A wind tunnel model part machining process parameter optimization method

Abstract

The application provides a wind tunnel model part machining process parameter optimization method, and belongs to the technical field of part machining. The method comprises the following steps: collecting working condition, material and equipment state data to formulate initial parameters; real-time acquisition of machining quality time sequence data, analysis to obtain quality indexes and deviation degrees; calculation of dynamic compensation parameters combined with disturbance types, adjustment to generate real-time parameters; threshold-based feedback optimization of parameters, parallel coupling analysis of working conditions and quality data to extract abnormal characteristics, and deep correction of parameters if the optimization does not achieve the expected result. Through full-process dynamic sensing, the application identifies the instantaneous changes in the machining process, and according to a multi-level response mechanism, it can quickly suppress regular disturbances and locate the abnormal source to implement targeted correction. The application converts experience rules into a quantitative correlation model, reduces the dependence on manual operation, improves the yield and consistency of precision component machining, and provides a reliable intelligent manufacturing scheme.

Description

Technical Field

[0001] This invention belongs to the field of parts processing technology, specifically a method for optimizing the processing parameters of wind tunnel model parts. Background Technology

[0002] In the field of high-end equipment manufacturing, precision components that undertake key testing tasks have almost stringent requirements for processing quality. Their geometric accuracy and surface integrity directly determine the validity of subsequent test data and the safety of equipment operation.

[0003] However, for a long time, the processing of such components has relied heavily on fixed parameters set by human experience. Static process solutions are often insufficient to cope with the constantly changing physical environment, subtle fluctuations in material properties, and dynamic wear and tear on the equipment itself. In actual production, changes in ambient temperature cause thermal deformation of components and equipment, mechanical fluctuations during cutting lead to vibration patterns, and progressive tool wear subtly alters machining accuracy. These factors intertwine, often resulting in untraceable precision deviations in the final product. Traditional quality control models focus primarily on post-production inspection and correction, lacking the ability to dynamically perceive and predict the entire processing cycle. Once a problem occurs, machine downtime for adjustments is often necessary, impacting production efficiency and failing to fundamentally eliminate potential quality issues, thus failing to meet the urgent need for mass production and standardized manufacturing of high-precision components. Summary of the Invention

[0004] This invention provides a method for optimizing the processing parameters of wind tunnel model parts, in order to overcome the deficiencies existing in the prior art.

[0005] This invention provides a method for optimizing the machining process parameters of wind tunnel model parts, including:

[0006] Collect processing condition data, material property data, and equipment status data, and formulate initial process parameters based on the material property data and wind tunnel model design specifications.

[0007] Real-time quality inspection data of the part processing process is collected, and the processing quality index is obtained by analyzing the time-series quality inspection data using the multi-modal quality state hierarchical analysis method. The corresponding processing quality deviation index is also calculated.

[0008] Dynamic compensation parameters are calculated based on the type of disturbance in the processing conditions and the deviation index of processing quality. The initial process parameters are then adjusted based on the dynamic compensation parameters to obtain the real-time process parameters.

[0009] Determine whether the current processing quality index has reached the preset quality threshold. If so, maintain the real-time process parameters; otherwise, perform proportional feedback optimization based on threshold deviation on the dynamic compensation parameters, and set the upper limit of single-step adjustment and the iterative convergence threshold.

[0010] During the processing, a parallel working condition coupling analysis method is used to analyze the processing working condition data and the time-series quality inspection data to obtain the correlation influence data. Feature extraction is performed on the current processing quality indicators to obtain processing anomaly characteristics. When the feedback optimization does not meet expectations, the dynamic compensation parameters are deeply adjusted based on the correlation influence data to obtain the process correction parameters.

[0011] According to the present invention, a method for optimizing the machining process parameters of wind tunnel model parts includes machining condition data such as temperature change data, cutting vibration data, and cutting force pulsation data. Material property data includes hardness data, elastic modulus data, coefficient of thermal expansion data, and residual stress distribution data. Equipment status data includes spindle rotation accuracy data, tool wear status data, guideway geometric error data, and cooling and lubrication status data.

[0012] According to the method for optimizing the machining process parameters of wind tunnel model parts provided by the present invention, the process of formulating initial process parameters includes: Based on the wind tunnel model design specifications, the differentiated process requirements of force measurement models, pressure measurement models, and special test models are distinguished and transformed into accuracy level indicators and surface quality level indicators. Cutting parameter constraint boundaries are then formulated based on material property data.

[0013] By combining the matching characteristics of tool material and workpiece material, suitable tool types are selected, and the cutting depth allocation for roughing, semi-finishing and finishing is determined according to the geometric features and accuracy level of the part.

[0014] The reference cutting speed is determined based on the spindle power range and the tool's limit cutting speed, and the feed strategy is selected according to the accuracy grade index and the surface quality grade index.

[0015] The tool path is planned based on the equipment stiffness characteristics and geometric tolerance requirements, and the tool wear warning threshold is set based on the material machinability index.

[0016] According to the method for optimizing the machining process parameters of wind tunnel model parts provided by the present invention, the process of analyzing and obtaining machining quality indicators includes: The data collected include part size detection signals, surface integrity data, and process parameters, which are then preprocessed to serve as time-series quality inspection data.

[0017] Dimensional deviation, geometric tolerance, and profile are extracted from part dimension inspection signals as dimensional accuracy indicators, while surface roughness, surface residual stress, and work-hardened layer depth are extracted from surface integrity data as surface quality indicators.

[0018] Tool wear rate, cutting force fluctuation amplitude, and spindle thermal drift are extracted from process parameters as process stability indicators.

[0019] The trend characteristics and working condition correlation indicators of dimensional accuracy indicators, surface quality indicators and process stability indicators are extracted to form a multi-dimensional and hierarchical processing quality index.

[0020] The processing quality deviation index is the absolute deviation value and relative deviation rate between the measured value of each processing quality index and the corresponding preset quality threshold.

[0021] According to the method for optimizing the machining process parameters of wind tunnel model parts provided by the present invention, the process of calculating dynamic compensation parameters includes: By integrating machining condition data, material property data, and equipment status data, disturbance types are classified according to thermal deformation disturbance, tool wear disturbance, cutting vibration disturbance, and material property fluctuation disturbance, and the disturbance level is quantified according to the degree of influence to obtain the condition disturbance data.

[0022] Determine the preset threshold for the processing quality deviation index, calculate the deviation value, classify the quality deviation level according to the degree of deviation, and perform time synchronization and correlation mapping with the working condition disturbance data.

[0023] Establish a disturbance type-process parameter mapping table to identify the dominant disturbance type in the complex disturbance and determine the strategy strength based on the disturbance level.

[0024] The parameter adjustment priority is determined based on the strategy strength, the adjustment range is determined based on the quality deviation, and the single-step adjustment amount is limited. Specific compensation parameters are added for the type of disturbance in the processing condition to obtain dynamic compensation parameters.

[0025] According to the method for optimizing the machining process parameters of wind tunnel model parts provided by the present invention, the process of adjusting and obtaining real-time process parameters includes: Based on equipment status data and historical process adjustment records, the impact of dynamic compensation parameters on processing quality is determined, and a sensitivity coefficient matrix is ​​formed.

[0026] The adjustment priorities are determined based on the directness of their impact on machining quality and the response speed, with the response speed from fastest to slowest being: tool compensation value adjustment, feed rate adjustment, spindle speed adjustment, and depth of cut adjustment.

[0027] A smooth transition mechanism is established for setting transition times and parameter activation conditions for continuous process parameters, preventing sudden parameter changes from causing processing chatter or surface quality jumps.

[0028] The current dynamic compensation parameters are superimposed on the initial process parameters according to the adjustment priority and sensitivity coefficient matrix. The rate of change of parameters before and after superposition and the reasons for the change are recorded to obtain the real-time process parameters.

[0029] According to the method for optimizing the processing parameters of wind tunnel model parts provided by the present invention, the process of obtaining correlated influence data includes: Multidimensional working condition features are extracted from processing condition data, and key quality features are extracted from time-series quality inspection data.

[0030] Correlation analysis was performed on multidimensional operating condition characteristics and key quality characteristics, and hysteresis response characteristics were marked, whereby hysteresis response characteristics reflect the time delay between changes in operating conditions and quality response.

[0031] Based on the combination of disturbance types, processing scenarios are divided. Under different processing scenarios, regression models of key quality features and multi-dimensional working condition features are constructed, the goodness of fit is calculated, and significant disturbance factors are screened.

[0032] The thermal correlation between temperature change and dimensional deviation, the vibrational correlation between cutting vibration and surface roughness, and the correlation between tool wear and profile accuracy degradation are analyzed, and a statistical analysis report is generated.

[0033] The statistical analysis report is converted into process adjustment execution rules. The confidence and support of all process adjustment execution rules are calculated, and reliable process adjustment execution rules that meet the preset threshold are retained to generate a working condition coupling rule library.

[0034] According to the method for optimizing the machining process parameters of wind tunnel model parts provided by the present invention, the process of generating a statistical analysis report includes: Plot a scatter plot of the spindle thermal drift and the corresponding dimensional deviation to identify the mapping relationship between the temperature gradient and the dimensional drift trend, and calculate the dimensional change rate corresponding to a unit temperature rise to obtain the thermal correlation.

[0035] The relationship between the root mean square value of cutting vibration acceleration and surface roughness is analyzed, a fitting curve is generated, and the correlation coefficient between the dominant vibration frequency and the surface texture spacing is calculated to obtain the vibration-mass correlation.

[0036] The correlation between the wear width of the tool's flank face and the profile deviation was statistically analyzed, and the correlation between the wear rate and the accuracy degradation rate was extracted to obtain the degradation correlation.

[0037] The thermal correlation, vibrational quality correlation, and degradation correlation are standardized and presented in the format of influencing factors - affected quality parameters - quantitative relationship - processing scenario adaptation suggestions to obtain a statistical analysis report.

[0038] According to the method for optimizing the machining process parameters of wind tunnel model parts provided by the present invention, the process of obtaining machining anomaly features includes: For the preset processing quality threshold, the abnormal threshold range of each processing quality indicator is determined, and for the time-series fluctuation, dynamic thresholds are set to adapt to the time-varying characteristics of the processing process.

[0039] Iterate through all time windows, and when any processing quality indicator exceeds the abnormal threshold range or dynamic threshold, it is determined to be an abnormal segment.

[0040] The corresponding abnormal segments were extracted from the dimensions of size abnormality, surface quality abnormality, and process stability abnormality, respectively.

[0041] By combining the associated machining condition data and material property data, the anomaly types of abnormal segments are marked. The anomaly types include thermal drift type, tool degradation type, vibration mark generation type, and material mutation type.

[0042] Calculate the Pearson correlation coefficient among all anomalous segments, retain the anomalous segments that reach the preset coefficient threshold, and combine them with the corresponding anomalous types as processing anomalous features.

[0043] According to the method for optimizing the machining process parameters of wind tunnel model parts provided by the present invention, the process of adjusting and obtaining the process correction parameters includes: Based on the associated impact data, the abnormal causes of the abnormal characteristics of the processing are located by feature matching, and a process adjustment mapping table is established by combining the priority adjustment parameters and strategy direction.

[0044] Conflicts in process parameter adjustments are identified and coordinated through preset constraint priorities. The preset constraint priorities are executed in the order that hardware constraints take precedence over process specification constraints, which in turn take precedence over efficiency optimization.

[0045] The parameter adjustment sequence is planned from fast to slow response speed. Sudden anomalies are given priority for rapid correction of the feed rate, followed by fine-tuning of the spindle speed and depth of cut.

[0046] By combining the correlation impact data, the tool compensation correction amount, cutting parameter correction index, and cooling parameter correction value are calculated as the process correction range, and the dynamic compensation parameters are adjusted to obtain the process correction parameters.

[0047] This invention provides a method for optimizing the processing parameters of wind tunnel model parts. By capturing multi-dimensional dynamic information from the processing site in real time, it can identify instantaneous changes in material properties, equipment status, and environmental conditions, providing comprehensive data support for process adjustments. By establishing a multi-level response mechanism, it can both rapidly mitigate conventional disturbances through real-time calculation of dynamic compensation parameters and locate the root cause of anomalies and implement targeted corrections when conventional adjustments fail through in-depth analysis of the correlation between operating conditions and quality.

[0048] By introducing a smooth transition and conflict coordination mechanism, continuous parameter optimization is achieved while ensuring processing safety, thus solving the chatter and quality step problems that are easily caused by traditional adjustment methods.

[0049] By transforming empirical process rules into quantifiable correlation models, the reliance on human experience is reduced, and the yield and consistency of high-precision component processing are improved, providing a reliable intelligent solution for the high-quality manufacturing of complex and precision parts. Attached Figure Description

[0050] The invention will now be further described with reference to the accompanying drawings.

[0051] Figure 1 This is a flowchart illustrating a method for optimizing the processing parameters of wind tunnel model parts according to the present invention. Figure 2 This is a schematic diagram of the process for determining the initial process parameters in this invention; Figure 3 This is a schematic diagram of the process for calculating dynamic compensation parameters in this invention. Detailed Implementation

[0052] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0053] like Figures 1 to 3 As shown in the figure, this invention provides a method for optimizing the machining process parameters of wind tunnel model parts. This method optimizes parameters for the five-axis CNC milling of 7075 aluminum alloy parts used in aerospace wind tunnel force measurement models. The machining equipment is a vertical five-axis machining center, equipped with a temperature-controlled cutting fluid system, an on-machine contact probe, a three-axis vibration acceleration sensor, a piezoelectric triaxial force gauge, and a tool vision inspection system. These types of parts have stringent requirements for form and position tolerances, surface integrity, and dimensional accuracy. The machining process is easily affected by multiple sources of disturbance, such as thermal deformation, tool wear, cutting vibration, and material property fluctuations. This method ensures machining quality through full-process control. The specific implementation steps are as follows: Data on machining conditions, material properties, and equipment status are collected. Initial process parameters are formulated based on material property data and wind tunnel model design specifications. During the acquisition of machining condition data, temperature change data is collected in real time using PT100 platinum resistance sensors located at the front end of the machine tool spindle, the worktable, and inside the protective cover of the cutting zone, with a temperature measurement range covering -20℃ to 150℃. Cutting vibration data is collected using a triaxial piezoelectric accelerometer installed on the spindle end housing, simultaneously acquiring time-domain signals of vibration acceleration in the X, Y, and Z orthogonal directions. Cutting force pulsation data is collected using a triaxial piezoelectric force gauge installed below the workpiece fixture, with a range covering the main cutting force from 0 to 10000N and the feed resistance and back force from 0 to 5000N, outputting real-time force value signals in three directions.

[0054] All operating data is transmitted to the industrial control computer via industrial Ethernet. A unified machine tool system clock is used to mark the timestamps, and data with different sampling frequencies are aligned using linear interpolation.

[0055] During the material property data acquisition process, hardness data was obtained by distributing at least five points on the surface of the workpiece blank using a Rockwell hardness tester, and the average value was taken as the reference hardness of the blank. The test points were evenly distributed within the machined area of ​​the part. The elastic modulus and coefficient of thermal expansion were obtained by using a universal testing machine with a high and low temperature test chamber, covering a temperature range of 20℃ to 120℃. A set of data was collected every 20℃, and a temperature-performance curve was fitted. Residual stress distribution data was obtained by using an X-ray stress analyzer along the depth direction of the machined surface of the part, with a depth detection interval of 0.05mm, and the residual stress distribution curve along the depth direction was output. All material property data were collected during the blank incoming inspection stage and entered into the process database as the basis for initial parameter setting.

[0056] During equipment status data acquisition, spindle rotation accuracy data is detected using a laser interferometer to measure radial runout, axial runout, and angular oscillation errors. The measured speeds cover the commonly used machining range of 500 r / min to 8000 r / min, with a set of data collected every 1000 r / min to form a speed-rotation accuracy mapping table. Tool wear status data is collected periodically after each tool change and during machining intervals using a tool vision inspection system. Images of the tool's flank face are captured, and the wear band width is extracted using image processing algorithms. The sampling interval is set to be once every 15 minutes based on the machining duration. Guide rail geometric error data is detected using a laser interferometer to measure the straightness, perpendicularity, pitch angle, and yaw angle errors of each axis guide rail. Calibration testing is completed during the equipment's regular maintenance phase. Cooling and lubrication status data is collected in real time using flow sensors, temperature sensors, and concentration sensors installed on the cutting fluid pipeline, outputting real-time cutting fluid flow rate, outlet temperature, and concentration values.

[0057] In the process of formulating initial process parameters, the different process requirements of force measurement model, pressure measurement model and special test model are distinguished according to the wind tunnel model design specifications. The tolerance requirements in the design drawings are transformed into accuracy level indicators and surface quality level indicators. Cutting parameter constraint boundaries are formulated based on material property data.

[0058] The core force-bearing surface of the force measurement model is required to reach the IT3 to IT4 level accuracy, and the surface roughness requirement is Ra≤0.4μm, with the most stringent cutting parameter constraints; the pressure measurement model must ensure the integrity of the surface around the pressure measurement hole, and the surface roughness requirement is Ra≤0.8μm.

[0059] When setting the boundary constraints for cutting parameters, for 7075 aluminum alloy, the upper limit of the reference cutting speed is set to 800 m / min, the upper limit of the feed per tooth is set to 0.2 mm / z, and the upper limit of the depth of cut in a single pass is set to 2 mm; for TC4 titanium alloy, the upper limit of the reference cutting speed is set to 120 m / min, the upper limit of the feed per tooth is set to 0.15 mm / z, and the upper limit of the depth of cut in a single pass is set to 1.5 mm.

[0060] When the measured hardness of the material is more than 5% higher than the standard hardness of the grade, the constraint boundary of the cutting speed and feed rate is reduced by 10%; when the hardness is more than 5% lower than the standard hardness, the constraint boundary is increased by 8% to ensure that the parameters are adapted to the actual performance of the material.

[0061] By combining the matching characteristics of tool material and workpiece material, suitable tool types are selected, and the cutting depth allocation for roughing, semi-finishing and finishing is determined according to the geometric features and accuracy level of the part.

[0062] During tool matching, carbide diamond-coated ball end mills and end mills are preferred for machining aluminum alloys, while carbide TiAlN coated tools are preferred for machining titanium alloys. Ball end mills are used for curved surface structures, long neck reduction end mills are used for deep cavity structures, and large core diameter close-tooth end mills are used for thin-walled structures.

[0063] The depth of cut distribution follows the principle of "roughing to remove excess material, semi-finishing to refine the contour, and finishing to ensure accuracy." In the roughing stage, 70% to 80% of the total machining allowance is removed, with a single depth of cut of 1mm to 3mm and a side depth of cut of 70% to 80% of the tool diameter. In the semi-finishing stage, 15% to 20% of the machining allowance is removed, with a single depth of cut of 0.3mm to 0.8mm, leaving 0.1mm to 0.2mm for finishing. In the finishing stage, the remaining finishing allowance is removed, with a single depth of cut of 0.1mm to 0.2mm, ensuring the final machining accuracy and surface quality.

[0064] The reference cutting speed is determined based on the spindle power range and the tool's limit cutting speed, and the feed strategy is selected according to the accuracy grade index and the surface quality grade index.

[0065] The reference cutting speed is taken as the midpoint of the intersection between the optimal power range of 60% to 80% of the machine tool's rated power and the tool's limit cutting speed, taking into account both machining efficiency and tool life. A lower cutting speed is selected in the roughing stage, and a higher cutting speed is selected in the finishing stage.

[0066] In terms of feed strategy selection, a constant feed rate strategy is adopted in the finishing stage where high precision is required, and the feed rate is kept constant throughout the tool travel to ensure uniform surface texture. In the roughing stage, a variable feed rate strategy is adopted. In corners and deep grooves with large cutting allowances, the feed rate is reduced to 60% to 80% of the set value, while in flat areas with small allowances, the feed rate is increased to 110% to 120% of the set value to balance machining efficiency and tool wear.

[0067] The tool path is planned based on the equipment stiffness characteristics and geometric tolerance requirements, and the tool wear warning threshold is set based on the material machinability index.

[0068] During the tool path planning process, reciprocating or unidirectional tool path is used for planar machining, and contour or spiral tool path is used for curved surface machining. The cantilever machining area with weak equipment rigidity is actively avoided. For mating surfaces and datum surfaces with high form and position tolerance requirements, a finishing tool path is added. The tool step distance is set to 10% to 30% of the tool diameter according to the surface roughness requirements.

[0069] The material machinability index is calculated by weighting four parameters: material hardness, strength, thermal conductivity, and plastic elongation. The value ranges from 0 to 1, with higher values ​​indicating better machinability. In the calculation process, hardness accounts for 40% of the weight, strength accounts for 30%, thermal conductivity accounts for 20%, and plastic elongation accounts for 10%. All parameters are standardized and then weighted and summed.

[0070] The machinability index of 7075 aluminum alloy is 0.75, and that of TC4 titanium alloy is 0.35. The corresponding tool wear warning threshold is set based on the material's machinability index. For every 0.1 decrease in the machinability index, the warning threshold is lowered by 0.05 mm. The warning threshold for tool flank wear is 0.3 mm for 7075 aluminum alloy and 0.2 mm for TC4 titanium alloy. When the wear reaches the warning threshold, a compensation adjustment reminder is triggered; when the wear reaches 0.5 mm, a forced tool change reminder is triggered.

[0071] Real-time quality inspection data of the part machining process is collected. A multi-modal quality state hierarchical analysis method is used to analyze the time-series quality inspection data to obtain machining quality indicators and calculate the corresponding machining quality deviation indicators. During the time-series quality inspection data acquisition process, part size detection signals are obtained point-by-point through machine contact probes. Inspection is performed once after each operation in the roughing stage, once after each layer of tool movement in the semi-finishing stage, and once after each tool movement in the finishing stage. The number of inspection points is set to 10 to 50 depending on the part size. Surface integrity data is detected periodically during machining intervals using a portable surface roughness tester and an X-ray stress analyzer. Simultaneously, surface integrity parameters are indirectly predicted through a pre-trained prediction model, combined with cutting vibration and cutting force signals. The detection interval is synchronized with tool wear detection. Process parameters, including real-time machine tool operating parameters such as spindle speed, feed rate, depth of cut, and cutting fluid flow rate, are directly acquired through the machine tool CNC system interface at a sampling frequency of 10Hz. The collected raw data is first preprocessed. The db5 wavelet basis is used for three-layer wavelet threshold filtering to remove high-frequency noise and random interference. A small number of missing data are filled in by linear interpolation. Outliers exceeding three times the standard deviation are removed according to the 3σ criterion. All preprocessed data are aligned with a unified timestamp to form a standard time series quality detection dataset.

[0072] The multi-modal quality state hierarchical analysis method analyzes multi-source heterogeneous quality inspection data layer by layer: dimensional accuracy layer, surface quality layer, and process stability layer. Each layer extracts corresponding feature indicators, which are then standardized and weighted and fused to form a comprehensive processing quality indicator system. Dimensional deviation, geometric tolerance, and profile are extracted from the part dimensional inspection signals as dimensional accuracy indicators. Dimensional deviation is the difference between the measured size and the theoretical design size, and is divided into absolute deviation and relative deviation. Geometric tolerance includes straightness, flatness, roundness, cylindricity, parallelism, and perpendicularity, and is obtained according to the testing and calculation methods specified in GB / T1182. Profile includes line profile and surface profile, calculated by the maximum normal distance from discrete inspection points to the theoretical profile. Surface roughness, residual stress, and work-hardened layer depth are extracted from surface integrity data as surface quality indicators. Surface roughness is evaluated using the arithmetic mean deviation Ra as the core indicator, supplemented by the maximum profile height Rz. Residual stress includes the peak values ​​and corresponding distribution depths of residual tensile and compressive stresses in the surface layer. The work-hardened layer depth is the depth of the area where the surface hardness is 10% higher than the substrate hardness after machining, obtained by point-by-point detection along the depth direction using a microhardness tester. Tool wear rate, cutting force fluctuation amplitude, and spindle thermal drift are extracted from process parameters as process stability indicators. Tool wear rate is the increment of the tool flank wear width per unit machining time, in μm / min. Cutting force fluctuation amplitude is the difference between the maximum and minimum cutting force values ​​within a set time window, supplemented by the cutting force coefficient of variation (i.e., the ratio of standard deviation to mean) as a stability evaluation indicator. Spindle thermal drift is the axial and radial displacement of the spindle relative to its initial cooling position during the heating process, in μm.

[0073] Trend characteristics and working condition correlation indicators of dimensional accuracy, surface quality, and process stability indicators are extracted to form a multi-dimensional, hierarchical processing quality index. Trend characteristics are obtained by linearly fitting each indicator to a 60-second sliding time window, with the fitting slope used as the trend feature value to reflect the increasing, decreasing, or stable changes in the quality index with processing time. Working condition correlation indicators are obtained by calculating the correlation coefficient between each quality index and the corresponding working condition parameters using Pearson correlation analysis, reflecting the degree of influence of working condition changes on processing quality. All hierarchical indicators are first subjected to range standardization, mapping the values ​​to the interval between 0 and 1. The standardization formula is x'=(x-x_min) / (x_max-x_min), where x is the original index value, x_min is the historical minimum value of the index, and x_max is the historical maximum value. After standardization, a weighted summation method is used to calculate the comprehensive processing quality index. The weights are determined by a combination of the analytic hierarchy process (AHP) and the entropy weighting method. First, subjective weights are obtained through the AHP, and then objective weights are calculated based on the information entropy of historical data using the entropy weighting method. Finally, the subjective and objective weights are weighted at a ratio of 0.6:0.4 to obtain the combined weight. In conventional processing scenarios, the combined weight of dimensional accuracy index accounts for 50%, surface quality index accounts for 30%, and process stability index accounts for 20%. For parts with special precision requirements, the weight ratio can be adjusted according to actual needs.

[0074] The processing quality deviation index is divided into individual index deviation and comprehensive deviation. The absolute deviation value of an individual index is the absolute value of the difference between the measured value and the preset threshold, and the relative deviation rate is the ratio of the absolute deviation value to the preset threshold, expressed as a percentage. The comprehensive processing quality deviation is obtained by weighted summation of the relative deviation rates of each individual index, with the weights consistent with those of the comprehensive quality index. The quality deviation levels are divided into four levels: qualified, slight deviation, moderate deviation, and severe deviation, corresponding to relative deviation rates of ≤5%, 5% < deviation ≤20%, 20% < deviation ≤50%, and deviation >50%, respectively.

[0075] Dynamic compensation parameters are calculated based on the type of disturbance in the machining condition and the deviation index of machining quality. The initial process parameters are then adjusted based on these dynamic compensation parameters to obtain real-time process parameters. In calculating the dynamic compensation parameters, machining condition data, material property data, and equipment status data are first integrated. Disturbances are categorized into thermally induced deformation disturbances, tool wear disturbances, cutting vibration disturbances, and material property fluctuation disturbances. The disturbance levels are then quantified according to their impact to obtain the condition disturbance data. Thermally induced deformation disturbances are caused by the accumulation of cutting heat and fluctuations in ambient temperature, leading to thermal expansion and contraction of the machine tool spindle, worktable, tool, and workpiece, resulting in machining dimensional deviations. Tool wear disturbances are caused by normal tool wear and abnormal chipping during cutting, leading to deterioration of the cutting edge profile and affecting machining accuracy and surface quality. Cutting vibration disturbances are caused by factors such as cutting force impact, insufficient machine tool rigidity, and excessive tool overhang, and are classified into free vibration, forced vibration, and chatter, resulting in surface texture and dimensional fluctuations. Material property fluctuation disturbances are caused by factors such as uneven workpiece blank hardness, uneven residual stress distribution, and component segregation, leading to inconsistencies between cutting force fluctuations and machining deformation.

[0076] The disturbance level was quantified using the fuzzy comprehensive evaluation method. First, the set of evaluation factors was determined, including five evaluation indicators: temperature change rate, root mean square value of vibration acceleration, coefficient of variation of cutting force, tool wear rate, and material hardness deviation rate. Second, the weight vectors of each evaluation factor were determined using the analytic hierarchy process (AHP), with weights of 0.25, 0.2, 0.2, 0.25, and 0.1. Then, membership functions corresponding to the four evaluation levels were established, with each indicator using a triangular membership function. The membership degree of each disturbance level was calculated based on the indicator values. Finally, the weight vector was multiplied by the membership matrix to obtain the comprehensive evaluation vector. The level corresponding to the highest membership degree was taken as the final disturbance level, corresponding to four levels: slight, moderate, significant, and severe, with corresponding quantitative scores of 1, 2, 3, and 4.

[0077] A preset threshold for the processing quality deviation index is determined, the deviation value is calculated, and the quality deviation level is classified according to the degree of deviation. Time synchronization and correlation mapping are then performed with the operating condition disturbance data. During the time synchronization process, cross-correlation analysis is used to calculate the time delay between the disturbance occurrence and the quality response. The cross-correlation function formula is as follows:

[0078] In the formula, For time series data of disturbance quantities, For quality indicator time series data, This is the amount of time delay. Given the total length of the data, take the value corresponding to the maximum value of the cross-correlation function. The value is used as the time delay between the disturbance and the quality index. Based on the calculated time delay, the operating condition disturbance data and quality deviation data are aligned to establish a disturbance-quality deviation correlation mapping matrix and identify the main quality deviation items corresponding to different disturbance types.

[0079] A disturbance type-process parameter mapping table is established to identify the dominant disturbance type in complex disturbances and determine the strategy intensity based on the disturbance level. The mapping table clarifies the priority adjustment parameters, adjustment direction, and adjustment strategy corresponding to each disturbance type. For thermally induced deformation disturbances, the priority is to adjust the tool compensation value and cutting fluid parameters to compensate for thermal deformation while reducing the cutting speed to reduce heat generation. For tool wear disturbances, the priority is to adjust the tool compensation value and feed rate to compensate for wear while reducing the feed rate to slow down the wear rate. For cutting vibration disturbances, the priority is to adjust the spindle speed and feed rate to avoid the resonance speed range while reducing the feed rate to reduce the cutting force amplitude. For material property fluctuation disturbances, the priority is to adjust the depth of cut and feed rate. In areas with increased hardness, the depth of cut and feed rate are reduced, while in areas with decreased hardness, the machining parameters are appropriately increased. In complex disturbance scenarios with multiple disturbances superimposed, principal component analysis is used to identify the dominant disturbance type. The influence degree of each disturbance source is extracted using principal components. The disturbance with a load ratio exceeding 60% in the first principal component is determined to be the dominant disturbance. The adjustment strategy for the dominant disturbance is the primary approach, supplemented by parameter adjustments for secondary disturbances, to avoid conflicting adjustment directions. The strategy strength coefficient corresponds one-to-one with the disturbance level. The coefficient ranges from 0.3 to 0.5 for slight disturbances, from 0.5 to 0.8 for moderate disturbances, from 0.8 to 1.2 for larger disturbances, and from 1.2 to 1.5 for severe disturbances. It is used to adjust the overall scaling parameter range.

[0080] The parameter adjustment priority is determined based on the strategy strength, the adjustment range is determined based on the quality deviation, and the single-step adjustment amount is limited. Specific compensation parameters are added for different types of machining disturbances to obtain a dynamic compensation parameter set. Parameter adjustment priority is sorted from fastest to slowest response speed: tool compensation value adjustment, feed rate adjustment, spindle speed adjustment, and depth of cut adjustment. Parameters with fast response and direct impact on quality are adjusted first to quickly suppress quality deviation. The adjustment range adopts a proportional adjustment method, directly proportional to the quality deviation. A single-step adjustment limit is set: 5μm for tool compensation value, 10% for feed rate, 8% for spindle speed, and 5% for depth of cut to prevent sudden parameter changes from causing machining chatter or surface quality abrupt changes. Specific compensation parameters are added for different disturbance types: cutting fluid flow and temperature adjustment parameters are added for thermally induced deformation disturbances; damper damping parameters and tool path adjustment parameters are added for cutting vibration disturbances; and regional variable parameter machining compensation parameters are added for material property fluctuation disturbances, forming a complete dynamic compensation parameter set.

[0081] The comprehensive compensation factor for the dynamic compensation parameters is calculated using the following formula:

[0082] In the formula, The comprehensive compensation factor is dimensionless and is used to characterize the overall adjustment ratio of process parameters under the combined effect of current disturbance and quality deviation. For the first The weighting coefficients for each quality indicator are calculated, and the sum of all weighting coefficients is 1. For the first The measured deviation of each quality indicator; For the first Reference deviation thresholds for each quality indicator; This represents the strategy strength coefficient corresponding to the current disturbance level.

[0083] After the comprehensive compensation factor is calculated, the target adjustment amount of each specific process parameter is obtained by mapping through the process parameter sensitivity coefficient matrix. The mapping formula is as follows:

[0084] In the formula, This is a vector of process parameter adjustments, including adjustments for four parameters: tool compensation, feed rate, spindle speed, and depth of cut. is the Moore-Penrose pseudo-inverse matrix of the sensitivity coefficient matrix. The element in the i-th row and j-th column of the sensitivity coefficient matrix represents the change in the i-th quality index caused by the unit change of the j-th process parameter. This is the quality index deviation vector, which includes the measured deviation values ​​of each individual quality index. This mapping relationship is verified by dimensions. The dimensions of the elements of the sensitivity coefficient matrix are mass dimension / parameter dimension, and the dimensions of its pseudo-inverse matrix are parameter dimension / mass dimension. After multiplying with the quality deviation vector, the parameter dimension is obtained. After multiplying by the dimensionless comprehensive compensation factor, the dimension of the final adjustment quantity is consistent with the dimension of the process parameter.

[0085] In the process of adjusting real-time process parameters, the impact of dynamic compensation parameters on machining quality is first determined based on equipment status data and historical process adjustment records, forming a sensitivity coefficient matrix. This sensitivity coefficient matrix is ​​calibrated through single-factor controlled variable experiments. During calibration, other process parameters and machining conditions are fixed, and the value of a specific process parameter is changed sequentially. After each adjustment, the change in the corresponding machining quality index is measured, and the sensitivity coefficient of that parameter to each quality index is obtained through linear regression fitting. Calibration experiments must be conducted in both cold and thermal equilibrium states of the machine tool to obtain sensitivity coefficient matrices for different temperature ranges. Recalibration is performed after every 10 batches of parts are machined to ensure the accuracy of the coefficients.

[0086] Adjustment priorities are established based on the directness of their impact on machining quality and their response speed, with the response speed from fastest to slowest as follows: tool compensation value adjustment, feed rate adjustment, spindle speed adjustment, and depth of cut adjustment. A smooth transition mechanism is established by setting transition times and parameter activation conditions for continuous process parameters to prevent machining chatter or surface quality abrupt changes caused by parameter mutations. Tool compensation values ​​are discrete values ​​that can be adjusted directly; feed rate, spindle speed, and depth of cut are continuously varying parameters, with linear transition times set from 1s to 5s. Parameters gradually change to the target value within this transition time, with the transition time adaptively set according to the magnitude of the parameter adjustment—the larger the adjustment magnitude, the longer the transition time. Parameter activation requires meeting preset conditions, including the current cutting position being on a non-critical machining surface, cutting force fluctuations being within a stable range, and the spindle speed being stable without sudden changes. If the activation conditions are not met, the adjustment is delayed until the next safe adjustment point to avoid tool marks or quality abrupt changes on critical surfaces of the part.

[0087] The current dynamic compensation parameters are superimposed on the initial process parameters according to the adjustment priority and sensitivity coefficient matrix. The rate of change and the reason for the change are recorded before and after superposition to obtain the real-time process parameters. After parameter superposition, the value is checked to see if it is within the boundary of the cutting parameters. If it exceeds the boundary, the boundary value is taken as the final parameter value. All parameter adjustment records are stored in the process database with timestamps, including the parameters before adjustment, the parameters after adjustment, the rate of change, the triggering reason, the disturbance type, the quality deviation, and other information. The finally generated real-time process parameters are sent to the machine tool CNC system via the OPCUA protocol. After receiving the parameters, the system sends back a confirmation signal. If no confirmation signal is received within 3 seconds, the parameters are resent. If the resentment fails three times in a row, a communication fault alarm is triggered, and the current parameters are kept unchanged to ensure the safety and stability of the machining process.

[0088] The minimum hardware configuration requirements for implementing this method are as follows: the industrial control computer should have a quad-core or higher processor with a main frequency of not less than 2.5GHz, a running memory of not less than 8GB, a storage space of not less than 256GB, and be equipped with an industrial Ethernet interface; the sensor measurement accuracy should meet the detection requirements, and the data acquisition delay should not exceed 100ms; the machine tool CNC system should support external parameter writing and real-time data reading functions, and be equipped with corresponding communication authorization.

[0089] The system determines whether the current processing quality indicators have reached the preset quality thresholds. If so, it maintains the real-time process parameters; otherwise, it performs proportional feedback optimization based on threshold deviation on the dynamic compensation parameters, setting an upper limit for single-step adjustment and an iterative convergence threshold. The preset quality thresholds are divided into a qualified threshold and an optimal threshold. The qualified threshold is the minimum standard for part acceptance, set according to 80% of the tolerance requirements in the part drawing, with a certain process margin reserved. The optimal threshold is the ideal target value for processing quality, set according to the requirement that the process capability index Cpk ≥ 1.33. The system compares the current processing quality indicators with the preset thresholds one by one. If all indicators reach the qualified threshold, the quality is deemed satisfactory, and processing continues with the current real-time process parameters. If any indicator fails to reach the qualified threshold, the quality is deemed unsatisfactory, triggering the proportional feedback optimization process.

[0090] Proportional feedback optimization based on threshold deviation employs a proportional-integral (PI) control algorithm. It takes the quality index deviation as input and the process parameter adjustment as output. The proportional coefficient is set according to the sensitivity coefficient matrix, and the integral coefficient is set to 0.1 to 0.2 times the proportional coefficient to eliminate steady-state deviation. During feedback optimization, the upper limit of the single-step adjustment of the dynamic compensation parameters is used to avoid processing oscillations caused by excessive single adjustments. Simultaneously, an iteration convergence threshold is set. When the quality index deviation after two consecutive iterations is less than 5% of the reference deviation threshold, optimization is considered converged, iteration stops, and the final optimized dynamic compensation parameters are output. The maximum number of iterations is set to 5. If convergence is not achieved after 5 iterations, conventional feedback optimization is considered to have failed to meet expectations.

[0091] During unstable machining phases such as start / stop, tool change, and idle feed, feedback optimization and dynamic compensation adjustments are paused, and the current parameters are kept unchanged. Once machining enters a stable cutting state, the optimization process is restarted after a 3-second delay to avoid erroneous adjustments caused by interference data from non-cutting phases. In the roughing stage, efficiency is prioritized, and the deviation threshold for feedback optimization is appropriately relaxed, while the upper limit of single-step adjustment is appropriately increased. In the finishing stage, accuracy is prioritized, the deviation threshold is strictly set, and the upper limit of single-step adjustment is narrowed to ensure the quality of the machined surface.

[0092] During the processing, the working condition coupling analysis method is used in parallel to analyze the processing working condition data and the time-series quality detection data to obtain the correlation influence data. The current processing quality index is used to extract the processing abnormal features. When the feedback optimization does not meet the expectations, the dynamic compensation parameters are deeply adjusted in combination with the correlation influence data to obtain the process correction parameters.

[0093] In obtaining the data on related influences, multi-dimensional operating condition features are first extracted from the machining condition data, and key quality features are extracted from the time-series quality inspection data. Multi-dimensional operating condition features include three categories: temperature change features, vibration features, and cutting force features. Temperature change features include average temperature, temperature gradient, and temperature rise rate; vibration features include the root mean square value of vibration acceleration, dominant vibration frequency, and quantile of the vibration amplitude probability distribution; cutting force features include average cutting force, cutting force pulsation frequency, cutting force coefficient of variation, and peak cutting force. Key quality features include three categories: dimensional accuracy features, surface quality features, and process stability features. Dimensional accuracy features include average dimensional deviation, dimensional fluctuation, and deviation change rate; surface quality features include average surface roughness, surface groove spacing, and peak residual stress; process stability features include tool wear rate, spindle thermal drift rate, and cutting force fluctuation amplitude. During feature extraction, time-domain features are directly obtained through statistical calculations, frequency-domain features are obtained through fast Fourier transform spectral analysis of the time-domain signal, and time-frequency domain features are obtained through wavelet packet transform. These multi-dimensional features comprehensively characterize the operating condition and quality status.

[0094] Correlation analysis was performed on multidimensional operating condition characteristics and key quality characteristics to label hysteresis response characteristics, which reflect the time delay between changes in operating conditions and the quality response. The correlation analysis simultaneously used Pearson correlation coefficient to calculate the degree of linear correlation and Spearman's rank correlation coefficient to calculate the degree of nonlinear correlation, resulting in a correlation coefficient matrix between operating condition characteristics and quality characteristics. Feature pairs with an absolute correlation coefficient greater than 0.6 were selected as strongly correlated feature pairs. Hysteresis response characteristics were calculated using sliding cross-correlation analysis to identify the time delay between changes in each operating condition characteristic and the corresponding quality characteristic change. The hysteresis time corresponding to thermally induced deformation was 2 to 5 minutes, the hysteresis time corresponding to tool wear was 5 to 30 minutes, and the hysteresis time corresponding to cutting vibration was less than 1 second. Hysteresis response characteristics are used for feedforward compensation control, which can predict quality change trends in advance and adjust process parameters before quality deviations occur.

[0095] Processing scenarios are categorized based on perturbation type combinations. For each scenario, regression models are constructed to integrate key quality features and multidimensional working condition characteristics. Goodness of fit is calculated, and significant perturbation factors are screened. Processing scenarios are categorized into 12 types based on perturbation type combinations, including single thermal deformation, single tool wear, single vibration, single material fluctuation, combined thermal deformation and tool wear, and combined thermal deformation and vibration. For each scenario, multiple linear regression, support vector regression, or backpropagation (BP) neural network regression algorithms are used, with working condition characteristics as input and quality characteristics as output, to train scenario-based quality prediction models. Historical processing datasets are used for model training, divided into training and testing sets in a 7:3 ratio. After training, the goodness of fit of the models is calculated. Models with a goodness of fit greater than 0.8 were retained; significance analysis was performed using t-tests and F-tests, and non-significant influencing factors with a significance level greater than 0.05 were removed to obtain the set of core significant perturbation factors for each scenario.

[0096] The analysis process includes analyzing the thermal correlation between temperature change and dimensional deviation, the vibrational-physical correlation between cutting vibration and surface roughness, and the correlation between tool wear and profile accuracy degradation, generating a statistical analysis report. In the thermal correlation analysis, a scatter plot of spindle thermal drift and corresponding dimensional deviation is plotted. A linear relationship curve is fitted using the least squares method to identify the mapping relationship between temperature gradient and dimensional drift trend. The dimensional change rate per unit temperature rise, i.e., the thermal deformation sensitivity coefficient, is calculated; for example, for every 1°C increase in spindle temperature, the dimensional drift in the X direction is 1.2 μm. In the vibrational-physical correlation analysis, the relationship between the root mean square value of cutting vibration acceleration and surface roughness is analyzed, generating a power function fitting curve. The correlation coefficient between the dominant vibration frequency and the surface texture spacing is calculated through spectral analysis. The surface texture spacing is equal to the ratio of the feed rate to the dominant vibration frequency; a larger vibration amplitude indicates higher surface roughness. During the degradation correlation analysis, the correspondence between the tool flank wear width and the profile deviation was statistically analyzed. A wear-deviation scatter plot was plotted and a linear relationship was fitted. The correlation coefficient between the wear rate and the accuracy degradation rate was extracted, showing a significant positive correlation between the two; the faster the tool wear rate, the higher the accuracy degradation rate. The results of the three types of correlation analysis were standardized and presented in a format that includes influencing factors, affected quality parameters, quantitative relationships, and machining scenario adaptation suggestions. Corresponding data charts and fitting formulas were provided to form a complete statistical analysis report.

[0097] The theoretical calculation model for the surface geometric residual height is as follows:

[0098] In the formula, The theoretical geometric residual height of the surface, in units of It is used to characterize the theoretical extreme value of the surface micro profile under the combined action of feed motion and tool tip arc, and can be used as a reference upper limit for surface roughness; This is the machinability coefficient of the material, dimensionless. Materials with high plasticity tend to have larger values, while materials with high brittleness tend to have smaller values. The typical value range is 0.8 to 1.5. Feed per revolution, in units of ; The radius of the tool tip arc, in units of This formula is derived based on the geometric residual height under ideal cutting conditions and corrected by the material's machinability coefficient. In actual machining, the surface roughness Ra is 1 / 5 to 1 / 3 of the theoretical residual height. The specific ratio is determined experimentally based on material properties and cutting conditions. It can be used for the matching design of feed rate and tool tip radius when setting initial parameters.

[0099] Statistical analysis reports are transformed into process adjustment execution rules. The confidence and support of all process adjustment execution rules are calculated, and reliable process adjustment execution rules meeting preset thresholds are retained to generate a working condition coupling rule library. Each process adjustment execution rule includes four components: preconditions, adjustment parameters, adjustment range, and expected effect. Support is the proportion of the rule's frequency of occurrence in historical processing data to the total number of samples, and confidence is the proportion of samples that meet quality standards after the rule's execution to the total number of executed samples. Rules with a support greater than 0.2 and a confidence greater than 0.8 are retained as reliable rules and stored according to disturbance type and processing scenario, forming a structured working condition coupling rule library. The rule library supports online incremental updates. After each batch of parts is processed, valid adjustment cases from this processing are added to the rule library, and the support and confidence of the corresponding rules are recalculated. Simultaneously, old rules that have not been triggered for a long time and have a confidence below 0.6 are discarded to ensure the timeliness and accuracy of the rule library.

[0100] In obtaining abnormal processing characteristics, the first step is to determine the abnormal threshold range for each processing quality indicator based on a preset processing quality threshold. To address time-series fluctuations, dynamic thresholds are set to adapt to the time-varying characteristics of the processing process. The abnormal threshold range is set to 1.2 to 1.5 times the acceptable threshold; indicators exceeding this range are considered significantly abnormal. The dynamic threshold is calculated based on a 300-second sliding time window, using the formula: dynamic threshold equals the window mean plus or minus three times the window standard deviation. This adapts to the slow time-varying trend of the processing process, avoiding misjudging normal process fluctuations as abnormalities. Independent sliding windows and dynamic thresholds are set for the roughing, semi-finishing, and finishing stages to accommodate the fluctuation characteristics of different processing stages.

[0101] The system iterates through all time windows, identifying any processing quality indicator that exceeds the abnormal threshold or dynamic threshold as an abnormal segment. The step size for time window traversal is set to 1 / 10 of the window length, i.e., 30 seconds, to ensure continuous identification of abnormal segments. When identifying an abnormal segment, the system records the segment's start time, end time, abnormal indicator name, abnormal peak value, and the magnitude of the exceedance. Adjacent abnormal segments with a time interval of less than 30 seconds are merged to form a complete abnormal event segment, avoiding fragmented segmentation of abnormal events.

[0102] Anomalies were extracted from three dimensions: dimensional anomalies, surface quality anomalies, and process stability anomalies. Dimensional anomalies included subtypes such as dimensional deviations, geometric tolerance deviations, and profile deviations; surface quality anomalies included subtypes such as surface roughness deviations, excessive residual surface stress, and excessively deep work-hardened layers; and process stability anomalies included subtypes such as excessively rapid tool wear, excessive cutting force fluctuations, and excessive spindle thermal drift. Each dimension's anomaly fragment was labeled with its corresponding index value and degree of deviation, forming a set of anomaly fragments for each dimension.

[0103] Combining associated machining condition data and material property data, the anomaly types of abnormal segments are labeled. These anomaly types include thermal drift, tool degradation, vibration mark generation, and material mutation. Each of the four types corresponds to a preset feature vector template. Thermal drift anomalies are characterized by a slow, monotonically increasing dimensional deviation with rising temperature, accompanied by a synchronous increase in spindle thermal drift, with no significant abrupt change in cutting force. Tool degradation anomalies are characterized by a gradual increase in profile deviation with machining time, accompanied by a continuous increase in tool flank wear, with the wear rate exceeding the normal range. Vibration mark generation anomalies are characterized by a sudden increase in surface roughness, the presence of periodic vibration marks on the surface, accompanied by a sudden increase in the amplitude of cutting vibration acceleration, and stable vibration frequency. Material mutation anomalies are characterized by sudden fluctuations in cutting force, along with abrupt changes in dimensional deviation and surface quality, corresponding to abrupt changes in workpiece material hardness or residual stress. By calculating the similarity between the feature vector of the abnormal segment and the template vectors of each type, the type with the highest similarity is selected as the labeling type for that abnormal segment, and the similarity value is output.

[0104] Calculate the Pearson correlation coefficient among all anomalous segments, retaining those that reach a preset threshold, and combine them with their corresponding anomalous types as the final processed anomaly features. The correlation coefficient threshold is set to 0.7. Anomalous segments with correlation coefficients higher than the threshold are considered associated anomalies and grouped into the same anomalous event; those with correlation coefficients lower than the threshold are considered isolated and occasional anomalies and are discarded to eliminate random noise interference. The final output processed anomaly features are structured data, including the start and end times of the anomalous event, anomaly type, anomaly dimension, core anomaly indicators, anomaly severity, and related operating conditions.

[0105] In the process of adjusting the process correction parameters, firstly, based on the associated impact data, feature matching is used to locate the abnormal causes of processing anomalies. Then, a process adjustment mapping table is established by combining the priority adjustment parameters and strategy direction. During the anomaly cause matching process, the extracted processing anomaly feature vectors are compared with the rule prerequisite vectors in the operating condition coupling rule base using cosine similarity calculation. The cosine similarity formula is:

[0106] In the formula, This is the current abnormal feature vector. This is the precondition vector for each rule in the rule base. The three rules with the highest similarity are selected as candidate triggers, sorted by similarity from highest to lowest. The rule with the highest similarity is the primary trigger, and the rest are secondary triggers. If the similarity of all rules is below 0.6, the match is considered a failure, triggering a manual diagnostic process. Simultaneously, a general conservative adjustment strategy is used to fine-tune parameters to prevent the aberration from escalating.

[0107] Conflicts in process parameter adjustments are identified and coordinated using preset constraint priorities, which are executed in the order of hardware constraints over process specification constraints over efficiency optimization. A conflict is identified when adjustment strategies corresponding to different causes result in opposite adjustment directions for the same parameter, or when the cumulative adjustment magnitude exceeds the parameter constraint boundary. The coordination process first verifies hardware constraints: all adjustments must not exceed the machine tool's rated parameter range, the tool's limit cutting parameters, or the equipment's safe operating threshold; adjustments violating hardware constraints are directly eliminated. Secondly, process specification constraints are verified: adjustments must ensure that the part's machining accuracy and surface quality meet process requirements; adjustments violating specifications are corrected. Finally, while satisfying the first two constraints, efficiency optimization objectives are considered, and adjustment schemes with minimal impact on machining efficiency are prioritized.

[0108] The parameter adjustment sequence is planned from fastest to slowest response speed. For sudden anomalies, the feed rate is quickly corrected first, followed by fine-tuning of the spindle speed and depth of cut. For sudden anomalies such as sudden chatter or sudden changes in cutting force, a hierarchical adjustment logic of suppression followed by optimization is adopted. First, the feed rate is quickly reduced to 60% to 70% of the current value to reduce the cutting force amplitude, rapidly attenuate vibration, and prevent the anomaly from escalating and causing part scrap. After the vibration is suppressed and the cutting state is stable, the spindle speed is fine-tuned to avoid the resonance range. Finally, the depth of cut is adjusted to optimize the cutting state and gradually restore machining efficiency. For slowly changing anomalies such as thermal deformation and tool wear, a gradual adjustment strategy is adopted, adjusting parameters in multiple small increments to avoid causing new quality fluctuations due to excessive adjustments in a single step.

[0109] By combining the correlation impact data, the tool compensation correction amount, cutting parameter correction index, and cooling parameter correction value are calculated as the process correction range, and the dynamic compensation parameters are adjusted to obtain the process correction parameters.

[0110] The tool compensation correction amount is calculated using the following formula:

[0111] In the formula, This is the tool compensation correction amount, in units of ; The tool wear sensitivity coefficient is dimensionless and represents the machining dimension deviation corresponding to a unit amount of tool wear. It is obtained by calibration through historical data and has a value range of 0.8 to 1.2. The width of the tool's flank wear, in units of ; This is the thermo-coupling influence factor, in units of This represents the correction factor for the tool wear compensation amount due to temperature changes. It takes into account the thermal deformation coupling effect between the tool and the workpiece at high temperatures, and its value ranges from 0.001 to 0.005℃. -1 ; This represents the deviation between the cutting zone temperature and the reference temperature, in units of... The formula's dimensionality is consistent; the dimensionless coefficient multiplied by the length dimension still yields a length dimension, matching the unit of the compensation quantity.

[0112] The cutting parameter correction index is calculated using the following formula: The cutting parameter correction index for a single working condition factor is:

[0113] In the formula, For the first The cutting parameter correction index corresponding to the working condition factor is dimensionless. This is a coefficient representing the severity of processing abnormalities. It is dimensionless and ranges from 0.5 to 1.5, with a larger value for more severe abnormalities. For the first The influence coefficient of the operating condition factor, its dimension is the first... The reciprocal of the dimension of the deviation value of the operating condition factor; For the first The quantitative deviation value of the operating condition factor. The dimensional verification of this formula is consistent. After multiplying the influence coefficient and the deviation value, the dimensions cancel each other out, resulting in a dimensionless correction index.

[0114] When multiple operating conditions act in combination, the total cutting parameter correction index is the weighted sum of the individual correction indices:

[0115] In the formula, For the first The influence weight of the operating condition factor, and satisfying The weights are determined based on the contribution of each working condition factor to the anomaly. The total cutting parameter correction index is limited to a range of 0.5 to 1.5. When the value exceeds the range, it is truncated and converted into the corresponding process parameter adjustment ratio according to the preset mapping relationship. When the correction index is greater than 1, it means to increase the parameter, and when it is less than 1, it means to decrease the parameter. For example, when the correction index is 0.8, the corresponding feed rate is adjusted to 80% of the current value.

[0116] The cooling parameter correction value is calculated based on the results of thermal correlation analysis, and the formula for adjusting the cutting fluid flow rate is as follows: ,in This is the flow rate temperature correction factor, in L / (min) (℃), with values ​​ranging from 0.2 to 0.5 L / (min) ℃); the formula for adjusting the cutting fluid temperature is: ,in This is a dimensionless fluid temperature correction factor, ranging from 0.3 to 0.6. By adjusting the cutting fluid flow rate and temperature, the temperature in the cutting zone is controlled within the optimal range, reducing the impact of thermal deformation.

[0117] If a multi-factor equilibrium composite scenario exists without a dominant disturbance, meaning the contribution of each disturbance factor is less than 60% and the difference is no more than 10%, a multi-parameter equilibrium adjustment strategy is adopted. The adjustment range is allocated according to the weight ratio of each disturbance factor, and the corresponding process parameters are adjusted synchronously to avoid excessive adjustment of a single parameter. If the quality index still fails to meet the standard after three consecutive tests following the deep adjustment, the deep adjustment is deemed to have failed, triggering a graded handling mechanism: for mild failure, the processing efficiency is reduced by 20%, and processing continues with conservative parameters; for moderate failure, processing is suspended, the tool is replaced, and a trial cut is performed; for severe failure, processing is stopped, a manual intervention alarm is triggered, and process personnel conduct on-site investigation and handling.

[0118] The final generated process correction parameters are sent to the machine tool for execution after constraint verification. At the same time, the parameter changes, quality response and processing results of the entire depth adjustment process are recorded and added to the working condition coupling rule base.

[0119] In summary, this embodiment provides a method for optimizing the machining process parameters of wind tunnel model parts. By capturing multi-dimensional dynamic information from the machining site in real time, it can identify instantaneous changes in material properties, equipment status, and environmental conditions, providing comprehensive data support for process adjustments. By establishing a multi-level response mechanism, it can both quickly mitigate conventional disturbances through real-time calculation of dynamic compensation parameters and locate the root cause of anomalies and implement targeted corrections when conventional adjustments fail through in-depth analysis of the correlation between operating conditions and quality.

[0120] By introducing a smooth transition and conflict coordination mechanism, continuous parameter optimization is achieved while ensuring processing safety, thus solving the chatter and quality step problems that are easily caused by traditional adjustment methods.

[0121] By transforming empirical process rules into quantifiable correlation models, the reliance on human experience is reduced, and the yield and consistency of high-precision component processing are improved, providing a reliable intelligent solution for the high-quality manufacturing of complex and precision parts.

[0122] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0123] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the machining process parameters of wind tunnel model parts, characterized in that, include: Collect processing condition data, material property data, and equipment status data, and formulate initial process parameters based on the material property data and wind tunnel model design specifications; Real-time quality inspection data of the part processing process is collected, and the processing quality index is obtained by analyzing the time-series quality inspection data using the multi-modal quality state hierarchical analysis method, and the corresponding processing quality deviation index is calculated. Dynamic compensation parameters are calculated based on the type of disturbance in the processing conditions and the deviation index of processing quality. The initial process parameters are then adjusted based on the dynamic compensation parameters to obtain the real-time process parameters. Determine whether the current processing quality index has reached the preset quality threshold. If so, maintain the real-time process parameters; otherwise, perform proportional feedback optimization based on threshold deviation on the dynamic compensation parameters, and set the upper limit of single-step adjustment and the iterative convergence threshold. During the processing, the processing condition coupling analysis method is used in parallel to analyze the processing condition data and the time-series quality detection data to obtain the associated influence data, and the current processing quality index is used to extract the processing abnormal features. When the feedback optimization fails to meet expectations, the dynamic compensation parameters are adjusted in depth based on the associated impact data to obtain process correction parameters.

2. The method for optimizing the processing parameters of a wind tunnel model part according to claim 1, characterized in that, The machining condition data includes temperature change data, cutting vibration data, and cutting force pulsation data; the material property data includes hardness data, elastic modulus data, thermal expansion coefficient data, and residual stress distribution data; the equipment status data includes spindle rotation accuracy data, tool wear status data, guideway geometric error data, and cooling and lubrication status data.

3. The method for optimizing the machining process parameters of a wind tunnel model part according to claim 1, characterized in that, The process of determining the initial process parameters includes: Based on the wind tunnel model design specifications, the differentiated process requirements of force measurement models, pressure measurement models, and special test models are distinguished and converted into accuracy level indicators and surface quality level indicators. Based on the material property data, cutting parameter constraint boundaries are formulated. By combining the matching characteristics of tool material and workpiece material, suitable tool types are selected, and the cutting depth allocation for roughing, semi-finishing and finishing is determined according to the geometric features and accuracy level of the part. The reference cutting speed is determined based on the spindle power range and the tool's limit cutting speed, and the feed strategy is selected according to the accuracy grade index and the surface quality grade index. The tool path is planned based on the equipment stiffness characteristics and geometric tolerance requirements, and the tool wear warning threshold is set based on the material machinability index.

4. The method for optimizing the machining process parameters of a wind tunnel model part according to claim 1, characterized in that, The process of analyzing and obtaining the aforementioned processing quality indicators includes: The part size detection signal, surface integrity data and process parameters are collected and preprocessed to serve as the time-series quality detection data. Dimensional deviation, geometric tolerance, and profile are extracted from the part size detection signal as dimensional accuracy indicators, and surface roughness, surface residual stress, and work hardening layer depth are extracted from the surface integrity data as surface quality indicators. Tool wear rate, cutting force fluctuation amplitude, and spindle thermal drift are extracted from the process parameters as process stability indicators. The trend characteristics and working condition correlation indicators of the dimensional accuracy index, the surface quality index and the process stability index are extracted to form the multi-dimensional and hierarchical processing quality index; The processing quality deviation index is the absolute deviation value and relative deviation rate between the measured value of each processing quality index and the corresponding preset quality threshold.

5. The method for optimizing the machining process parameters of a wind tunnel model part according to claim 1, characterized in that, The process of calculating the dynamic compensation parameters includes: Based on the processing condition data, material property data, and equipment status data, the disturbance types are classified according to thermal deformation disturbance, tool wear disturbance, cutting vibration disturbance, and material property fluctuation disturbance, and the disturbance level is quantified according to the degree of influence to obtain the processing condition disturbance data; Determine the preset threshold of the processing quality deviation index, calculate the deviation value, classify the quality deviation level according to the degree of deviation, and perform time synchronization and correlation mapping processing with the working condition disturbance data; Establish a disturbance type-process parameter mapping table, identify the dominant disturbance type in the compound disturbance, and determine the strategy strength according to the disturbance level; The parameter adjustment priority is determined based on the strength of the strategy, the adjustment range is determined based on the quality deviation, and the single-step adjustment amount is limited. The dynamic compensation parameters are obtained by supplementing specific compensation parameters for the disturbance type of the processing condition.

6. The method for optimizing the machining process parameters of a wind tunnel model part according to claim 1, characterized in that, The process of adjusting and obtaining the real-time process parameters includes: Based on the equipment status data and historical process adjustment records, the impact of the dynamic compensation parameters on processing quality is determined, and a sensitivity coefficient matrix is ​​formed. Prioritize adjustments based on the directness of their impact on processing quality and the speed of response. A smooth transition mechanism is achieved by setting transition times and parameter activation conditions for continuous process parameters, preventing sudden parameter changes from causing processing chatter or surface quality abrupt changes. The current dynamic compensation parameters are superimposed on the initial process parameters according to the adjustment priority and the sensitivity coefficient matrix, and the parameter change rate and the reason for the change before and after superposition are recorded to obtain the real-time process parameters.

7. The method for optimizing the machining process parameters of a wind tunnel model part according to claim 1, characterized in that, The process of obtaining the aforementioned correlation impact data includes: Multidimensional working condition features are extracted from the processing condition data, and key quality features are extracted from the time-series quality inspection data. A correlation analysis is performed between the multidimensional operating condition characteristics and the key quality characteristics to label the hysteresis response characteristics, wherein the hysteresis response characteristics reflect the time delay between the change in operating condition and the quality response. Based on the combination of disturbance types, processing scenarios are divided. Under different processing scenarios, regression models of the key quality features and the multidimensional working condition features are constructed, the goodness of fit is calculated, and significant disturbance factors are screened. Analyze the thermal correlation between temperature change and dimensional deviation, the vibrational-physical correlation between cutting vibration and surface roughness, and the correlation between tool wear and profile accuracy degradation, and generate a statistical analysis report; The statistical analysis report is converted into process adjustment execution rules. The confidence and support of all process adjustment execution rules are calculated, and reliable process adjustment execution rules that meet the preset threshold are retained to generate a working condition coupling rule library.

8. The method for optimizing the processing parameters of a wind tunnel model part according to claim 7, characterized in that, The process of generating the statistical analysis report includes: Plot a scatter plot of the spindle thermal drift and the corresponding dimensional deviation, identify the mapping relationship between the temperature gradient and the dimensional drift trend, and calculate the dimensional change rate corresponding to a unit temperature rise to obtain the thermal correlation. The relationship between the root mean square value of cutting vibration acceleration and surface roughness is analyzed, a fitting relationship curve is generated, and the correlation coefficient between the dominant vibration frequency and the surface texture spacing is calculated to obtain the vibration-material correlation. The correlation between the wear width of the tool flank and the profile deviation is statistically analyzed, and the correlation between the wear rate and the accuracy degradation rate is extracted to obtain the degradation correlation. The statistical analysis report is obtained by standardizing the description of the thermal correlation, the vibrational quality correlation, and the degradation correlation according to the format of influencing factors-affected quality parameters-quantified relationship-processing scenario adaptation suggestions.

9. The method for optimizing the machining process parameters of a wind tunnel model part according to claim 1, characterized in that, The process of obtaining the aforementioned processing anomaly features includes: For the preset processing quality threshold, the abnormal threshold range of each processing quality indicator is determined, and for the time-series fluctuation, dynamic thresholds are set to adapt to the time-varying characteristics of the processing process. Traverse all time windows, and when any processing quality indicator exceeds the abnormal threshold range or the dynamic threshold, it is determined to be an abnormal segment; The corresponding abnormal segments were extracted from the dimensions of size abnormality, surface quality abnormality, and process stability abnormality, respectively. By combining the associated machining condition data and material property data, the anomaly type of the abnormal segment is marked. The anomaly types include thermal drift type, tool degradation type, vibration mark generation type, and material mutation type. Calculate the Pearson correlation coefficient among all abnormal segments, retain the abnormal segments that reach the preset coefficient threshold, and combine them with the corresponding abnormal type as the processing abnormal feature.

10. The method for optimizing the machining process parameters of a wind tunnel model part according to claim 1, characterized in that, The process of adjusting and obtaining the process correction parameters includes: Based on the associated impact data, the abnormal causes of the processing abnormality are located by feature matching, and a process adjustment mapping table is established by combining the priority adjustment parameters and strategy direction. Conflicts in process parameter adjustments are identified and coordinated through preset constraint priorities, which are executed in the order of hardware constraints taking precedence over process specification constraints taking precedence over efficiency optimization. The parameter adjustment sequence is planned from fast to slow response speed. Sudden anomalies are first executed to quickly correct the feed rate, and then the spindle speed and depth of cut are fine-tuned. The tool compensation correction amount, cutting parameter correction index, and cooling parameter correction value are calculated based on the aforementioned correlation influence data as the process correction range. The dynamic compensation parameters are then adjusted to obtain the process correction parameters.

Images