task, and generate N subtasks.

[0055] Furthermore, the real-time monitoring module 16 is configured to perform the following steps: Collect mineral casting data of mineral material machine tools based on maintenance cycles and establish mineral casting data sets; Performing abnormal mineral composition collection on the mineral casting data set, performing component abnormality analysis on the abnormal collection results through an early warning network, and establishing a mineral casting monitoring result; The mineral casting monitoring results and the machine tool abnormality warning are used to complete the real-time monitoring of the mineral material machine tool.

[0056] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0057] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0058] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A real-time monitoring method for a mineral material machine tool, characterized in that: The method comprises: Establishing a machine tool execution task for a mineral material machine tool, decomposing the machine tool execution task into N subtasks, wherein the task decomposition includes tool task decomposition and feed task decomposition; Obtain the quality inspection parameters of the workpiece and establish an association mapping between the quality inspection parameters and N subtasks, where one quality inspection parameter corresponds to one or more subtasks; Use displacement sensors to monitor the motion of N subtasks within the machine tool range within the task scope and establish a motion data set for the subtasks; Obtaining the start and stop nodes of N subtasks, collecting images of the tool after coolant cleaning using a visual acquisition sensor at the start and stop nodes, and establishing a verification image set for the subtasks; Obtaining a quality inspection result of the workpiece, and establishing an execution exception of the subtask based on the association mapping; The motion data set, the verification image set and the execution anomaly are input into an anomaly fusion authentication model, machine tool anomaly authentication is performed, a machine tool anomaly warning is established, and real-time monitoring of mineral material machine tools is completed based on the machine tool anomaly warning.

2. The method according to claim 1, wherein The method further comprises: Configuring a vibration monitoring sensor to perform full-process vibration monitoring of the machine tool while performing a task, and establishing a vibration monitoring data set; Performing a global vibration evaluation on the vibration monitoring data set and establishing a global abnormal node; Calling the implicit cutting force monitoring data of the node window through the global abnormal node, performing abnormal authentication of the global abnormal node with the implicit cutting force monitoring data, and generating an abnormal authentication result; The abnormality authentication result is input into the abnormality fusion authentication model after incremental learning to perform machine tool abnormality authentication update.

3. The method according to claim 1, wherein The obtaining of the quality inspection result of the workpiece and establishing the execution exception of the subtask based on the association mapping further includes: Establishing a quality abnormality level based on the quality test results; Performing an abnormal matching probability evaluation of the corresponding subtasks and quality inspection results according to the association mapping; An execution exception of the subtask is established based on the quality exception level and the exception matching probability evaluation.

4. The method according to claim 1, wherein The image acquisition of the tool after coolant cleaning is performed by the visual acquisition sensor at the start-stop node to establish a verification image set for the subtask, including: Establishing a start-stop node association, and capturing images of the tool after coolant cleaning at the start node and the stop node, respectively, to create a start image and a stop image, wherein the start image and the stop image are images captured at the same angle; Perform image alignment of the start image and the stop image based on the start-stop node association, perform perspective recognition based on the alignment result, and establish an abnormal feature identification; A verification image set is established according to the abnormal feature identifier and the start image and the stop image.

5. The method according to claim 4, wherein The image alignment of the start image and the stop image based on the association of the start and stop nodes, and perspective recognition based on the alignment result to establish an abnormal feature identification further includes: Acquire tool data, perform feature classification of the tool based on the tool data, and establish an important index for the tool partition; Obtain the recognition position of perspective recognition, perform recognition position mapping based on the important index of the tool partition, and obtain the important index mapping result; The position abnormality degree of perspective recognition is obtained, an abnormal value is calculated based on the position abnormality degree and the important index mapping result, and abnormal feature identification is completed based on the abnormal value calculation result.

6. The method according to claim 1, wherein Decomposing the machine tool execution task into N subtasks includes: Establish a task decomposition model, input the machine tool execution task into the task decomposition model, and perform task decomposition, specifically including: S1: Perform tool call decomposition of the machine tool execution task through the tool decomposition unit and establish the initial decomposition result; S2: Synchronizing the initial decomposition result to the feed decomposition unit, decomposing the feed amount task of the same tool through the feed decomposition unit, and establishing an updated decomposition result; S3: Synchronize the update decomposition result to the continuous decomposition unit, perform continuous analysis of the action of the update decomposition result through the continuous decomposition unit, further decompose the discontinuous action task, and generate N subtasks.

7. The method according to claim 1, wherein The method further comprises: Collect mineral casting data of mineral material machine tools based on maintenance cycles and establish mineral casting data sets; Performing abnormal mineral composition collection on the mineral casting data set, performing component abnormality analysis on the abnormal collection results through an early warning network, and establishing a mineral casting monitoring result; The mineral casting monitoring results and the machine tool abnormality warning are used to complete the real-time monitoring of the mineral material machine tool.

8. A real-time monitoring system for a mineral material machine tool, characterized in that: The system comprises: A subtask construction module is used to establish a machine tool execution task for a mineral material machine tool, decompose the machine tool execution task into N subtasks, wherein the task decomposition includes tool task decomposition and feed task decomposition; An association mapping establishment module is used to obtain the quality inspection parameters of the workpiece and establish an association mapping between the quality inspection parameters and N subtasks, wherein one quality inspection parameter corresponds to one or more subtasks; A motion data set establishment module is used to monitor the motion of N subtasks within the machine tool range within the task scope through a displacement sensor and establish a motion data set for the subtasks; A verification image set establishment module is used to obtain the start and stop nodes of N subtasks, and to acquire images of the tool after coolant cleaning at the start and stop nodes using a visual acquisition sensor to establish a verification image set for the subtasks; an execution exception establishment module, configured to obtain a quality inspection result of a workpiece and establish an execution exception of a subtask based on the association mapping; The real-time monitoring module is used to input the motion data set, the verification image set and the execution anomaly into the anomaly fusion authentication model, perform machine tool anomaly authentication, establish a machine tool anomaly warning, and complete real-time monitoring of the mineral material machine tool based on the machine tool anomaly warning.

Images