Machining factory management method and system based on digital twinning

Abstract

The invention relates to the technical field of digital twinning modeling, in particular to a machining factory management method and system based on digital twinning, and the method comprises the following steps: a workshop sensor collects the temperature data of a machining area, extracts path nodes, marks an over-limit segment in combination with a thermal deformation threshold value to generate an early warning group, and binds threads according to an offset sequence to generate a list; and extracting main shaft load and feed rate correction simulation parameters to generate an instruction set, and loading a three-dimensional model to record prediction and actual deviation to generate a verification report. Through linkage analysis of a temperature sensing network and a tool path, dynamic identification and early warning of a thermally induced error area are realized, path offset sensing precision is improved, a resource priority binding mechanism is established based on calculation thread load and offset severity, resource scheduling efficiency is enhanced, simulation iteration parameters are corrected in combination with processing feedback, and the method is suitable for large-scale popularization and application. Visual feedback and error records are corrected through a three-dimensional path, dynamic consistency verification of the machining process is achieved, and process stability and control precision are enhanced.

Description

Technical Field

[0001] This invention relates to the field of digital twin modeling technology, and in particular to a method and system for managing machining plants based on digital twins. Background Technology

[0002] The field of digital twin modeling technology encompasses the synchronization and interaction of information between physical and digital spaces by constructing virtual mapping models of physical entities. The core of this technology lies in establishing a two-way mapping relationship between digital models and real-world objects, enabling real-time perception, analysis, and control of the state, behavior, and changes of physical objects. This field involves modeling theory, real-time data acquisition and synchronization, virtual-physical fusion computing mechanisms, model-driven simulation and optimization, and is widely applied in scenarios such as industrial manufacturing, equipment maintenance, and urban management. Digital twin modeling achieves dynamic mirroring of physical systems through multi-source data fusion, semantic modeling, and system simulation, making it a key foundational technology for realizing intelligent, information-based, and refined management of manufacturing systems.

[0003] Among them, the digital twin-based machining factory management method refers to constructing a digital twin model of machining equipment, production line processes, and workshop layout. This model comprehensively models and represents the spatial configuration, process paths, work sequences, and state changes of physical resources within the factory. Combined with actual process parameters and real-time sensor data input, the physical machining process in the model is mapped and virtually driven in real time. Specifically, this method encompasses multi-dimensional scheduling modeling of machining tasks, virtual visualization construction of machining paths, temporal logic arrangement of process nodes, real-time annotation of machining resource usage status, and the use of rule-based reasoning and temporal control logic to manage the machining process, forming a configurable simulation management unit that can be used in factory management systems. The above process achieves information fusion and expression using a unified data format and model structure, typically combining finite state control logic, task-driven methods oriented towards structured flowcharts, and coordinate mapping-based 3D modeling techniques.

[0004] While current digital twin modeling technology possesses the capability to model and represent physical systems in real time, it exhibits a lag in responding to environmental disturbances during machining processes. Thermal field changes, a significant disturbance factor in machining, exhibit random and dynamic spatial diffusion. Existing technologies lack proactive identification and labeling mechanisms for the interaction between temperature distribution and machining paths, making it difficult to provide timely warnings of thermally induced errors and easily leading to deviations in the dimensional accuracy of machined parts. Furthermore, existing solutions often employ static thread binding strategies at the computational resource allocation level, lacking dynamic awareness of the simulation resource load status. This results in slow response when handling highly complex path correction tasks, hindering efficient resource utilization. Regarding virtual-real coupling, existing models largely rely on initially set process parameters and fixed correction strategies, making it difficult to flexibly adjust simulation behavior based on real-time data feedback. This leads to continuous discrepancies between model predictions and actual equipment execution. For example, in a workshop performing precision hole machining in a high-temperature environment, the failure to detect tool displacement caused by localized temperature rises resulted in hole position deviations exceeding the tolerance limit, severely impacting product assembly accuracy. This demonstrates the shortcomings of existing solutions in high-precision control and environmental adaptability. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a digital twin-based method and system for managing machining plants.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a machine shop management method based on digital twins, comprising the following steps:

[0007] S1: Through the workshop-wide temperature sensor network, real-time temperature data of the processing area is collected, the tool movement trajectory is compared with the temperature distribution change trend, the range of abnormal temperature fluctuations within the trajectory coverage area is identified, and a dynamic temperature field influence range boundary map is generated.

[0008] S2: Based on the boundary map of the influence range of the dynamic temperature field, extract the set of coordinates of the processing path nodes within the boundary line, combine them with the preset equipment thermal deformation tolerance threshold, mark the starting point and directional offset of the path segment that exceeds the tolerance range, and generate a thermal path offset warning mark group.

[0009] S3: Invoke the thermal path offset warning marker group, scan the thread load status in the digital twin computing resource pool, sort the idle computing threads according to the severity of the offset of the warning path segment, establish the priority binding relationship between thread resources and path segments, and generate a dynamic allocation list of virtual entity resources.

[0010] S4: Based on the aforementioned virtual and real computing resource dynamic allocation list, extract the spindle load data and feed rate data fed back in real time by the physical processing equipment, correct the simulation iteration number and motion trajectory control parameters of the corresponding path segment in priority order, and generate a virtual and real synchronous path correction instruction set.

[0011] As a further aspect of the present invention, the dynamic temperature field influence range boundary map includes an abnormal temperature distribution area, a trajectory-related spatial range, and a boundary change trend graph; the thermally induced path deviation early warning marker group includes a path segment start point marker, a direction deviation marker, and a set of out-of-tolerance coordinates; the virtual entity resource dynamic allocation list includes a thread resource number, a deviation severity level, and a resource path segment binding relationship; and the virtual-real synchronous path correction instruction set includes an iteration number correction value, a motion trajectory control parameter adjustment value, and a path segment matching number.

[0012] As a further aspect of the present invention, the specific steps of S1 are as follows:

[0013] S101: Acquire the temperature values ​​of the collection points in the temperature sensor network covering the entire workshop, establish a mapping relationship between the temperature data and the processing area grid based on the sensor spatial coordinates, and generate the temperature change trend value of the processing grid by combining the temperature time series change of the grid area.

[0014] S102: Based on the temperature change trend value of the machining grid, call the path segment coordinate sequence in the tool motion trajectory and perform spatial matching with the positional relationship of the machining grid, extract the temperature change rate in the area covered by the trajectory path, perform difference analysis on the temperature sequence of the matched area, and obtain the abnormal temperature interval value of the trajectory path.

[0015] S103: Based on the abnormal temperature range values ​​of the trajectory path, extract the corresponding interval boundary coordinate points, connect the boundary points in spatial order to form a polyline structure, and construct a boundary closed path by combining the positional relationship of adjacent path segments to generate a boundary map of the dynamic temperature field influence range.

[0016] As a further aspect of the present invention, the specific steps of S2 are as follows:

[0017] S201: Based on the boundary map of the influence range of the dynamic temperature field, extract the set of coordinates of the processing path nodes within the boundary line, combine the time information and temperature data of each coordinate point, perform correlation analysis on the temperature change trend and spatial location between nodes, and generate a temperature influence correlation coefficient value.

[0018] S202: Based on the temperature influence correlation coefficient value, cluster the node coordinates, identify the temperature change peak and position offset peak in the cluster points, and combine the thermal variation deviation threshold set in the twin model to filter out path segments that exceed the threshold and generate path segment offset marker coordinate values.

[0019] S203: For the path segment direction change corresponding to each set of coordinates in the path segment offset marker coordinate values, extract the angle difference between the spatial direction of the coordinate point and the original movement direction of the path, establish a reference table of the relationship between the starting position and the direction change of the path segment, and generate a thermal path offset warning marker group.

[0020] As a further aspect of the present invention, the specific steps of S3 are as follows:

[0021] S301: Based on the thermal path offset warning marker group, call the path segment offset severity and marker number, extract the thread load data, running status and idle cycle length, filter thread resources that meet the conditions according to the running status and load threshold, and generate an idle thread resource list.

[0022] S302: Call the idle thread resource list and thermal path offset warning flag group, prioritize the path segment offset severity, sort according to the thread idle cycle, establish the correspondence between path segments and thread resources, and generate a path segment resource priority binding relationship table;

[0023] S303: Based on the path segment resource priority binding relationship table, integrate the path segment direction offset, space number, priority level and thread resource label to construct a resource allocation structure and generate a dynamic allocation list of virtual entity resources.

[0024] As a further aspect of the present invention, the specific calculation formula for the extracted thread's load data, running status, and idle cycle length is as follows:

[0025]

[0026] Among them, L′ i ω represents the overall load fluctuation value of the i-th thread. ij f represents the weight value of the load of the i-th thread in the overall thread pool during the j-th monitoring. ij τ represents the instantaneous load value of the i-th thread in the j-th monitoring. ij u represents the idle period length of the i-th thread in the j-th monitoring. ij d represents the task running status value of the i-th thread in the j-th monitoring. ij y represents the task processing time of the i-th thread in the j-th monitoring, and y represents the total number of data records within the monitoring period. This represents the average load value of the i-th thread during the monitoring period.

[0027] As a further aspect of the present invention, the specific steps of S4 are as follows:

[0028] S401: Obtain the thread resource number, path segment number and priority order value from the virtual and real computing resource dynamic allocation list, extract the spindle load time series data and feed rate synchronization data corresponding to the path segment, complete the pairing and sorting of the two types of data, and generate load and feed paired sequence value group.

[0029] S402: Based on the spindle load value and feed rate value of the path segment in the load and feed pairing sequence value group, and combined with the motion trajectory control parameters and thread occupancy status of the path segment, extract the parameter offset information and obtain the path segment control parameter offset coefficient group.

[0030] S403: Based on the offset value and path segment number in the path segment control parameter offset coefficient group, and combined with the priority order in the resource allocation list, the simulation iteration number and motion trajectory control parameters of the corresponding path segments are corrected in sequence to generate a virtual-real synchronous path correction instruction set.

[0031] As a further aspect of the present invention, the specific calculation formula for the priority order in the resource allocation list is as follows:

[0032]

[0033] in, W represents the priority adjustment value required for the i-th path segment. i R represents the initial weight coefficient of the i-th path segment in the path set. j Δ represents the available score of the j-th resource in the resource allocation list. ij This represents the offset value indicating the importance of resource j to path segment i. This represents the sample variance of the number of iterations for path segment i in the historical simulation. μ represents the weighted average of all resource scores assigned to path segment i. i This represents the average resource score of path segment i among similar paths.

[0034] As a further aspect of the present invention, the method further includes:

[0035] S5: Based on the virtual-real synchronization path correction instruction set, load the three-dimensional visualization model of the correction path in the twin interface, record the deviation value between the twin model prediction data and the actual execution data of the device, and generate a dynamic path virtual-real consistency verification report.

[0036] The dynamic path virtual-real consistency verification report includes a three-dimensional correction model, prediction and execution deviation records, and path execution offset trend chart.

[0037] The specific steps of S5 are as follows:

[0038] S501: Based on the path segment coordinates, offset values ​​and direction parameters in the virtual-real synchronous path correction instruction set, call the twin platform modeling resources, extract the path segment spatial node data and processing position information, load the path graphic system structure according to the node arrangement and geometric constraints, and obtain the three-dimensional path structure primitive set.

[0039] S502: Based on the three-dimensional path structure primitive set, obtain the node trajectory values ​​in the twin model and the real-time trajectory data fed back by the device, match the two types of trajectory values ​​according to the node index, identify the nodes whose offset amplitude exceeds the dynamic error warning value, and obtain the path offset numerical feature group by combining direction and frequency information.

[0040] S503: Call the path offset numerical feature group and path segment number, overlay the three-dimensional path structure primitive morphology data, construct the offset annotation layer in the twin interface, record the difference between the predicted and actual values ​​according to the processing time sequence of the path segment, and generate the path virtual-real consistency deviation trend value.

[0041] A digital twin-based machining factory management system includes:

[0042] The temperature monitoring module collects global temperature sensor data and tool trajectory coordinates, interpolates to generate a temperature distribution surface, calculates the absolute difference between the temperature change rate at the trajectory coordinate points and the surface change rate, and if the difference exceeds the temperature fluctuation threshold, extracts the coordinates of abnormal temperature points and generates a dynamic temperature field influence range boundary map.

[0043] The path warning module extracts the vertex coordinates of the polygon in the boundary map of the influence range of the dynamic temperature field, traverses the coordinates of the path nodes, calculates the Euclidean distance between the nodes and the boundary, compares it with the thermal distortion deviation threshold, filters nodes with excessive distance, marks the starting point coordinates and three-axis offset of the corresponding path segment, and generates a thermal path deviation warning mark group.

[0044] The resource allocation module calls the offset of the thermal path offset warning marker group, detects and calculates the resource pool thread occupancy rate, filters idle threads and sorts them in descending order of offset, binds the largest offset path segment to the thread with the largest idle memory, establishes a thread-path segment mapping table, and generates a virtual entity resource dynamic allocation list.

[0045] The parameter correction module extracts the spindle load current and feed rate pulses according to the virtual entity resource dynamic allocation list, converts the load current into a torque coefficient, converts the pulses into displacement increments, adjusts the torque and displacement parameters of the simulation model path segment according to priority, and generates a virtual-real synchronous path correction instruction set.

[0046] The verification feedback module loads the virtual-real synchronous path correction instruction set into the 3D model, extracts the coordinate positioning deviation value, speed following deviation value, and trajectory curvature deviation value of the correction path segment, compares the digital twin prediction data with the actual execution data of the physical device, integrates the distribution range, mean, and standard deviation of the deviation value, and generates a dynamic path virtual-real consistency verification report.

[0047] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0048] In this invention, dynamic identification and early warning of thermal error areas are achieved through the linkage analysis of temperature sensing network and tool trajectory, thereby improving the accuracy of path deviation perception. A resource priority binding mechanism is established based on the computing thread load and deviation severity to enhance resource scheduling efficiency. Simulation iteration parameters are corrected in conjunction with machining feedback to improve the accuracy of virtual-real synchronization. Dynamic consistency verification of the machining process is achieved through three-dimensional path correction visual feedback and error recording, thereby strengthening process stability and control accuracy. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the steps of the present invention.

[0050] Figure 2 This is a system module diagram of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0052] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0053] Please see Figure 1 A digital twin-based method for managing machining plants includes the following steps:

[0054] S1: Through the workshop-wide temperature sensor network, real-time temperature data of the processing area is collected, the tool movement trajectory is compared with the temperature distribution change trend, the range of abnormal temperature fluctuations within the trajectory coverage area is identified, and a dynamic temperature field influence range boundary map is generated.

[0055] Temperature field, a well-known term, refers to the spatial distribution of temperature.

[0056] S2: Based on the boundary map of the influence range of the dynamic temperature field, extract the set of coordinates of the processing path nodes within the boundary line, combine the preset equipment thermal deformation tolerance threshold in the twin model, mark the starting point and directional offset of the path segment that exceeds the tolerance range, and generate a thermal path offset warning mark group.

[0057] The thermal deformation tolerance range is a well-known parameter in the field of machining, representing the allowable limit of material deformation under heat.

[0058] S3: Call the thermal path offset warning flag group, scan the thread load status in the digital twin computing resource pool, sort the idle computing threads according to the severity of the offset of the warning path segment, establish the priority binding relationship between thread resources and path segments, and generate a dynamic allocation list of virtual entity resources.

[0059] Thread Load Status is a well-known parameter in the computer field, representing the utilization rate of processor threads;

[0060] Virtual-Physical entities conform to the definition of digital twin technology;

[0061] S4: Based on the dynamic allocation list of virtual and real computing resources, extract the spindle load data and feed rate data fed back by the physical processing equipment in real time, and correct the simulation iteration number and motion trajectory control parameters of the corresponding path segment in the twin model according to priority order to generate a virtual and real synchronous path correction instruction set.

[0062] Spindle load data and feed rate data are standard output parameters of CNC machining equipment.

[0063] Simulation iterations and kinematic parameters are well-known concepts in the field of digital twins.

[0064] S5: Execute the virtual-real synchronization path correction instruction set, load the three-dimensional visualization model of the correction path in the twin interface, record the deviation value between the twin model prediction data and the actual execution data of the device, and generate a dynamic path virtual-real consistency verification report.

[0065] 3D visualization model and trajectory verification program are well-known technology modules in the manufacturing industry.

[0066] The dynamic temperature field influence range boundary map includes temperature anomaly distribution areas, trajectory-related spatial range, and boundary change trend graphs. The thermal path deviation early warning marker group includes path segment start point markers, direction deviation markers, and out-of-tolerance coordinate sets. The virtual entity resource dynamic allocation list includes thread resource numbers, deviation severity levels, and resource path segment binding relationships. The virtual-real synchronization path correction instruction set includes iteration number correction values, motion trajectory control parameter adjustment values, and path segment matching numbers. The dynamic path virtual-real consistency verification report includes a three-dimensional correction model, prediction and execution deviation records, and path execution deviation trend graphs.

[0067] The specific steps of S1 are as follows:

[0068] S101: Acquire the temperature values ​​of the collection points in the temperature sensor network covering the entire workshop, establish a mapping relationship between the temperature data and the processing area grid based on the sensor spatial coordinates, and generate the temperature change trend value of the processing grid by combining the temperature time series change of the grid area.

[0069] The workshop is equipped with numerous temperature acquisition nodes, fixed in various spatial locations such as corners, overhead beams, or around equipment. The acquisition frequency is set to once per minute, and each node carries spatial coordinate information, numbered sequentially from T001 to T100. The spatial locations of the sensors are mapped to a two-dimensional grid of the processing area, typically with a grid size of 5 cm square. For example, sensor number T045 has coordinates of (2.5 m, 1.5 m, 3 m), and its two-dimensional projection in the processing area corresponds to the 50th row and 30th column of the grid. Each sensor continuously records temperature time-series data. For instance, starting at 2 PM, a node records temperatures of 27.5℃, 27.7℃, 28.1℃, 28.4℃, 28.7℃, and 28.9℃ up to 2:05 PM. These temperature records are then populated into the corresponding grid cells, forming a multi-time, multi-grid temperature change map. The temperature value sequence within each grid area is modeled for its trend, typically selecting changes over a 10-minute period, and the simplest linear trend method is used to calculate the minute-by-minute temperature change. For example, if the temperature of a certain grid rises from 26.8℃ to 29.3℃ within 10 minutes, the trend is an increase of 0.25℃ per minute, which is recorded as the trend value of that grid. For grids without sensor coverage, the trend values ​​of adjacent grids can be interpolated to complete the trend map of the entire area, ultimately forming a two-dimensional temperature change trend map covering the entire processing area. Once constructed, this provides reference data for various trajectory analyses.

[0070] S102: Based on the temperature change trend value of the machining grid, call the path segment coordinate sequence in the tool motion trajectory and perform spatial matching with the positional relationship of the machining grid. Extract the temperature change rate in the area covered by the trajectory path, perform difference analysis on the temperature sequence of the matched area, and obtain the abnormal temperature interval value of the trajectory path.

[0071] The tool movement trajectory information originates from the coordinate instructions in the machining program and can be broken down into multiple path segments, each consisting of a start point and an end point. For example, a path segment from point (1.0, 1.0) to point (2.0, 1.5) represents a linear machining path. Based on the spatial position of each point in the path segment, it is determined which machining grid areas the path segment traverses. The areas traversed by the path may cover grids numbered G12, G13, and G14. The temperature change trend values ​​of these grids are extracted, for example, 0.2℃ per minute, 0.4℃ per minute, and 0.3℃ per minute, respectively. This set of values ​​is used as the thermal information set of the path segment, and the degree of fluctuation between these values ​​is analyzed. Fluctuation can be determined by whether the degree of change exceeds a set threshold, for example, a fluctuation judgment limit of 0.1℃ per minute. When the temperature trend difference within the area covered by a path segment exceeds this value, it can be preliminarily determined that the temperature environment of the path segment is unstable and belongs to an abnormal trajectory segment. By further combining time tags, the time and grid location of the prominent temperature changes are extracted. For example, the temperature of G13 rose from 26.8℃ to 29.5℃ between 14:01 and 14:03, which is a significant increase in a very short period of time. This time period is identified as an abnormal temperature range. The corresponding grid number and time range are extracted as the key area for subsequent analysis.

[0072] S103: Based on the abnormal temperature interval values ​​of the trajectory path, extract the corresponding interval boundary coordinate points, connect the boundary points in spatial order to form a polyline structure, combine the positional relationship of adjacent path segments to construct a boundary closed path, and generate a dynamic temperature field influence range boundary map.

[0073] After determining the abnormal temperature range, the spatial coordinates of the abnormal grids need to be further extracted. Each grid corresponds to a center position; for example, G13 represents coordinates (2.0 m, 1.5 m). If multiple grids are continuously abnormal, multiple center points can be aggregated to form a point set, such as containing coordinates (2.0 m, 1.5 m), (2.05 m, 1.5 m), and (2.1 m, 1.5 m). These coordinate points are sorted according to the direction of the processing trajectory, and adjacent points are connected sequentially to form a polygonal line structure, initially depicting the path distribution shape of the abnormal temperature trajectory. If the trajectory moves first along the X direction and then along the Y direction, the polygonal line forms a tortuous structure. Based on the position and turning relationship between adjacent path segments, the abnormal point set is constructed as a closed boundary. The convex hull algorithm is used to encapsulate the point set, forming a closed region graphic. This closed path can serve as a graphic identifier representing the range of influence of temperature abrupt changes. The boundary line is marked on the two-dimensional coordinate graph. The area contained within the region can be estimated by the number of grids; for example, if it contains a grid of 2 rows and 5 columns, the total area is 0.05 square meters. Once the boundary map is completed, it can be saved as a data format file and applied to processing monitoring and parameter correction to form visualized information on the boundary effects of temperature anomalies.

[0074] The specific steps of S2 are as follows:

[0075] S201: Based on the boundary map of the influence range of dynamic temperature field, extract the set of coordinates of processing path nodes within the boundary line, combine the time information and temperature data of each coordinate point, perform correlation analysis on the temperature change trend and spatial location between nodes, and generate temperature influence correlation coefficient value.

[0076] The boundary map of the dynamic temperature field influence range can be constructed using continuously recorded thermal imaging data. In image processing, edge detection methods are used to identify the contour lines of the hot zone, and the image pixel coordinates are converted into actual two-dimensional coordinate points within the processing area. This is then combined with machine tool motion trajectory data to match the positions of path nodes within the hot zone. Each path node is attached to the temperature information recorded by the temperature monitoring system according to a time series, constructing a temperature change sequence over time for each node. Temperature changes between adjacent nodes can be calculated using the difference between the two points, while spatial position changes are measured by the distance between them. The ratio of the temperature change value to the spatial change value is processed to obtain the rate of temperature change with space. This rate of change is statistically analyzed for each pair of adjacent nodes and used as a primary analysis item. To further eliminate local measurement errors, a sliding interval containing multiple nodes can be set. Within this interval, all rate of change values ​​are weighted by time to obtain the temperature influence correlation value of the node. This weight can be set based on the influence of neighboring nodes, such as assigning a higher proportion to points closer to the center node. Taking a certain region as an example, if the temperature change between nodes is 3.7 degrees and the corresponding spatial change is 1.4 millimeters, then the rate of change is 2.6 degrees per millimeter. Further processing by moving average yields a temperature influence correlation coefficient of approximately 2.75 degrees per millimeter for this node. This coefficient can be used for subsequent thermal influence aggregation analysis.

[0077] S202: Based on the temperature influence correlation coefficient value, cluster the node coordinates, identify the temperature change peak and position offset peak in the cluster points, combine the thermal variation deviation threshold set in the twin model, filter the path segments that exceed the threshold, and generate the path segment offset marker coordinate value.

[0078] Based on the temperature influence correlation coefficient, density clustering can be used to identify clusters of points close to each other along a path with similar temperature change characteristics. The clustering process requires setting appropriate neighborhood distances and minimum sample sizes. The distance can be selected as an appropriate multiple of the average distance between points on the path, and the sample size must ensure sufficient statistical stability. Within each cluster, the temperature influence correlation of its constituent nodes is extracted, the extreme values ​​are identified, and the range of variation is calculated. For path segments within the same cluster region, the actual position coordinates on the processing path are compared point-by-point with the theoretical design path coordinates to calculate the spatial offset length, forming a series of offset value sequences. These sequences are then compared with the set thermal deformation error limit to filter out path segments whose offset values ​​exceed the tolerance limit. This error limit is derived from material properties or manufacturing process specifications; for example, the allowable deformation limit for aluminum profiles can be set to ±0.25 mm. If the maximum offset value of a certain path segment is 0.32 mm, it is determined to be an abnormal offset segment. The coordinates of the starting and ending nodes of this path segment are marked as the output of the thermal deformation offset result.

[0079] S203: For the path segment direction change corresponding to each set of coordinates in the path segment offset marker coordinate value, extract the angle difference between the spatial direction of the coordinate point and the original movement direction of the path, establish a reference table of the relationship between the starting position and the direction change of the path segment, and generate a thermal path offset warning marker group.

[0080] For identified path deviation segments, the original direction of movement can be obtained by analyzing the path direction between the starting and ending points. Based on the coordinates of the actual path during thermal deformation, the current actual direction of movement is then determined. The difference in direction angle between the two is the path direction change angle. This angle difference serves as a basis for judging the path distortion or deviation trend and can be classified into different levels. For example, a change in direction angle within five degrees is considered a slight deviation, five to fifteen degrees a moderate deviation, and more than fifteen degrees a severe deviation. In practical implementation, if the original path direction is forty degrees and the deformed path direction is fifty-four degrees, the direction change is fourteen degrees, which can be considered a moderate deviation. This change angle is recorded along with the corresponding starting and ending coordinates of the path segment, forming a direction change checklist. If the direction change exceeds a set warning threshold, such as ten degrees, the path segment is marked as a thermally induced path deviation warning segment, ultimately forming a warning marker coordinate group containing multiple path segments for subsequent path correction and adjustment.

[0081] The specific steps for S3 are as follows:

[0082] S301: Based on the thermal path offset warning flag group, call the path segment offset severity and flag number to extract the thread's load data, running status and idle cycle length, filter thread resources that meet the conditions according to the running status and load threshold, and generate an idle thread resource list.

[0083] The specific formulas for calculating the thread's load data, running status, and idle period length are as follows:

[0084]

[0085] Among them, L′ i ω represents the overall load fluctuation value of the i-th thread. ij f represents the weight value of the load of the i-th thread in the overall thread pool during the j-th monitoring. ij τ represents the instantaneous load value of the i-th thread in the j-th monitoring. ij u represents the idle period length of the i-th thread in the j-th monitoring. ij d represents the task running status value (numerical discrete task status) of the i-th thread in the j-th monitoring. ij y represents the task processing time of the i-th thread in the j-th monitoring, and y represents the total number of data records within the monitoring period. This represents the average load value of the i-th thread during the monitoring period;

[0086] Detailed explanation of the formula and its calculation derivation:

[0087] Formula Parameter Details

[0088] y represents the total number of data records within the monitoring period. In practical applications, if the monitoring period is one working day, then n may be 24 (recording once per hour).

[0089] ω ij These are weight values, reflecting the relative importance of threads within different time periods. Weights are typically set based on the thread's business priority; for example, a thread for a critical task might have a weight of 1.5, while a thread for a regular task might have a weight of 1.0.

[0090] f ij It is the instantaneous load value of the i-th thread in the j-th monitoring, which is usually measured by the CPU utilization of the thread.

[0091] τ ij It is the idle period length of the i-th thread in the j-th monitoring, which can be obtained by recording the time the thread is inactive.

[0092] u ij These are task running status values, obtained through numerical representations of thread states (such as running, waiting, and sleeping). For example, running state = 2, waiting state = 1, and sleeping state = 0.

[0093] d ij Task processing time is the time required to complete a task, usually measured in milliseconds.

[0094] The average load value is calculated based on all f values ​​within the monitoring period. ij The average value.

[0095] Formula derivation and examples

[0096] The example data is set as follows:

[0097] Substitute the above data into the formula to perform the calculation:

[0098]

[0099] Through specific calculations, the overall load fluctuation value of each thread at different time points can be obtained. This result indicates that, for a thread resource management system, a low overall load fluctuation value signifies that the thread performs stably in terms of load and state maintenance, and is suitable to be considered a usable resource.

[0100] S302: Call the list of idle thread resources and the thermal path offset warning flag group, prioritize the path segment offset severity, sort according to the thread idle cycle, establish the correspondence between path segments and thread resources, and generate a path segment resource priority binding relationship table;

[0101] After generating the list of idle thread resources, it needs to be used in conjunction with path segment offset warning information. Resource allocation and matching are based on the severity of path segment offset as the priority level. First, a priority mapping relationship is established according to the path segment offset level, with the level decreasing sequentially from the highest severity level, and the priority increasing sequentially from level one. Next, all idle thread resources are sorted according to their predicted idle period length, with longer periods having higher priority. The sorting method can use a conventional order. After completion, starting from the path segment with the highest priority, thread resources that meet the idle period requirements are searched one by one. Only one pair of path segments and thread resources is bound at a time. When multiple path segments have the same offset level, their offset values ​​can be compared, with the one with the higher offset being matched first. The binding relationship adopts a one-to-one correspondence, establishing an association between the path segment number and the thread number, and recording its corresponding priority level. If the resource list is insufficient, the remaining path segments will not participate in this round of scheduling. The final generated path segment resource priority binding relationship table will list in detail the number, priority level, bound thread resource identifier, and idle period of each path segment, providing specific reference data for the subsequent scheduling module.

[0102] S303: Based on the path segment resource priority binding relationship table, integrate the path segment direction offset, space number, priority level and thread resource label to construct the resource allocation structure and generate a dynamic allocation list of virtual entity resources;

[0103] Based on existing path segment resource binding information, core elements such as path segment directional offset values, space numbers, priority levels, and thread resource identifiers are further integrated to construct a complete resource allocation structure. First, the directional offset is read from the path segment information. This data comes from continuous sampling by displacement sensors, and the average of multiple measurements is used to form a directional offset reference value. For example, by obtaining 2.1 degrees, 2.2 degrees, 2.4 degrees, 2.5 degrees, and 2.3 degrees from five measurements, the final directional offset can be calculated as 2.3 degrees. Subsequently, the corresponding space number is queried from the topology table according to the path segment number. For example, if a path segment belongs to the fifth space module, its space number is 05. Based on this, priority levels and thread resource numbers are added according to the binding relationship table. The above information is uniformly integrated into a resource allocation structure, which includes path segment identifiers, directional offset values, space numbers, priority levels, and corresponding thread resource numbers. Each set of information represents an independent resource binding item. By processing all path segment data one by one, a unified and data-clear dynamic resource allocation list is formed, in which each item is the direct basis for calling schedulable targets.

[0104] The specific steps of S4 are as follows:

[0105] S401: Obtain the thread resource number, path segment number and priority order value from the dynamic allocation list of virtual and real computing resources, extract the spindle load time series data and feed rate synchronization data corresponding to the path segment, complete the pairing and sorting of the two types of data, and generate load and feed paired sequence value groups.

[0106] First, extract the thread resource ID, path segment ID, and priority order value from the dynamic allocation list of computing resources. Associate each thread resource ID with its corresponding path segment ID and initially rank them according to the priority order value. For example, thread resource ID T001 is associated with path segments P015 and P016, with priority order values ​​of 1 and 3, indicating that the processing priority of path segment P016 is higher than that of P015. Subsequently, for each path segment, extract the spindle load time series and feed rate synchronization data during the machining process from the monitoring records or simulation logs of the relevant equipment. The spindle load value is usually expressed as a percentage to represent the power consumption of the spindle in each time period, while the feed rate represents the physical displacement speed of the path segment per unit time. For example, in path segment P015, the spindle load values ​​are recorded as 50%, 52%, and 48% at three time points during machining, and the feed rates are recorded as 1200, 1180, and 1210 mm / min. These two types of data are matched one-to-one at each time point to form a synchronization sequence. Each pair of values ​​represents the combination of spindle load and feed rate at a certain time point. Next, according to the priority order of the path segments, the paired data of each path segment are arranged in sequence, with the path segments with higher priority placed first. If P016 has a higher priority than P015, then the data group of P016 should be arranged first in the final pairing sequence. Then, the paired data of other path segments are spliced ​​in sequence to complete the construction of a complete pairing sequence of load and feed rate in the entire machining path segment.

[0107] S402: Based on the spindle load value and feed rate value of the path segment in the load and feed pairing sequence value group, combined with the motion trajectory control parameters and thread occupancy status of the path segment, extract the parameter offset information and obtain the path segment control parameter offset coefficient group;

[0108] First, extract the thread resource ID, path segment ID, and priority order value from the dynamic allocation list of computing resources. Associate each thread resource ID with its corresponding path segment ID and initially rank them according to the priority order value. For example, thread resource ID T001 is associated with path segments P015 and P016, with priority order values ​​of 1 and 3, indicating that the processing priority of path segment P016 is higher than that of P015. Subsequently, for each path segment, extract the spindle load time series and feed rate synchronization data during the machining process from the monitoring records or simulation logs of the relevant equipment. The spindle load value is usually expressed as a percentage to represent the power consumption of the spindle in each time period, while the feed rate represents the physical displacement speed of the path segment per unit time. For example, in path segment P015, the spindle load values ​​are recorded as 50%, 52%, and 48% at three time points during machining, and the feed rates are recorded as 1200, 1180, and 1210 mm / min. These two types of data are matched one-to-one at each time point to form a synchronization sequence. Each pair of values ​​represents the combination of spindle load and feed rate at a certain time point. Next, according to the priority order of the path segments, the paired data of each path segment are arranged in sequence, with the path segments with higher priority placed first. If P016 has a higher priority than P015, then the data group of P016 should be arranged first in the final pairing sequence. Then, the paired data of other path segments are spliced ​​in sequence to complete the construction of a complete pairing sequence of load and feed rate in the entire machining path segment.

[0109] S403: Based on the offset value and path segment number in the path segment control parameter offset coefficient group, and combined with the priority order in the resource allocation list, the simulation iteration number and motion trajectory control parameters of the corresponding path segment are corrected in sequence to generate a virtual-real synchronous path correction instruction set.

[0110] The specific calculation formula for the priority order in the resource allocation list is as follows:

[0111]

[0112] in, W represents the priority adjustment value required for the i-th path segment. i R represents the initial weight coefficient of the i-th path segment in the path set. j Δ represents the available score of the j-th resource in the resource allocation list. ij This represents the offset value indicating the importance of resource j to path segment i. This represents the sample variance of the number of iterations for path segment i in the historical simulation. μ represents the weighted average of all resource scores assigned to path segment i. i This represents the average resource score of path segment i among similar paths;

[0113] Detailed explanation of the formula and its calculation derivation:

[0114] Given the priority adjustment value of path segment i Obtained through actual data, including:

[0115] W i : The initial weight coefficient for path i, determined by historical data analysis. It is set to 0.85, based on the average value derived from historical task completion speed and quality assessments.

[0116] R j The available score for resource j is obtained based on resource usage frequency and importance monitoring. Assume the scores for resource 1 and resource 2 are 4.5 and 3.7 respectively.

[0117] Δ ij The importance offset of resource j to path segment i is obtained through regression analysis of past project data. For resource 1 and resource 2, the offset values ​​are set to 0.8 and 0.65, respectively.

[0118] The sample variance of the number of iterations of path segment i in historical simulations is obtained through statistical analysis of past data. This value is 1.2.

[0119] The weighted average of all resource scores assigned to path segment i is calculated by multiplying each resource score by its corresponding weight and then dividing by the sum of those weights. Consider resource 1 and resource 2, with weights of 0.6 and 0.4, respectively.

[0120] μ i The average resource score of path segment i among similar paths, obtained through market research and internal resource usage reports, is set to 3.8.

[0121] Based on the parameters set above, the calculation process is as follows:

[0122] First, calculate the impact of resource allocation, i.e.

[0123]

[0124] Then calculate

[0125]

[0126] Next, we calculate the denominator.

[0127]

[0128] Final calculation

[0129]

[0130] The result indicates that the path segment priority adjustment value is 4.4, meaning that this path segment is 4.4 times more important than other path segments when adjusting priorities. This value was derived by comprehensively considering resource scores, resource importance offsets, and historical data variance, providing a quantitative basis for resource priority adjustment and ensuring more scientific and reasonable resource allocation.

[0131] The specific steps of S5 are as follows:

[0132] S501: Based on the path segment coordinates, offset values ​​and direction parameters in the virtual-real synchronous path correction instruction set, call the twin platform modeling resources, extract the path segment spatial node data and processing position information, load the path graphic system structure according to the node arrangement and geometric constraints, and obtain the three-dimensional path structure primitive set.

[0133] During execution, data containing path segment coordinates, offset values, and direction parameters is first extracted from the virtual-real synchronized path correction instruction set. This data typically describes the spatial positioning information of each segment in the processing path, such as the coordinates of the path segment's start and end points, the direction of offset, and the specific value of the offset. After extraction, the path segment coordinates undergo weighted offset processing, and the original path is corrected according to the offset values ​​and direction parameters. The corrected path is then transmitted to the digital twin platform in data structure form. The modeling resource module is invoked within the twin platform to perform node decomposition on the corrected path segment information, converting the path segments into a sequence of nodes with topological relationships in space. These nodes... In three-dimensional space, there are clear coordinates, sequential numbering, and geometric connection relationships. Then, by reading the node arrangement order and the geometric constraints between them, a structured path graphics system is constructed. This graphics system not only preserves the geometric attributes of the path, but also includes the logical connection relationships between path segments. For example, a path segment consists of 12 spatial nodes, with a constant spacing between each pair of nodes and consistent direction. The system uses this as the basic unit to construct a complete three-dimensional path structure primitive set. This primitive set includes basic parameters such as spatial type, segment length, node sequence, inter-segment angle, and spatial attitude. Each primitive represents the spatial representation of a processing path, providing a structural basis for subsequent trajectory comparison and deviation analysis.

[0134] S502: Based on the three-dimensional path structure primitive set, obtain the node trajectory values ​​in the twin model and the real-time trajectory data fed back by the device, match the two types of trajectory values ​​by node index, identify nodes whose offset amplitude exceeds the dynamic error warning value, and obtain the path offset numerical feature group by combining direction and frequency information.

[0135] Based on the constructed 3D path structure primitive set, the trajectory values ​​of key nodes in each primitive are extracted. These trajectory values ​​are theoretical path information provided by the static model. Then, real-time trajectory data collected by the equipment during actual processing is obtained. The latter is periodically recorded at processing nodes by sensors and control systems, forming a continuous dynamic path set. The two types of trajectory values ​​are matched by node index. That is, the node numbered 5 in the model corresponds to the actual trajectory point numbered 5 in the feedback trajectory of the processing equipment. After the matching is completed, the spatial offset amplitude between the theoretical value and the actual value is compared. If the offset amplitude exceeds the pre-set dynamic error warning value, the node is marked as an offset node. The offset node not only includes coordinate differences, but also needs to combine direction information to determine the directional attribute of the offset trend. For example, if the offset is concentrated along the positive X-axis, it means that the offset direction is positive X. At the same time, the frequency of this type of offset in multiple periods is recorded to form a path offset feature group with parameters such as number, direction, amplitude, and frequency. Each item in the feature group can be traced back to the specific processing trajectory segment and node number, providing a data foundation for primitive offset layer superposition and path difference trend analysis.

[0136] S503: Call the path offset numerical feature group and path segment number, overlay the three-dimensional path structure primitive morphology data, construct the offset annotation layer in the twin interface, record the difference between the predicted and actual values ​​according to the processing time sequence of the path segment, and generate the path virtual-real consistency deviation trend value.

[0137] Using path offset feature groups and path segment numbers, the path offset values ​​are mapped one by one to the spatial positions of the original graphic elements. Following the original path segment numbering structure, the position coordinates of all nodes in each graphic element segment are adjusted. During the adjustment process, the specific direction of the offset needs to be determined based on the spatial direction of the offset. For example, if the offset value is 1.2 mm in the positive Y-axis direction, then the Y-coordinate of each node in the graphic element needs to be increased by 1.2 mm. The offset graphic element positions form a new layer structure. This layer structure is graphically overlaid on the original path layer in the digital twin interface, forming an offset annotation layer. In the offset annotation layer, offset level ranges are divided according to the node offset magnitude, and colors are used for visual marking. For example, offset... Aspects within 1 mm are marked in yellow, those between 1 and 2 mm in orange, and those exceeding 2 mm in red. The color changes represent the spatial distribution of the overall path offset. The actual trajectory data during processing is then compared with the original predicted values. Differences within each processing cycle are recorded to form a serialized deviation data set. The path difference degree of each node in different processing cycles is recorded sequentially. The average deviation trend value of each node in multiple cycles is accumulated and calculated. If the trend value gradually increases within the statistical period and exceeds the dynamic deviation increment threshold set by the historical offset average, it is determined to be a significant increase in path consistency deviation, indicating that there is a continuous spatial drift risk in the long-term operation of the processing path.

[0138] Please see Figure 2 A digital twin-based machining factory management system includes:

[0139] The temperature monitoring module collects global temperature sensor data and tool trajectory coordinates, interpolates to generate a temperature distribution surface, calculates the absolute difference between the temperature change rate at the trajectory coordinate points and the surface change rate, and if the difference exceeds the temperature fluctuation threshold, extracts the coordinates of abnormal temperature points and generates a dynamic temperature field influence range boundary map.

[0140] The path early warning module extracts the vertex coordinates of the polygon in the boundary map of the influence range of the dynamic temperature field, traverses the coordinates of the path nodes, calculates the Euclidean distance between the nodes and the boundary, compares it with the thermal distortion deviation threshold, filters nodes with excessive distance, marks the starting point coordinates and three-axis offset of the corresponding path segment, and generates a thermally induced path deviation early warning mark group.

[0141] The resource allocation module calls the offset of the thermal path offset warning flag group, detects and calculates the thread occupancy rate of the resource pool, filters idle threads and sorts them in descending order of offset, binds the largest offset path segment to the thread with the largest free memory, establishes a mapping table between threads and path segments, and generates a dynamic allocation list of virtual entity resources.

[0142] The parameter correction module extracts the spindle load current and feed rate pulses from the virtual entity resource dynamic allocation list, converts the load current into a torque coefficient, converts the pulses into displacement increments, adjusts the torque and displacement parameters of the simulation model path segment according to priority, and generates a virtual-real synchronous path correction instruction set.

[0143] The verification feedback module loads the virtual-real synchronous path correction instruction set into the 3D model, extracts the coordinate positioning deviation value, speed following deviation value, and trajectory curvature deviation value of the correction path segment, compares the digital twin prediction data with the actual execution data of the physical equipment, integrates the distribution range, mean, and standard deviation of the deviation value, and generates a dynamic path virtual-real consistency verification report.

[0144] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for managing a machining factory based on digital twins, characterized in that, Includes the following steps: S1: Through the workshop-wide temperature sensor network, real-time temperature data of the processing area is collected, the tool movement trajectory is compared with the temperature distribution change trend, the range of abnormal temperature fluctuations within the trajectory coverage area is identified, and a dynamic temperature field influence range boundary map is generated. S2: Based on the boundary map of the influence range of the dynamic temperature field, extract the set of coordinates of the processing path nodes within the boundary line, combine it with the preset equipment thermal deformation tolerance threshold, mark the starting point and directional offset of the path segment that exceeds the tolerance range, and generate a thermal path offset warning mark group. S3: Invoke the thermal path offset warning flag group, scan the thread load status in the digital twin computing resource pool, sort the idle computing threads according to the severity of the offset of the warning path segment, establish the priority binding relationship between thread resources and path segments, and generate a dynamic allocation list of virtual entity resources. S4: Based on the aforementioned virtual and real computing resource dynamic allocation list, extract the spindle load data and feed rate data fed back in real time by the physical processing equipment, correct the simulation iteration number and motion trajectory control parameters of the corresponding path segment in priority order, and generate a virtual and real synchronous path correction instruction set.

2. The machining factory management method based on digital twins according to claim 1, characterized in that, The dynamic temperature field influence range boundary map includes temperature anomaly distribution areas, trajectory-related spatial ranges, and boundary change trend graphs. The thermally induced path deviation early warning marker group includes path segment start point markers, direction deviation markers, and out-of-tolerance coordinate sets. The virtual entity resource dynamic allocation list includes thread resource numbers, deviation severity levels, and resource path segment binding relationships. The virtual-real synchronous path correction instruction set includes iteration number correction values, motion trajectory control parameter adjustment values, and path segment matching numbers.

3. The machining factory management method based on digital twins according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Acquire the temperature values ​​of the collection points in the temperature sensor network covering the entire workshop, establish a mapping relationship between the temperature data and the processing area grid based on the sensor spatial coordinates, and generate the temperature change trend value of the processing grid by combining the temperature time series change of the grid area. S102: Based on the temperature change trend value of the machining grid, call the path segment coordinate sequence in the tool motion trajectory and perform spatial matching with the positional relationship of the machining grid, extract the temperature change rate in the area covered by the trajectory path, perform difference analysis on the temperature sequence of the matched area, and obtain the abnormal temperature interval value of the trajectory path. S103: Based on the abnormal temperature range values ​​of the trajectory path, extract the corresponding interval boundary coordinate points, connect the boundary points in spatial order to form a polyline structure, and construct a boundary closed path by combining the positional relationship of adjacent path segments to generate a boundary map of the dynamic temperature field influence range.

4. The machining factory management method based on digital twins according to claim 3, characterized in that, The specific steps of S2 are as follows: S201: Based on the boundary map of the influence range of the dynamic temperature field, extract the set of coordinates of the processing path nodes within the boundary line, combine the time information and temperature data of each coordinate point, perform correlation analysis on the temperature change trend and spatial location between nodes, and generate a temperature influence correlation coefficient value. S202: Based on the temperature influence correlation coefficient value, cluster the node coordinates, identify the temperature change peak and position offset peak in the cluster points, and combine the thermal variation deviation threshold set in the twin model to filter out path segments that exceed the threshold and generate path segment offset marker coordinate values. S203: For the path segment direction change corresponding to each set of coordinates in the path segment offset marker coordinate values, extract the angle difference between the spatial direction of the coordinate point and the original movement direction of the path, establish a reference table of the relationship between the starting position and the direction change of the path segment, and generate a thermal path offset warning marker group.

5. The machining factory management method based on digital twins according to claim 4, characterized in that, The specific steps for S3 are as follows: S301: Based on the thermal path offset warning marker group, call the path segment offset severity and marker number, extract the thread load data, running status and idle cycle length, filter thread resources that meet the conditions according to the running status and load threshold, and generate an idle thread resource list. S302: Call the idle thread resource list and thermal path offset warning flag group, prioritize the path segment offset severity, sort according to the thread idle cycle, establish the correspondence between path segments and thread resources, and generate a path segment resource priority binding relationship table; S303: Based on the path segment resource priority binding relationship table, integrate the path segment direction offset, space number, priority level and thread resource label to construct a resource allocation structure and generate a dynamic allocation list of virtual entity resources.

6. The machining factory management method based on digital twins according to claim 5, characterized in that, The specific calculation formulas for the load data, running status, and idle cycle length of the extracted thread are as follows: Among them, L′ i ω represents the overall load fluctuation value of the i-th thread. ij f represents the weight value of the load of the i-th thread in the overall thread pool during the j-th monitoring. ij τ represents the instantaneous load value of the i-th thread in the j-th monitoring. ij u represents the idle period length of the i-th thread in the j-th monitoring. ij d represents the task running status value of the i-th thread in the j-th monitoring. ij y represents the task processing time of the i-th thread in the j-th monitoring, and y represents the total number of data records within the monitoring period. This represents the average load value of the i-th thread during the monitoring period.

7. The machining factory management method based on digital twins according to claim 5, characterized in that, The specific steps of S4 are as follows: S401: Obtain the thread resource number, path segment number and priority order value from the virtual and real computing resource dynamic allocation list, extract the spindle load time series data and feed rate synchronization data corresponding to the path segment, complete the pairing and sorting of the two types of data, and generate load and feed paired sequence value group. S402: Based on the spindle load value and feed rate value of the path segment in the load and feed pairing sequence value group, and combined with the motion trajectory control parameters and thread occupancy status of the path segment, extract the parameter offset information and obtain the path segment control parameter offset coefficient group. S403: Based on the offset value and path segment number in the path segment control parameter offset coefficient group, and combined with the priority order in the resource allocation list, the simulation iteration number and motion trajectory control parameters of the corresponding path segments are corrected in sequence to generate a virtual-real synchronous path correction instruction set.

8. The machining factory management method based on digital twins according to claim 7, characterized in that, The specific calculation formula for the priority order in the resource allocation list is as follows: in, W represents the priority adjustment value required for the i-th path segment. i R represents the initial weight coefficient of the i-th path segment in the path set. j Δ represents the available score of the j-th resource in the resource allocation list. ij This represents the offset value indicating the importance of resource j to path segment i. This represents the sample variance of the number of iterations for path segment i in the historical simulation. μ represents the weighted average of all resource scores assigned to path segment i. i This represents the average resource score of path segment i among similar paths.

9. The machining factory management method based on digital twins according to claim 1, characterized in that, The method further includes: S5: Based on the virtual-real synchronization path correction instruction set, load the three-dimensional visualization model of the correction path in the twin interface, record the deviation value between the twin model prediction data and the actual execution data of the device, and generate a dynamic path virtual-real consistency verification report. The dynamic path virtual-real consistency verification report includes a three-dimensional correction model, prediction and execution deviation records, and path execution offset trend chart. The specific steps of S5 are as follows: S501: Based on the path segment coordinates, offset values ​​and direction parameters in the virtual-real synchronous path correction instruction set, call the twin platform modeling resources, extract the path segment spatial node data and processing position information, load the path graphic system structure according to the node arrangement and geometric constraints, and obtain the three-dimensional path structure primitive set. S502: Based on the three-dimensional path structure primitive set, obtain the node trajectory values ​​in the twin model and the real-time trajectory data fed back by the device, match the two types of trajectory values ​​according to the node index, identify the nodes whose offset amplitude exceeds the dynamic error warning value, and obtain the path offset numerical feature group by combining direction and frequency information. S503: Call the path offset numerical feature group and path segment number, overlay the three-dimensional path structure primitive morphology data, construct the offset annotation layer in the twin interface, record the difference between the predicted and actual values ​​according to the processing time sequence of the path segment, and generate the path virtual-real consistency deviation trend value.

10. A machining factory management system based on digital twins, characterized in that, The machining plant management method based on digital twins according to any one of claims 1-9, wherein the system comprises: The temperature monitoring module collects global temperature sensor data and tool trajectory coordinates, interpolates to generate a temperature distribution surface, calculates the absolute difference between the temperature change rate at the trajectory coordinate points and the surface change rate, and if the difference exceeds the temperature fluctuation threshold, extracts the coordinates of abnormal temperature points and generates a dynamic temperature field influence range boundary map. The path warning module extracts the vertex coordinates of the polygon in the boundary map of the influence range of the dynamic temperature field, traverses the coordinates of the path nodes, calculates the Euclidean distance between the nodes and the boundary, compares it with the thermal distortion deviation threshold, filters nodes with excessive distance, marks the starting point coordinates and three-axis offset of the corresponding path segment, and generates a thermal path deviation warning mark group. The resource allocation module calls the offset of the thermal path offset warning marker group, detects and calculates the resource pool thread occupancy rate, filters idle threads and sorts them in descending order of offset, binds the largest offset path segment to the thread with the largest idle memory, establishes a thread-path segment mapping table, and generates a virtual entity resource dynamic allocation list. The parameter correction module extracts the spindle load current and feed rate pulses according to the virtual entity resource dynamic allocation list, converts the load current into a torque coefficient, converts the pulses into displacement increments, adjusts the torque and displacement parameters of the simulation model path segment according to priority, and generates a virtual-real synchronous path correction instruction set. The verification feedback module loads the virtual-real synchronous path correction instruction set into the 3D model, extracts the coordinate positioning deviation value, speed following deviation value, and trajectory curvature deviation value of the correction path segment, compares the digital twin prediction data with the actual execution data of the physical device, integrates the distribution range, mean, and standard deviation of the deviation value, and generates a dynamic path virtual-real consistency verification report.

Images