Cooperative control method and system for multi-process standard part integrated machining

Abstract

The invention discloses a cooperative control method and system for multi-process standard part integrated machining, and relates to the technical field related to cooperative control, and the method comprises the steps: building a process machining state characterization model through collecting multi-source data such as geometric offset, machining response and clamping stability of each process; extracting a state component and an irreversible change component which can be transmitted across processes, and generating a state inheritance vector; analyzing the influence of the preorder process on the adjustable space and the control degree of freedom of the subsequent process, and forming a cooperative influence mapping relation; and combined optimization of a processing parameter configuration strategy is carried out in combination with the overall processing consistency constraint, and cooperative control management is executed. The technical problems that in the prior art, in the multi-process machining process, state information is isolated, the inter-process coupling influence is difficult to quantify, and collaborative decision and control based on the global state inheritance relation are lacked are solved, and the technical effects that accurate evaluation of cross-process collaborative influence is achieved, and the overall machining consistency, precision and efficiency of standard parts are improved are achieved.

Description

Technical Field

[0001] This invention relates to the field of collaborative control technology, specifically to a collaborative control method and system for integrated processing of standard parts across multiple processes. Background Technology

[0002] Multi-process integrated machining of standard parts is the key to improving the overall performance and production efficiency of products. In multi-process continuous machining, the machining status of each process, such as geometric accuracy, stress response, and clamping stability, not only directly affects the machining quality of the process itself, but also transmits its influence to subsequent processes through changes in the physical state of the workpiece, forming a cross-process "state flow" that affects the stability and output accuracy of the subsequent process system. However, existing machining control often treats each process as an independent link, adopting "island-style" monitoring and parameter adjustment. It lacks systematic modeling and proactive coordination of the state transmission and coupling effects between processes. On the one hand, when a standard part undergoes a series of processes such as turning, milling, grinding, and measuring, its state attributes such as geometric shape, internal stress, and surface integrity will change continuously with the machining process. Offline detection after the completion of the process is difficult to trace the source of machining deviations and provide feedforward compensation. On the other hand, the sensors of the intelligent machining unit lack cross-process correlation analysis and information integration, which restricts collaborative decision-making based on global state insight. In addition, the machining results such as uneven allowance distribution, datum surface error, and changes in surface characteristics in the preceding processes will directly constrain the adjustment space and control freedom of the subsequent processes, resulting in a decrease in overall machining consistency and pass rate.

[0003] Therefore, current technologies face technical challenges such as isolated state information during multi-process manufacturing, difficulty in quantifying the coupling effects between processes, and a lack of collaborative decision-making and control based on global state inheritance relationships. Summary of the Invention

[0004] This application provides a collaborative control method and system for integrated processing of standard parts across multiple processes. This solves the technical problems in the prior art, such as isolated state information, difficulty in quantifying the coupling effects between processes, and lack of collaborative decision-making and control based on global state inheritance relationships during multi-process processing. It achieves the technical effect of accurately assessing the collaborative effects across processes and improving the overall consistency, accuracy, and efficiency of standard part processing.

[0005] This application provides a collaborative control method for integrated machining of standard parts across multiple processes. The method includes: collecting multi-source machining state data representing the geometric offset, machining response, and clamping stability of the standard part under each process during at least two machining processes; inputting the multi-source machining state data into a machining state representation model for each process of the component; performing cross-process correlation analysis using the machining state representation model to extract state components that can be transmitted between adjacent processes and components that undergo irreversible changes during supply and demand switching, generating a state inheritance vector representing the cross-process inheritance relationship of the standard part's machining state; evaluating the degree of influence of the machining results of the preceding process on the adjustable space and control degrees of freedom of the subsequent process based on the state inheritance vector, forming a collaborative influence mapping relationship between multiple processes; performing joint optimization of machining parameter configuration strategies according to the collaborative influence mapping relationship and overall machining consistency constraints; and performing collaborative control management based on the joint optimization results.

[0006] In a possible implementation, cross-process correlation analysis is performed using the machining state characterization model, including: the machining state characterization model includes a multi-source feature encoding layer, a unified state tensor construction layer, and a cross-process inheritance decomposition operator, wherein: the multi-source feature encoding layer decomposes the multi-source machining state data to obtain aggregated measurement data, cutting power or torque data, vibration data, clamping force data, and temperature data, and constructs a decomposed feature set; frequency domain peak energy, spectral entropy, temperature rise slope, clamping force attenuation rate, and aggregated deviation principal direction coefficient are extracted from the decomposed feature set to construct a modal feature vector; the unified state tensor construction layer is activated to integrate the modal feature vectors with the same machining reference. Aligning the coordinates and aggregating them according to a sliding time window generates a process processing state tensor; activating the cross-process inheritance decomposition operator, and constructing a corresponding cross-process state transition operator based on the process difference type and state baseline inheritance relationship between adjacent processing processes; using the cross-process state transition operator to predict and map the processing state tensor of the preceding process, constructing an inheritable predictable state tensor of the processing state of the preceding process in the subsequent process; performing difference calculation on the actual process processing state tensor obtained in the subsequent process and the inheritable predictable state tensor to obtain a state residual tensor representing the irreversible changes that occur during process switching; constructing a state inheritance vector based on the state residual tensor.

[0007] In a possible implementation, constructing a state inheritance vector based on the state residual tensor includes: using the energy proportion of the state residual tensor in each modal feature dimension and the stability index within the sliding time window to effectively screen the corresponding state components in the inheritable predictable state tensor, and determining the effective inherited state components that can be stably transferred across processes; using the sensitivity of each state component in the state residual tensor to the adjustment results of subsequent processing parameters to configure transfer weights for the effective inherited state components; and combining the effective inherited state components, transfer weights, and residual constraint components that characterize the intensity of irreversible changes to construct a state inheritance vector that characterizes the degree of inheritance and constraint relationship of the standard part processing state between adjacent processes.

[0008] In a possible implementation, the cross-process inheritance decomposition operator includes: a process difference determination unit, used to generate a state identifier matrix to identify the inheritance relationship of each state dimension before and after process switching based on the process type differences and clamping reference inheritance relationship of adjacent processing processes; a state mapping unit, used to perform reference alignment and mapping operations on each state dimension in the processing state tensor of the preceding process according to the state identifier matrix to form inheritable state components; a non-inheritable separation unit, used to separate state dimensions that do not meet the inheritance conditions from the processing state tensor of the preceding process to form state residual components; and an output combination unit, used to combine the inheritable state components with the state residual components to form the output tensor of the cross-process inheritance decomposition operator.

[0009] In a possible implementation, joint optimization of the processing parameter configuration strategy based on the synergistic influence mapping relationship and the overall processing consistency constraint includes: dividing the processing parameter set into multi-dimensional parameter subspaces based on the adjustable space and control degrees of freedom of each processing step, with each parameter subspace corresponding to different process adjustment degrees of freedom and processing sensitivity; mapping the synergistic influence mapping relationship to the multi-dimensional parameter subspaces to apply synergistic constraints, limiting the parameter adjustment range and allowable variation amplitude; applying constraints to the candidate parameter groups in each parameter subspace based on the overall processing consistency constraint, performing joint optimization calculations, and outputting the joint optimization results.

[0010] In a possible implementation, after imposing constraints on candidate parameter groups in each parameter subspace according to the overall processing consistency constraint, joint optimization calculation is performed, including: performing local optimization calculation in the neighborhood of each candidate parameter group to obtain the local optimal parameter combination and the corresponding performance index; determining whether to trigger a jump operation based on the performance gradient of the performance index, the sensitivity of the state inheritance vector, and the cooperative influence mapping constraint, wherein the jump operation is used to escape the local optimum; if a jump operation is triggered, the jump direction and magnitude are configured based on the minimization of processing error, the maximization of processing efficiency, and the satisfaction of multi-process consistency constraints, and a cross-subspace jump operation is performed to perform joint optimization.

[0011] In a possible implementation, collaborative control management is performed based on the joint optimization result, including: obtaining a trust level identifier of the joint optimization result, configuring an adjustable margin space based on the trust level identifier, and using the margin space to perform margin self-optimization update of the collaborative control process.

[0012] This application also provides a collaborative control system for integrated machining of multi-process standard parts. The system includes: a machining state data acquisition module, used to acquire multi-source machining state data representing the geometric offset, machining response, and clamping stability of the standard part under each process during at least two machining processes, and input the multi-source machining state data into the machining state representation model of each process of the component; a cross-process correlation analysis module, used to perform cross-process correlation analysis using the machining state representation model, extracting state components that can be transmitted between adjacent processes and components that undergo irreversible changes during supply and demand switching, and generating a state inheritance vector representing the cross-process inheritance relationship of the standard part's machining state; an influence degree assessment module, used to assess the influence degree of the machining results of the preceding process on the adjustable space and control freedom of the subsequent process based on the state inheritance vector, forming a collaborative influence mapping relationship between multiple processes; and a collaborative control management module, used to perform joint optimization of machining parameter configuration strategies according to the collaborative influence mapping relationship and overall machining consistency constraints, and perform collaborative control management according to the joint optimization results.

[0013] This application proposes a collaborative control method and system for integrated machining of multi-process standard parts. This method involves collecting multi-source data, including geometric offsets, machining responses, and clamping stability, from each process to establish a process machining state characterization model. It extracts state components that can be transferred across processes and irreversible change components to generate a state inheritance vector. The method analyzes the impact of preceding processes on the adjustable space and control degrees of freedom of subsequent processes, forming a collaborative influence mapping relationship. Finally, it combines overall machining consistency constraints to jointly optimize machining parameter configuration strategies and implement collaborative control management. This addresses the technical problems in existing technologies, such as isolated state information, difficulty in quantifying inter-process coupling effects, and a lack of collaborative decision-making and control based on global state inheritance relationships during multi-process machining. It achieves accurate assessment of cross-process collaborative influences and improves the overall machining consistency, accuracy, and efficiency of standard parts. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0015] Figure 1 This is a schematic flowchart of a collaborative control method for integrated processing of standard parts across multiple processes, provided in an embodiment of this application.

[0016] Figure 2 This is a schematic diagram of a collaborative control system for integrated processing of standard parts across multiple processes, provided in an embodiment of this application.

[0017] Figure labeling: 10 processing status data acquisition module, 20 cross-process correlation analysis module, 30 impact assessment module, 40 collaborative control management module. Detailed Implementation

[0018] To further illustrate the technical means and effects adopted by the present invention in order to achieve the intended purpose, the following detailed description is provided in conjunction with the accompanying drawings and preferred embodiments, based on the specific implementation methods, structures, features and effects of the present invention.

[0019] This application provides a collaborative control method for integrated machining of standard parts across multiple processes, such as... Figure 1 As shown, the method includes:

[0020] Step S100: During the process of the standard part undergoing at least two processing steps, multi-source processing state data representing the geometric offset, processing response and clamping stability of the standard part under each process are collected, and the multi-source processing state data are input into the processing state characterization model of the component under each process.

[0021] Preferably, as the standard part undergoes each independent machining process, an integrated sensing system synchronously captures three types of data: multi-source machining state data characterizing the geometric offset, machining response, and clamping stability of the standard part under each process. Specifically, online measurement devices such as in-machine probes and laser interferometers are used to obtain three-dimensional deviation data between the actual size, shape, and position of the workpiece and the theoretical model to characterize the geometric offset of the standard part; dynamic parameters such as cutting force / torque spectrum, vibration spectrum, acoustic emission signal, and spindle power transient curve are recorded using force sensors, accelerometers, acoustic emission sensors, and power monitoring units to characterize the machining response of the standard part; and clamping force decay curve and micro-displacement of the workpiece-fixture contact interface are monitored using strain gauge clamping force sensors and micro-displacement sensors to characterize the clamping stability of the standard part. Then, data preprocessing operations such as timestamp alignment, coordinate system unification, noise filtering, and outlier removal are performed on the multi-source machining state data. The preprocessed data is then reorganized according to the process feature dimension to form a standardized input data matrix for the process, which is then input into the machining state characterization model of each process of the component. Through feature extraction, a quantitative description of the machining state of the process is output.

[0022] Step S200: Perform cross-process correlation analysis using the processing state characterization model, extract state components that can be transmitted between adjacent processes and components that undergo irreversible changes during supply and demand switching, and generate a state inheritance vector that characterizes the cross-process inheritance relationship of the standard part's processing state.

[0023] Step S200 further includes the following: the machining state characterization model includes a multi-source feature encoding layer, a unified state tensor construction layer, and a cross-process inheritance decomposition operator, wherein: the multi-source feature encoding layer decomposes the multi-source machining state data to obtain aggregated measurement data, cutting power or torque data, vibration data, clamping force data, and temperature data, and constructs a decomposed feature set; the multi-source feature set is used to extract frequency domain peak energy, spectral entropy, temperature rise slope, clamping force attenuation rate, and aggregated deviation principal direction coefficient to construct a modal feature vector; the unified state tensor construction layer is activated to align the modal feature vectors in the same machining reference coordinate system and according to the sliding time... Inter-process window aggregation generates a process processing state tensor; the cross-process inheritance decomposition operator is activated, and a corresponding cross-process state transition operator is constructed based on the process difference type and state baseline inheritance relationship between adjacent processing processes; the cross-process state transition operator is used to predict and map the processing state tensor of the preceding process, and an inheritable predictable state tensor of the processing state of the preceding process under the subsequent process is constructed; the difference calculation is performed between the actual process processing state tensor obtained by the subsequent process and the inheritable predictable state tensor to obtain the state residual tensor representing the irreversible change during the process switching process; a state inheritance vector is constructed based on the state residual tensor.

[0024] Preferably, the processing state characterization model, through a multi-source feature encoding layer and a unified state tensor construction layer, transforms the real-time collected raw data from multiple sensors into a structured process processing state tensor containing multi-dimensional modal features and time-series information. When it is necessary to analyze the correlation between processes, the cross-process inheritance decomposition operator is invoked to perform decoupling analysis on the process processing state tensor. This allows for the generation of state inheritance vectors that can be transmitted between adjacent processes and components that undergo irreversible changes during supply and demand switching. This generates a state inheritance vector that characterizes the cross-process inheritance relationship of the standard part's processing state, thereby representing the inheritance relationship between adjacent processing processes and the transmittable state features, and quantifying the relative importance or influence of each transmittable state component.

[0025] Preferably, the machining state characterization model includes a multi-source feature encoding layer, a unified state tensor construction layer, and a cross-process inheritance decomposition operator. Specifically, the multi-source feature encoding layer is used to decompose multi-source machining state data, that is, to decompose the synchronously acquired raw mixed data stream according to its physical source and meaning, and obtain a set of measurement data, cutting power or torque data, vibration data, clamping force data, and temperature data. For example, the voltage signal from the probe is parsed into dimensional measurement data, the spindle current sensor signal is parsed into cutting power data, and the accelerometer signal is parsed into vibration data. The basic feature calculation is performed on the decomposed data to generate multiple low-level features. For example, the effective value (RMS) of a vibration time-domain signal is calculated, the instantaneous value of the clamping force data is calculated, and the decomposed feature set is obtained.

[0026] Preferably, advanced modal feature extraction is performed on the decomposed feature set using signal processing and statistical algorithms. This includes performing a fast Fourier transform on the vibration or acoustic emission signal to calculate the peak energy of the main frequency components, reflecting the intensity of the dominant mode shape or tool wear characteristic frequency; calculating the entropy value of the signal power spectrum to measure the disorder of the signal spectrum and to assess the stability of the machining process; performing linear fitting on the temperature sensor data sequence, with the slope representing the rate of heat accumulation in the machining area; performing exponential or linear fitting on the clamping force data sequence to calculate its attenuation coefficient and quantitatively assess the relaxation speed of the fixture; and performing principal component analysis on the geometric deviation data from multiple measurement points, with the direction and variance contribution rate of the first principal component representing the most important error direction and magnitude. Then, the calculated feature scalar values ​​are arranged in order to form a modal feature vector, which highly summarizes the current instantaneous machining state.

[0027] Preferably, a unified state tensor construction layer aligns modal feature vectors under the same machining reference coordinate system and aggregates them according to a sliding time window. This includes transforming the physical coordinates of features calculated by different probes and force sensors to the workpiece design reference or machine tool coordinate system to ensure spatial consistency. Then, according to a preset time window, such as the most recent 10 seconds, multiple modal feature vectors generated continuously within the window are stacked to generate a process machining state tensor. This may be a three-dimensional data array containing feature type dimension, time step dimension, and possible spatial / channel dimension, used to characterize the multi-faceted complete state evolution information of the process in the recent period. Using a cross-process inheritance decomposition operator, based on the process difference type between adjacent machining processes, such as from turning to grinding, and combined with the state reference inheritance relationship, a mathematical transformation is determined, including coordinate transformation matrix and error propagation and stress relaxation model. A corresponding cross-process state transition operator is constructed. Then, the machining state tensor of the preceding process is used as input, and the cross-process state transition operator is applied for prediction mapping to obtain the inheritable predictable state tensor of the preceding process's machining state under the subsequent process.

[0028] Preferably, tensor difference calculation is performed on the actual process processing state tensor obtained in subsequent processes and the inheritable predicted state tensor to obtain the state residual tensor representing irreversible changes during process switching. That is, after process switching, the difference between the actual measured state and the state predicted solely by inheritance from the previous state includes new states that did not exist in the previous process due to material removal, thermal / mechanical effects introduced by new processes, new clamping deformations, etc., i.e., irreversible changes. Finally, the energy in different feature dimensions is extracted from the state residual tensor and its stability is evaluated to construct a structured state inheritance vector, encoding both the inheritable and irreversible parts.

[0029] Furthermore, step S200 also includes the cross-process inheritance decomposition operator comprising: a process difference determination unit, used to generate a state identifier matrix to identify the inheritance relationship of each state dimension before and after process switching based on the process type differences and clamping reference inheritance relationship of adjacent processing processes; a state mapping unit, used to perform reference alignment and mapping operations on each state dimension in the processing state tensor of the preceding process according to the state identifier matrix to form inheritable state components; a non-inheritable separation unit, used to peel off state dimensions that do not meet the inheritance conditions from the processing state tensor of the preceding process to form state residual components; and an output combination unit, used to combine the inheritable state components with the state residual components to form the output tensor of the cross-process inheritance decomposition operator.

[0030] Preferably, the cross-process inheritance decomposition operator includes a process difference determination unit, a state mapping unit, a non-inheritable separation unit, and an output combination unit. This unit decomposes the machining state tensor of the input preceding process into inheritable state components and state residual components. The process difference determination unit establishes the mapping relationship for state transfer between processes. Specifically, the input includes process type descriptions of two adjacent processes, such as turning → grinding, milling → drilling, and clamping scheme information for the two processes, particularly the inheritance relationship of positioning datums, such as whether the same set of datum surfaces is used or whether the datum is converted. Based on a pre-set process knowledge base, it determines the machining state tensor under given process and clamping changes. Each state dimension in the quantity has the physical possibility and logical consistency to be passed from the previous process to the subsequent process. The process knowledge base includes the following: if the reference surface is completely consistent, the geometric error dimension related to the position may be inherited; if the subsequent process completely removes the machined surface of the previous process, the residual stress state dimension related to that surface cannot be inherited; the vibration mode of the tool spindle is completely changed due to different machine tools, so the vibration characteristic dimension cannot be inherited; and a state label matrix is ​​generated to identify the inheritance relationship of each state dimension before and after the process switch. The rows of this matrix correspond to the state dimensions, and the value 1 or 0 clearly indicates whether the dimension can be inherited or not during the process switch.

[0031] Preferably, the state mapping unit is used to perform coordinate and physical quantity transformations of the inheritable part. Specifically, it performs reference alignment and mapping operations on each state dimension in the processing state tensor of the preceding process according to the state identification matrix. That is, it extracts all state dimensions marked as inheritable from the processing state tensor to form a subset, and applies geometric coordinate transformation and physical quantity transfer model corresponding to the physical meaning of the dimension to each dimension. Then, it recombines all the dimension values ​​after mapping transformation into inheritable state components, representing the parts of the processing state tensor of the preceding process that can and have been mathematically transformed to the context of the subsequent process. The non-inheritable separation unit is used to separate state dimensions that do not meet the inheritance conditions from the processing state tensor of the preceding process, that is, state dimensions that cannot be directly transferred to the subsequent process, and finally forms the predicted state residual components. The output combination unit is used to combine the inheritable state components and the state residual components to form the output tensor of the cross-process inheritance decomposition operator, so as to facilitate prediction mapping and residual analysis, and ensure that the irreversible new state changes that actually occur at the process switching point are quantified.

[0032] Furthermore, step S200 also includes: using the energy proportion of the state residual tensor in each modal feature dimension and the stability index within the sliding time window to screen the state components corresponding to the inheritable predictable state tensor for validity, and determining the valid inherited state components that can be stably transferred across processes; using the sensitivity of each state component in the state residual tensor to the adjustment results of subsequent processing parameters to configure the transfer weights for the valid inherited state components; and combining the valid inherited state components, the transfer weights, and the residual constraint components that characterize the intensity of irreversible changes to construct a state inheritance vector that characterizes the degree of inheritance and constraint relationship of the standard part processing state between adjacent processes.

[0033] Preferably, for the state residual tensor, its energy in each modal feature dimension, such as peak energy, spectral entropy, and temperature rise slope in the frequency domain, is calculated. This includes calculating the sum of squares or L2 norm of the residual values ​​within all time windows of that dimension, and then calculating the percentage of energy in each feature dimension relative to the total energy. If the residual energy percentage in a certain dimension is very low, it indicates that the actual observed value in that dimension is very close to the predicted value, meaning that the state of the preceding process in that dimension is accurately predicted and inherited by the subsequent process, and the predicted component corresponding to that dimension is effective. The state residual tensor is then analyzed for its energy distribution over time in each feature dimension. The fluctuation of window changes, such as calculating variance or the strength of its trend over time, if the residual energy of a certain dimension is small and the fluctuation is gentle and the stability is low, it further indicates that its stable state transmission relationship is stable. Set a threshold, such as the residual energy ratio <5% and the stability index is less than a certain value. For each state component corresponding to each feature dimension in the inheritable predictable state tensor, check whether its corresponding residual meets the threshold condition. The components that pass the screening are marked as valid inherited state components that can be stably transmitted across processes, that is, the previous process state that can be reliably predicted and exists stably after process switching.

[0034] Preferably, the influence sensitivity refers to the degree of impact on the adjustment results of machining parameters or key control objectives of subsequent processes when each state component in the state residual tensor undergoes a unit change, such as final dimensional error, surface roughness, and tool life. This is gradient information that needs to be learned in advance through experiments, simulations, or historical data. Then, for each effective inherited state component, its influence sensitivity coefficient is calculated by analyzing the correlation between the historical data of the state residual tensor and the machining results of subsequent processes. Based on the calculated sensitivity coefficient, a transfer weight is assigned to each effective inherited state component, where the higher the sensitivity, the larger the assigned transfer weight, used to distinguish the relative importance of different inherited components in collaborative optimization. Finally, the effective inherited state components, transfer weights, and residual constraint components representing the intensity of irreversible changes are combined and spliced ​​according to a predefined data format to generate a structured state inheritance vector. This vector is used to represent the degree of inheritance and constraint relationship of the standard part machining state between adjacent processes, including the relative importance of each transferred state to the results of subsequent processes, and the intensity pattern of irreversible state changes generated at process switching points.

[0035] Step S300: Based on the state inheritance vector, evaluate the degree of influence of the processing results of the preceding process on the adjustable space and control freedom of the subsequent process, and form a collaborative influence mapping relationship between multiple processes.

[0036] Preferably, the state inheritance vector is used to evaluate the influence of the machining results of the preceding process on the adjustable space and control degrees of freedom of the subsequent process, and to establish a collaborative influence mapping relationship among the subsequent multiple processes. The adjustable space refers to the range of parameter adjustments in the subsequent process that can be used to compensate or correct the errors introduced by the preceding process while ensuring the machining quality of the current process. Examples include the geometric compensation space of toolpath offset, the process parameter window such as the feed rate of cutting speed, and the clamping adjustment margin of the fixture fine-tuning mechanism. The control degrees of freedom refer to the number and effectiveness of independent control variables that the subsequent process can independently adjust to cope with state changes.

[0037] Preferably, for each effective inherited state component in the state inheritance vector, a corresponding influence transfer function is determined to describe the unit change of the effective inherited component, which requires the consumption of the dimensional tolerance zone, form tolerance zone, or positional tolerance zone of the subsequent process. The amount consumed is the degree of influence of the effective inherited state component on the adjustable space. For each residual constraint component in the state inheritance vector, representing an irreversible new state, such as stiffness changes caused by the removal of new material or the generation of new surfaces, it may change the dynamic characteristics of the process system and limit the available range of control parameters. For example, after a process switch, the workpiece wall thickness becomes thinner and the rigidity decreases, causing subsequent processes to be more difficult to select the maximum allowable depth of cut. When cutting parameters such as feed are adjusted, the upper limit of the gain of the vibration suppression control loop is reduced, meaning that the degree of freedom for adjustment in this control dimension is compressed. By analyzing the residual constraint components to change the constraint equations or feasible region of the process system, the reduction ratio of the feasible range of each dimension of the control parameter vector is quantitatively calculated. Then, the evaluation results of the previous process on the subsequent process are organized into a structured synergistic influence mapping relationship, which may be a set of constraint equations or an influence coefficient matrix. This accurately describes the degree of compression and constraint of the feasible region of the subsequent process by the processing results of the previous process, and quantitatively demonstrates how the processing results of the previous process quantitatively reduce, limit, or change the adjustable space and control degree of freedom of the subsequent process.

[0038] Step S400: Perform joint optimization of the processing parameter configuration strategy based on the collaborative influence mapping relationship and the overall processing consistency constraint, and perform collaborative control management based on the joint optimization result.

[0039] Step S400 further includes dividing the processing parameter set into multi-dimensional parameter subspaces according to the adjustable space and control degrees of freedom of each processing step, with each parameter subspace corresponding to different process adjustment degrees of freedom and processing sensitivity; mapping the synergistic influence mapping relationship to the multi-dimensional parameter subspaces to apply synergistic constraints, limiting the parameter adjustment range and allowable variation amplitude; applying constraint conditions to the candidate parameter groups in each parameter subspace according to the overall processing consistency constraint, performing joint optimization calculations, and outputting the joint optimization results.

[0040] Preferably, the independent process parameters that can be adjusted to control the machining quality in each machining process are listed. For example, the cutting speed, feed rate, and tool compensation in turning, and the grinding wheel speed, workpiece speed, feed rate, and dressing interval in grinding. The adjustable parameters of all processes constitute a global parameter space. According to the physical affiliation and coupling relationship of the parameters, it is divided into multiple subspaces according to the process. For example, all parameters belonging to the turning process constitute the first subspace. The dimension of each parameter subspace corresponds to different process adjustment degrees of freedom, that is, the number of variables that can be adjusted independently. For each parameter dimension in each parameter subspace, its machining sensitivity is determined by evaluating through process models or experimental data, that is, the degree of influence of a small change in the parameter on the final machining quality or target. Then, the collaborative influence mapping relationship is transformed into specific mathematical constraints on the corresponding parameter subspace and mapped to the multi-dimensional parameter subspace to apply collaborative constraints. The allowed parameter value range of each parameter subspace is the range after dynamic correction by the collaborative influence mapping relationship, including limiting the parameter adjustment range and the allowable change amplitude. The allowed parameter value range is the collaborative constraint.

[0041] Preferably, a multi-objective comprehensive function is defined as the overall processing consistency constraint, such as minimizing the total processing error and maximizing the overall production efficiency while ensuring that the key quality characteristics of each process are within tolerance. Constraints are applied to candidate parameter groups in each parameter subspace according to the overall processing consistency constraint. A specific parameter value is selected from each parameter subspace to form a complete cross-process processing scheme and construct a candidate parameter group. Each candidate parameter group is checked to see if it satisfies the applied collaborative constraints and the overall processing consistency function. Then, an improved algorithm with "jump operation" is used to perform joint optimization calculation of the processing parameter configuration strategy. The search is carried out in the high-dimensional space formed by the Cartesian product of all parameter subspaces but strictly limited to the feasible region defined by the collaborative constraints. Numerous candidate parameter groups are evaluated to find the candidate parameter group that makes the overall objective function optimal, such as the comprehensive balance point with the best processing quality, highest efficiency, and lowest cost. Finally, the joint optimization result, i.e., the globally optimal cross-process processing parameter configuration scheme, is output.

[0042] Furthermore, step S400 also includes performing local optimization calculations in the neighborhood of each candidate parameter group to obtain the local optimal parameter combination and the corresponding performance index; determining whether to trigger a jump operation based on the performance gradient of the performance index, the sensitivity of the state inheritance vector, and the cooperative influence mapping constraint, wherein the jump operation is used to escape the local optimum; if a jump operation is triggered, the jump direction and magnitude are configured based on minimizing the processing error, maximizing the processing efficiency, and satisfying the consistency constraint of the multi-process, and a cross-subspace jump operation is performed to perform joint optimization.

[0043] Preferably, for each candidate parameter group, its neighborhood refers to the set of all parameter points in the multidimensional parameter space whose parameter values ​​differ from those of the candidate parameter group within a small step size ε. Gradient descent optimization is used to perform local optimization calculations within the neighborhood of the candidate parameter group to find the optimal solution within this small range, obtain the locally optimal parameter combination, i.e., the parameter point with the best performance in the current neighborhood, and evaluate the corresponding performance indicators, such as comprehensive processing error, efficiency score, etc.; then, the performance gradient of the performance indicator with respect to each parameter is calculated. If it is close to zero, it indicates that a local extremum has been reached; the sensitivity of each component in the state inheritance vector is obtained. Sensitivity weights are used to calculate whether highly sensitive state components are adequately optimized or compensated in the current local optimum parameter combination. If the error or cost corresponding to the highly sensitive inherited state components is high, it indicates that the quality of the current local optimum solution does not meet the core process requirements. The distance between the current local optimum parameter combination and the boundary of the cooperative influence mapping is evaluated. If the solution is squeezed into the corner of the boundary, it indicates that the adjustable space is extremely compressed, and it may be a forced, suboptimal local solution. If the performance gradient is close to zero and the highly sensitive components do not meet the standards or the local solution is on the boundary, it is determined that a jump operation needs to be triggered to escape the local optimum.

[0044] Preferably, if a jump operation is triggered, the jump direction and magnitude are configured based on minimizing machining error, maximizing machining efficiency, and satisfying the consistency constraints of multiple processes. Specifically, the main sources of current error are analyzed, and the jump direction tends to relax the parameter constraints causing the error or strengthen the parameter adjustments that compensate for the error. For example, if the roundness error is large, the jump direction may point to the region that allows for larger tool offset adjustments or changes in cutting parameters; the jump direction may point to the parameter region that can significantly shorten machining time, such as increasing the spindle speed or feed; the jump direction should be conducive to balancing the contradictions between processes, for example, jumping from a point that is extremely optimal for process A but strained for process B to the parameter region where both are at an acceptable level. The jump magnitude needs to be large enough to escape the current local optimum. The jump magnitude is calculated based on the adaptive covariance matrix, the probe step size of historical search information, or the geometric dimensions of the constraint space; then, according to the determined jump direction and magnitude, a cross-subspace jump operation is performed to generate a new candidate parameter set for joint optimization until the global convergence condition is met, ensuring that the generated cooperative control scheme achieves comprehensive optimality in terms of quality, efficiency, and robustness.

[0045] Furthermore, step S400 also includes obtaining a trust level identifier of the joint optimization result, configuring an adjustable margin space based on the trust level identifier, and using the margin space to perform margin self-optimization update of the collaborative control process.

[0046] Preferably, collaborative control management is implemented based on the joint optimization results, and the reliability and robustness of the joint optimization results are quantitatively evaluated. For example, it is determined whether the optimization process converges stably, the average deviation between the predicted values ​​of the cutting force model and the error propagation model and the historical actual values, the tension of collaborative constraints, and the uncertainty of the state inheritance vector. A trust level indicator is obtained, which may be classified as high / medium / low. The margin space is an adjustment range set around the theoretical optimal parameter value that allows for real-time fluctuations. Each dimension of the multi-dimensional interval corresponds to a machining parameter. Then, an adjustable margin space is configured according to the trust level indicator. A high trust level, such as >0.8, indicates that the optimization result is reliable, and the margin space is configured narrower, allowing the actual control parameters to make small adjustments near the optimal value. A low trust level, such as <0.4, indicates that the optimization result has high uncertainty, and the margin space is configured wider, leaving more room for adaptive adjustment in actual control. To avoid machining failures due to model inaccuracies, the margin space is used for self-optimization updates in the collaborative control process. This involves using the joint optimization result as the setpoint, allowing the adaptive controllers of each process to fine-tune within the preset margin space based on real-time feedback. This addresses issues such as instantaneous tool wear and material hardness fluctuations. During actual machining, the deviation of the actual adjustment amount of each underlying controller from the theoretical optimal value, as well as the real-time evaluation results of key machining states, are continuously monitored. If each controller maintains a good state with small adjustments within the margin space and the deviation is evenly distributed, the margin space is gradually reduced. If a controller continuously adjusts its parameters to the boundary of the margin space or the machining state exhibits unexpected drift, it is determined that the current disturbance is large or there is a deviation. Therefore, while ensuring machining quality, the minimum necessary margin is used as much as possible to improve the overall machining consistency, accuracy, and efficiency of the standard parts.

[0047] In the above text, refer to Figure 1 A collaborative control method for integrated machining of standard parts in multiple processes according to embodiments of the present invention is described in detail. Next, reference will be made to... Figure 2 A collaborative control system for integrated machining of standard parts in multiple processes, according to an embodiment of the present invention, is described.

[0048] The collaborative control system for integrated processing of standard parts according to embodiments of the present invention addresses the technical problems in existing technologies, such as isolated state information, difficulty in quantifying inter-process coupling effects, and lack of collaborative decision-making and control based on global state inheritance relationships during multi-process processing. It achieves the technical effect of accurately assessing cross-process collaborative effects and improving the overall consistency, accuracy, and efficiency of standard part processing. Figure 2 As shown, the collaborative control system for integrated processing of standard parts in multiple processes includes: a processing status data acquisition module 10, a cross-process correlation analysis module 20, an impact assessment module 30, and a collaborative control management module 40.

[0049] The processing status data acquisition module 10 is used to collect multi-source processing status data representing the geometric offset, processing response, and clamping stability of the standard part during at least two processing steps, and input the multi-source processing status data into the processing status representation model of each process of the component; the cross-process correlation analysis module 20 is used to perform cross-process correlation analysis using the processing status representation model, extract the state components that can be transmitted between adjacent processes and the components that undergo irreversible changes during supply and demand switching, and generate a state inheritance vector representing the cross-process inheritance relationship of the standard part's processing status; the influence degree assessment module 30 is used to assess the influence degree of the processing results of the preceding process on the adjustable space and control degree of freedom of the subsequent process based on the state inheritance vector, and form a collaborative influence mapping relationship between multiple processes; the collaborative control management module 40 is used to perform joint optimization of the processing parameter configuration strategy according to the collaborative influence mapping relationship and the overall processing consistency constraint, and perform collaborative control management according to the joint optimization result.

[0050] The specific configuration of the cross-process correlation analysis module 20 will be described in detail below. The cross-process correlation analysis module 20 further includes: the machining state characterization model includes a multi-source feature encoding layer, a unified state tensor construction layer, and a cross-process inheritance decomposition operator, wherein: the multi-source feature encoding layer decomposes the multi-source machining state data to obtain aggregated measurement data, cutting power or torque data, vibration data, clamping force data, and temperature data, constructing a decomposed feature set; extracting frequency domain peak energy, spectral entropy, temperature rise slope, clamping force attenuation rate, and aggregate deviation principal direction coefficient from the decomposed feature set to construct a modal feature vector; activating the unified state tensor construction layer, aligning the modal feature vectors in the same machining reference coordinate system and according to... Sliding time window aggregation generates process processing state tensors; the cross-process inheritance decomposition operator is activated, and corresponding cross-process state transition operators are constructed based on the process difference types and state baseline inheritance relationships between adjacent processing processes; the cross-process state transition operators are used to predict and map the processing state tensors of the preceding processes, constructing inheritable predictable state tensors of the processing states of the preceding processes in subsequent processes; the difference calculation is performed between the actual process processing state tensors obtained in subsequent processes and the inheritable predictable state tensors to obtain state residual tensors representing irreversible changes during process switching; and a state inheritance vector is constructed based on the state residual tensors.

[0051] The specific configuration of the cross-process correlation analysis module 20 will be described in detail below. The cross-process correlation analysis module 20 further includes: using the energy proportion of the state residual tensor in each modal feature dimension and the stability index within the sliding time window to effectively screen the corresponding state components in the inheritable predictable state tensor, determining the effective inherited state components that can be stably transferred across processes; configuring transfer weights for the effective inherited state components using the sensitivity of each state component in the state residual tensor to the adjustment results of subsequent processing parameters; and combining the effective inherited state components, transfer weights, and residual constraint components representing the intensity of irreversible changes to construct a state inheritance vector characterizing the degree of inheritance and constraint relationship of the standard part's processing state between adjacent processes.

[0052] The specific configuration of the cross-process correlation analysis module 20 will be described in detail below. The cross-process correlation analysis module 20 further includes: a process difference determination unit, used to generate a state identifier matrix to identify the inheritance relationship of each state dimension before and after process switching based on the process type differences and clamping datum inheritance relationship of adjacent processing processes; a state mapping unit, used to perform datum alignment and mapping operations on each state dimension in the processing state tensor of the preceding process according to the state identifier matrix to form inheritable state components; a non-inheritable separation unit, used to peel off state dimensions that do not meet the inheritance conditions from the processing state tensor of the preceding process to form state residual components; and an output combination unit, used to combine the inheritable state components with the state residual components to form the output tensor of the cross-process inheritance decomposition operator.

[0053] The specific configuration of the collaborative control management module 40 will be described in detail below. The collaborative control management module 40 further includes: dividing the processing parameter set into multi-dimensional parameter subspaces based on the adjustable space and control degrees of freedom of each processing step, with each parameter subspace corresponding to different process adjustment degrees of freedom and processing sensitivity; mapping the collaborative influence mapping relationship to the multi-dimensional parameter subspaces to apply collaborative constraints, limiting the parameter adjustment range and allowable variation amplitude; and after applying constraints to the candidate parameter groups in each parameter subspace based on the overall processing consistency constraints, performing joint optimization calculations and outputting the joint optimization results.

[0054] The specific configuration of the collaborative control management module 40 will be described in detail below. The collaborative control management module 40 further includes: performing local optimization calculations within the neighborhood of each candidate parameter group to obtain the locally optimal parameter combination and its corresponding performance index; determining whether to trigger a jump operation based on the performance gradient of the performance index, the sensitivity of the state inheritance vector, and the collaborative influence mapping constraints, wherein the jump operation is used to escape the local optimum; if a jump operation is triggered, configuring the jump direction and magnitude based on minimizing processing error, maximizing processing efficiency, and satisfying the multi-process consistency constraints, and performing a cross-subspace jump operation to perform joint optimization.

[0055] The specific configuration of the collaborative control management module 40 will be described in detail below. The collaborative control management module 40 further includes: obtaining a trust level identifier of the joint optimization result; configuring an adjustable margin space based on the trust level identifier; and using the margin space to perform margin self-optimization updates in the collaborative control process.

[0056] The collaborative control system for integrated processing of standard parts in multiple processes provided in this embodiment of the invention can execute the collaborative control method for integrated processing of standard parts in multiple processes provided in this embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0057] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention are all within the scope of the present invention.

Claims

1. A method for the coordinated control of the integrated processing of multiple workpieces, characterized in that, The method includes: During the process of the standard part undergoing at least two processing steps, multi-source processing state data representing the geometric offset, processing response and clamping stability of the standard part under each process are collected, and the multi-source processing state data is input into the processing state characterization model of the component under each process. The processing state characterization model is used to perform cross-process correlation analysis, extract the state components that can be transmitted between adjacent processes and the components that undergo irreversible changes during the supply and demand switching process, and generate a state inheritance vector that characterizes the cross-process inheritance relationship of the standard part processing state. Based on the state inheritance vector, the degree of influence of the processing results of the preceding process on the adjustable space and control degree of freedom of the subsequent process is evaluated, and a collaborative influence mapping relationship between multiple processes is formed. Based on the aforementioned collaborative influence mapping relationship and overall processing consistency constraints, a joint optimization of the processing parameter configuration strategy is performed, and collaborative control management is executed based on the joint optimization results.

2. The collaborative control method for integrated machining of standard parts across multiple processes as described in claim 1, characterized in that, Cross-process correlation analysis is performed using the aforementioned processing state characterization model, including: The processing state characterization model includes a multi-source feature encoding layer, a unified state tensor construction layer, and a cross-process inheritance decomposition operator, wherein: The multi-source feature coding layer decomposes the multi-source machining state data to obtain a set of measurement data, cutting power or torque data, vibration data, clamping force data, and temperature data, and constructs a decomposed feature set. Extract frequency domain peak energy, spectral entropy, temperature rise slope, clamping force attenuation rate, and set deviation principal direction coefficient from the decomposed feature set to construct a modal feature vector; Activate the unified state tensor construction layer, align the modal feature vectors in the same machining reference coordinate system and aggregate them according to the sliding time window to generate the process machining state tensor; Activate the cross-process inheritance decomposition operator and construct the corresponding cross-process state transition operator based on the process difference type and state benchmark inheritance relationship between adjacent processing processes; The cross-process state transition operator is used to predict and map the processing state tensor of the preceding process, and an inheritable predictable state tensor of the processing state of the preceding process in the subsequent process is constructed. The difference between the process processing state tensor actually obtained in subsequent processes and the inheritable predictable state tensor is calculated to obtain the state residual tensor that represents the irreversible changes that occur during process switching. Construct a state inheritance vector based on the state residual tensor.

3. The collaborative control method for integrated machining of standard parts in multiple processes as described in claim 2, characterized in that, Constructing a state inheritance vector based on the state residual tensor includes: By utilizing the energy proportion of the state residual tensor in each modal feature dimension and the stability index within the sliding time window, the state components corresponding to the inheritable predictive state tensor are effectively screened to determine the effective inheritable state components that can be stably transferred across processes. By utilizing the sensitivity of each state component in the state residual tensor to the adjustment results of subsequent processing parameters, a transfer weight is configured for the effective inherited state component. The effective inherited state components, the transfer weights, and the residual constraint components that characterize the intensity of irreversible changes are combined to construct a state inheritance vector that characterizes the degree of inheritance and constraint relationship of the standard part processing state between adjacent processes.

4. The collaborative control method for integrated machining of standard parts in multiple processes as described in claim 2, characterized in that, Cross-process inheritance decomposition operators include: The process difference determination unit is used to generate a state identifier matrix to identify the inheritance relationship of each state dimension before and after the process switching, based on the process type differences between adjacent processing operations and the clamping datum inheritance relationship. The state mapping unit is used to perform reference alignment and mapping operations on each state dimension in the processing state tensor of the previous process according to the state identifier matrix to form inheritable state components. Non-inheritable separation units are used to separate state dimensions that do not meet the inheritance conditions from the state tensor of the preceding process, forming state residual components. The output combination unit is used to combine the inheritable state components and the state residual components to form the output tensor of the cross-process inheritance decomposition operator.

5. The collaborative control method for integrated machining of standard parts across multiple processes as described in claim 1, characterized in that, Joint optimization of processing parameter configuration strategies based on the aforementioned synergistic influence mapping relationship and overall processing consistency constraints includes: Based on the adjustable space and control degrees of freedom of each processing step, the set of processing parameters is divided into multi-dimensional parameter subspaces, each parameter subspace corresponding to different process adjustment degrees of freedom and processing sensitivity; The synergistic influence mapping relationship is mapped to a multidimensional parameter subspace to apply synergistic constraints, limiting the parameter adjustment range and allowable variation magnitude; After applying constraints to the candidate parameter groups in each parameter subspace based on the overall processing consistency constraint, joint optimization calculation is performed, and the joint optimization result is output.

6. The collaborative control method for integrated machining of standard parts in multiple processes as described in claim 5, characterized in that, After applying constraints to the candidate parameter groups in each parameter subspace based on the overall processing consistency constraint, joint optimization calculations are performed, including: Local optimization calculations are performed in the neighborhood of each candidate parameter group to obtain the locally optimal parameter combination and the corresponding performance index. Based on the performance gradient of the performance index, the sensitivity of the state inheritance vector, and the cooperative influence mapping constraint, it is determined whether to trigger a jump operation, which is used to escape local optima. If a jump operation is triggered, the jump direction and magnitude are configured based on minimizing processing error, maximizing processing efficiency, and satisfying the consistency constraints of multiple processes. A cross-subspace jump operation is then performed to conduct joint optimization.

7. The collaborative control method for integrated machining of standard parts in multiple processes as described in claim 1, characterized in that, Based on the joint optimization results, collaborative control management is implemented, including: Obtain a trust level identifier for the joint optimization results, and configure an adjustable margin space based on the trust level identifier; The margin space is used for margin self-optimization update of the collaborative control process.

8. A collaborative control system for integrated machining of standard parts across multiple processes, characterized in that, The system is used to implement the collaborative control method for integrated processing of multi-process standard parts as described in any one of claims 1 to 7, the system comprising: The processing status data acquisition module is used to collect multi-source processing status data representing the geometric offset, processing response and clamping stability of the standard part during at least two processing steps, and input the multi-source processing status data into the processing status characterization model of the component under each process. The cross-process correlation analysis module is used to perform cross-process correlation analysis using the processing state characterization model, extract the state components that can be transmitted between adjacent processes and the components that undergo irreversible changes during the supply and demand switching process, and generate a state inheritance vector that characterizes the cross-process inheritance relationship of the standard part processing state. The impact assessment module is used to assess the degree of impact of the processing results of the preceding process on the adjustable space and control freedom of the subsequent process based on the state inheritance vector, and to form a collaborative impact mapping relationship between multiple processes. The collaborative control management module is used to perform joint optimization of the processing parameter configuration strategy based on the collaborative influence mapping relationship and the overall processing consistency constraint, and to perform collaborative control management based on the joint optimization result.

Images