A bearing grinding processing parameter adaptive control method and system

Abstract

The application provides a bearing grinding process parameter adaptive control method and system, and relates to the field of industrial system control; the method acquires the set processing parameters and the in-process tolerance range of inner ring grinding, outer ring grinding, raceway grinding and superfinishing, and establishes the correlation constraint relation of the whole process parameters. Under the condition that the actual parameters of each process do not exceed the tolerance range, the actual processing parameters are collected in real time and the normalized deviation is calculated, the relation is substituted to solve the whole process comprehensive performance deviation; when it exceeds the allowable performance deviation threshold, the single deviation contribution value of each process is calculated and sorted from large to small, and each process is step by step adjusted in the direction of decreasing deviation in turn until the comprehensive performance deviation is lower than the threshold, so that the subsequent processing is controlled by the finally adjusted parameters. The application realizes the adaptive optimization control of the whole process of bearing grinding.

Description

Technical Field

[0001] This invention relates to the field of industrial system control, specifically to an adaptive control method and system for bearing grinding parameters. Background Technology

[0002] In bearing grinding, existing production control methods typically set independent processing parameters and tolerance ranges for each process, such as inner ring grinding, outer ring grinding, raceway grinding, and ultra-precision machining, and perform discrete control based on the dimensions, geometric tolerances, or surface quality indicators of each process.

[0003] In actual production, each process is executed sequentially according to predetermined parameters, and the overall performance of the final bearing largely depends on offline sampling inspection after processing. The currently commonly used control strategies only focus on whether the parameters of each process meet their own tolerance requirements, without fully considering the complex interaction between processing deviations and the nonlinear coupling effect on the overall performance of the finished bearing. In particular, when the actual parameters of multiple processes fluctuate within their respective tolerance ranges, existing methods are unable to quantify the coupling relationship between these deviations and their nonlinear superposition effect on the overall performance indicators such as bearing dynamic vibration, rotational accuracy, and fatigue life. Furthermore, when deviations in the final bearing performance are detected, there is a lack of a control mechanism that can distinguish the primary and secondary responsibilities of the deviation factors in each process and make hierarchical adjustments according to reasonable priorities. This often leads to directional conflicts or repeated oscillations during the adjustment process, making it difficult to efficiently restore the overall performance deviation of the entire process to the allowable level while keeping the parameters of each process within the tolerance range. Summary of the Invention

[0004] To achieve the above objectives, this invention proposes an adaptive control method and system for bearing grinding parameters.

[0005] On the one hand, this invention proposes an adaptive control method for bearing grinding parameters, including:

[0006] S1. Obtain the set processing parameters and corresponding tolerance ranges of each process in the bearing grinding process, wherein the processes include inner ring grinding, outer ring grinding, raceway grinding and ultra-precision machining in sequence;

[0007] S2. Establish the correlation constraint relationship between the entire process processing parameters and bearing performance. The correlation constraint relationship is used to characterize the comprehensive influence of the main effect of a single process, the interactive coupling effect between processes, and the nonlinear coupling effect of the entire process on the overall performance of the bearing.

[0008] S3. Under the condition that each process follows its own set processing parameters and the actual parameters do not exceed the tolerance range of each process, collect the actual processing parameters of each process in real time and calculate their normalized deviation, and substitute them into the correlation constraint relationship to calculate the current overall performance offset of the whole process.

[0009] S4. Determine whether the overall performance deviation of the entire process exceeds the preset allowable performance deviation threshold: If it does not exceed the threshold, maintain the current processing parameters of each process and continue processing; if it exceeds the threshold, proceed to step S5.

[0010] S5. Calculate the individual deviation contribution value of each process, where the individual deviation contribution value is the weighted deviation term corresponding to the process in the associated constraint relationship; sort the individual deviation contribution values ​​of each process from largest to smallest, and adjust each process in sequence according to the sorting: adjust the set processing parameters of the process in the direction of decreasing deviation by a preset step size, and recalculate the overall performance offset of the entire process; if it is still higher than the allowable performance offset threshold, continue to adjust the same process in the same direction until its set processing parameters reach the boundary of the tolerance range within the process; if it is still not lower than the threshold after reaching the boundary, proceed to the next process according to the sorting; repeat the above process until the overall performance offset of the entire process is lower than the allowable performance offset threshold; then control the subsequent bearing grinding process with the finally adjusted set processing parameters of each process.

[0011] As a further technical solution, the processing parameters set for each process in step S1 include: grinding depth of the inner ring grinding process, grinding depth of the outer ring grinding process, raceway roundness deviation of the raceway grinding process, and surface roughness of the ultra-precision machining process; the units for the inner ring grinding depth and outer ring grinding depth are millimeters, the unit for raceway roundness deviation is micrometers, and the unit for surface roughness is micrometers; the tolerance range within each process is predetermined by the bearing design drawings or process documents, and for each process, its tolerance range half-width is determined by half the difference between the upper and lower tolerance limits. The machining parameters are taken as the median of the tolerance range or the nominal design value; the above data are read and stored in the controller through the CNC system parameter table or the manufacturing execution system interface; the grinding depth directly determines the amount of material removed, which in turn affects the dimensional accuracy and form and position tolerance of the bearing; the raceway roundness deviation directly affects the contact uniformity between the rolling elements and the raceway; the surface roughness determines the friction and wear characteristics and vibration and noise level of the bearing; these four parameters constitute the core process control variables in the entire bearing grinding process. By collecting and adjusting them, the minimum set of variables can cover most of the factors affecting the comprehensive performance of the bearing.

[0012] As a further technical solution, the correlation constraint relationship in step S2 is specifically as follows: Let the actual machining parameters of each process, including inner ring grinding, outer ring grinding, raceway grinding, and ultra-precision machining, be denoted as p1, p2, p3, and p4, respectively, and the corresponding set values ​​be... The tolerance range half-width within each process is t1, t2, t3, t4, and the normalized deviation is... Where i = 1 to 4; then the association constraint relationship is:

[0013]

[0014] Where i and j represent process indices, α i β is the single-process main effect coefficient determined by orthogonal experiments combined with analysis of variance, based on the weighted proportion of the influence of each process parameter on the overall bearing performance. ij λ is the interaction coupling coefficient between each pair of processes, γ is the full coupling coefficient of the four processes, and λ is the coupling coefficient between each pair of processes. i The nonlinear saturation factor; the square term δ in this correlation constraint relationship. i 2 Characterizing the nonlinear degradation of performance due to deviations in a single process, the cross term δ i δ j The product term represents the synergistic or antagonistic effect between two processes, and the supercoupling effect that occurs when all four processes deviate simultaneously, which is beyond linear superposition. The exponential structure after introducing absolute values ​​can simulate the saturation phenomenon in actual physical processes where performance deteriorates sharply when the deviation approaches the tolerance boundary. To ensure the computability of this relationship in engineering, all coefficients are obtained through standardized offline test calibration. After calibration, they are solidified into executable code in the controller. During real-time calculation, only the currently collected normalized deviation needs to be substituted to output the overall performance offset of the entire process.

[0015] As a further technical solution, β ij The calculation process is as follows: Two extreme value combination experiments are set for each pair of processes i and j, while keeping the deviation of the remaining processes at 0; in the first experiment, δ... i and δ j If both are set to +1 or both are set to -1, the overall performance offset of the entire process under this combination is measured and recorded as ΔP. ij同 In the second type of experiment, δ i and δ j Let the tolerance boundary values ​​have opposite signs, i.e., one is +1 and the other is -1, and the measured offset is denoted as ΔP. ij异 ;but The calculation process for γ is as follows: The deviations δi of the four processes are simultaneously set to the same sign as their respective tolerance boundary values, either all +1 or all -1. The overall performance offset of the entire process at this time is measured and recorded as ΔP. ​​This is then calculated using the formula... The solution yields, where This is the sum of the main effect terms of a single process. It is the sum of all pairwise interaction terms; ΔP involved in the above calculation process ij同 ΔP ij异Both β and ΔP are comprehensive performance offsets obtained through actual measurement on a bearing performance comprehensive testing bench. This bearing performance comprehensive testing bench can apply standard loads to the bearing and measure its dynamic vibration acceleration, rotational accuracy, temperature rise, and fatigue life conversion factor. The weighted geometric average of these four indicators is then normalized according to industry standards, and the result is supplemented to obtain the dimensionless offset. ij The denominator contains α i With α j Its physical significance lies in normalizing the interaction effect to the order of magnitude of the single-process effect, avoiding the influence of α. i Too small or too large a value can lead to unreasonable amplification of the interaction coefficient; the calculation of γ removes the residual full coupling effect after stripping away all low-order terms, ensuring the consistency of the correlation constraint relationship in predicting different combinations of deviations.

[0016] As a further technical solution, the calculation process of the overall performance offset in step S3 is as follows: During the sequential processing of each process, the actual processing parameters of each process are collected, so that the collection timestamps of the four processes in this processing correspond one-to-one with the processing time of the corresponding process; all actual processing parameters under the same timestamp are substituted into the correlation constraint relationship to solve for the overall performance offset of the entire process at that time. This overall performance offset is a real-time scalar value characterizing the degree of deviation of the bearing's overall performance; in specific implementation, a laser displacement sensor is installed on the inner ring grinding machine to measure the change in the inner diameter of the workpiece before and after grinding in real time. The grinding depth p1 is obtained through a synchronous data acquisition card; the outer ring grinding depth p2 is measured in the same way; the online measurement method for raceway roundness deviation p3 is as follows: after the raceway grinding process is completed but before the workpiece leaves the machine, a roundness measuring head installed at the same station performs a 360-degree scan to obtain the raceway profile, and the root mean square of a specific harmonic amplitude is extracted by Fourier transform as the roundness deviation. The measurement result is linked to the end of grinding; the online measurement of surface roughness p4 uses a contact-type automatic roughness measurement unit to obtain the roughness value after ultra-precision machining; after each process is completed, the actual measured value p is immediately recorded. i The process, along with the completion timestamp of the process, is packaged and sent to the central controller. After receiving the timestamps and numerical pairs of the four processes, the central controller determines whether the time difference is less than a preset window. If so, it is considered that the four process parameters are for the same workpiece. The normalization formula is substituted into the parameters to calculate the normalization deviation, and then the correlation constraint relationship is substituted into the formula to solve for ΔP. The value of ΔP is updated once after each workpiece is processed, realizing real-time monitoring of the performance of the entire process piece by piece without waiting for offline sampling inspection results.

[0017] As a further technical solution, the judgment logic for the overall performance deviation in step S4 is as follows: the preset allowable performance deviation threshold is 80% of the bearing's factory comprehensive performance qualification threshold; after each calculation of the overall performance deviation, a threshold comparison is immediately performed; only when the value of the overall performance deviation is strictly greater than the allowable performance deviation threshold is it determined to be exceeded and the parameter adjustment process in step S5 is triggered; if it is equal to or less than the threshold, it is determined to be not exceeded, and the current processing parameters remain unchanged; the calibration method for the bearing's factory comprehensive performance qualification threshold is as follows: 200 bearings are randomly selected from the batch of finished products on the production line, and their actual comprehensive performance deviation ΔP is measured. At the same time, the qualification of each bearing is determined according to the national standard; the ΔP measurements of all qualified bearings are arranged in ascending order, and the 95th percentile is taken as the qualification threshold; setting the threshold to 80% of the qualification threshold is equivalent to retaining a safety margin, so that ΔP is triggered for adjustment before reaching the qualification threshold, thereby effectively preventing the occurrence of defective products; the comparison is performed immediately after each workpiece is completed, so that the adjustment decision has the shortest feedback cycle and avoids the generation of batch scrap.

[0018] As a further technical solution, the calculation and sorting rules for the individual deviation contribution values ​​of each process in step S5 are as follows: extract the main effect term of the corresponding process and the sum of all pairwise interactive coupling terms containing that process from the correlation constraint relationship, and use it as the individual deviation contribution value of that process; for process i, its individual deviation contribution value C i =α i δ i 2 +Σ j≠i β ij ×δ i ×δ j The individual deviation contribution values ​​of the four processes—inner ring grinding, outer ring grinding, raceway grinding, and ultra-precision machining—are arranged in descending order to generate a process adjustment priority sequence. The physical meaning of this calculation method lies in the fact that each interactive coupling term β... ij ×δ i ×δ j The additional performance degradation caused by the joint deviation of process i and process j is attributed entirely to the contribution value of process i, which reflects the comprehensive responsibility of process i for the overall performance deterioration under the current deviation state. The larger the contribution value of a process, the more serious the negative impact of its current deviation and its coupling effect with other processes on the overall performance, so it should be adjusted first. After sorting, the processes are adjusted in sequence to avoid the control direction conflict or oscillation convergence problem caused by the coupling effect when adjusting multiple processes at the same time.

[0019] As a further technical solution, the preset step size in step S5 is 5% to 10% of the half-width of the tolerance range within the process. The specific ratio is selected according to the stability and controllability of the process parameters: for grinding depth controlled by a high-rigidity servo system, the stability is high, so a step size of 5% is used to avoid over-adjustment; for roundness deviation or roughness that is greatly affected by grinding wheel wear, the fluctuation is large, so a step size of 10% is used to speed up the response. The specific value of the step size is dynamically adjusted before each adjustment based on the difference between the current ΔP and the threshold: when the difference is large, the step size moves closer to 10%, and when the difference is close to zero, the step size converges to 5%. The adjustment direction is always kept in the direction of decreasing deviation, that is, if the current normalized deviation is positive, the set processing parameter is increased, and if it is negative, the set processing parameter is decreased. If ΔP does not improve after two consecutive adjustments to the same process but increases or remains unchanged, it is determined that the contribution value calculation of the process is interfered with by other processes. At this time, the process is skipped and the next process is started, and its priority weight is reduced in subsequent rounds.

[0020] On the other hand, the present invention proposes an adaptive control system for bearing grinding parameters, comprising:

[0021] The data acquisition module is configured to acquire the set processing parameters for each process in bearing grinding and their corresponding tolerance ranges within the process.

[0022] The relational construction module is configured to establish the correlation constraint relationship between the entire process processing parameters and the bearing performance. This correlation constraint relationship is used to characterize the comprehensive impact of the main effect of a single process, the interactive coupling effect between processes, and the nonlinear coupling effect of the entire process on the overall performance of the bearing.

[0023] The real-time acquisition and calculation module is configured to acquire the actual processing parameters of each process in real time and calculate their normalized deviation under the condition that each process follows its own set processing parameters and the actual parameters do not exceed the tolerance range of each process. The normalized deviation is then substituted into the associated constraint relationship to solve for the current overall performance offset of the entire process.

[0024] The comparison and judgment module is configured to determine whether the overall performance offset of the entire process exceeds a preset allowable performance offset threshold: if it does not exceed the threshold, the current processing parameters of each process are maintained and processing continues; if it exceeds the threshold, the adjustment execution module is triggered.

[0025] The execution module is adjusted to calculate the individual deviation contribution value of each process. The individual deviation contribution value is the weighted deviation term corresponding to the process in the associated constraint relationship. The individual deviation contribution values ​​of each process are sorted from largest to smallest, and adjustments are performed on each process in sequence according to the sorting: the set processing parameters of the process are adjusted by a preset step size in the direction of decreasing deviation, and the overall performance offset of the entire process is recalculated. If it is still higher than the allowable performance offset threshold, the same process is adjusted in the same direction until its set processing parameters reach the boundary of the tolerance range within the process. If it is still not lower than the threshold after reaching the boundary, the process is moved to the next process according to the sorting. The above process is repeated until the overall performance offset of the entire process is lower than the allowable performance offset threshold. Then, the subsequent bearing grinding process is controlled by the finally adjusted set processing parameters of each process.

[0026] This invention provides an adaptive control method and system for bearing grinding parameters. It offers the following advantages:

[0027] 1. This invention solves the problem of the inability to quantify the nonlinear impact of the complex coupling effect between deviations of multiple processes on the overall bearing performance by constructing an association constraint relationship containing parameters of the entire process, thereby improving the real-time online evaluation accuracy of the overall performance of the entire process.

[0028] 2. This invention solves the problem of difficulty in distinguishing the primary and secondary responsibilities of each deviation factor and the tendency for adjustments to cause directional conflicts when multiple process parameters fluctuate simultaneously by extracting the individual deviation contribution value of each process from the correlation constraint relationship and generating a priority sequence, thereby improving the pertinence and convergence speed of parameter adjustment.

[0029] 3. This invention solves the problem of lack of systematic coordination strategy when the overall performance still fails to meet the standard after a single process is adjusted to the tolerance boundary by gradually adjusting a single process according to priority and then moving to the next process after reaching the boundary. This achieves adaptive optimization control in each process. Attached Figure Description

[0030] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0031] Figure 1 This is a schematic diagram of the process described in this invention. Detailed Implementation

[0032] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] like Figure 1 As shown, this invention illustrates the adaptive control system and control method for bearing grinding parameters proposed in this invention, using a running bearing grinding production line as an example. This embodiment uses a fully automated bearing grinding production line of a precision bearing manufacturing company as an application scenario. The production line is sequentially equipped with an inner ring grinding machine, an outer ring grinding machine, a raceway grinding machine, and an ultra-precision machining machine. Each machine tool is equipped with an independent CNC system and connected to a central industrial controller via a fieldbus. The system adopts a programmable automation controller based on the PLCopen standard, integrating a data storage unit, a relational calculation unit, and a parameter adjustment logic unit. All machine tools are equipped with online measurement devices for real-time acquisition of actual processing parameters for each process.

[0034] Before the production line is officially put into operation, operators first need to complete the system initialization configuration. Operators then obtain the set machining parameters and corresponding tolerance ranges for the four processes—inner ring grinding, outer ring grinding, raceway grinding, and ultra-precision machining—from the bearing design drawings and process documents. Taking the inner ring grinding process as an example, the design drawings specify a nominal grinding depth of 0.15 mm and a tolerance range of 0.14 mm to 0.16 mm. Therefore, operators set the machining parameter to 0.15 mm and set the tolerance range half-width to 0. The system inputs data in 0.01 mm increments; the outer ring grinding process has a set machining parameter of 0.18 mm and a tolerance range of 0.01 mm half-width; the raceway grinding process has a set machining parameter of 1.5 μm raceway roundness deviation, a tolerance range of 1.0 μm to 2.0 μm, and a half-width of 0.5 μm; the ultra-precision machining process has a set machining parameter of 0.04 μm surface roughness, a tolerance range of 0.03 μm to 0.07 μm, and a half-width of 0.02 μm; the above data is automatically read and stored in the system's data acquisition module through the system interface.

[0035] After initial parameter configuration, the system enters the relational construction phase. The relational construction module is responsible for establishing the correlation constraint relationship between the entire process processing parameters and bearing performance. In this embodiment, the system calibrates all coefficients in the relational relationship through pre-conducted offline orthogonal experiments. Technicians design nine sets of experiments according to the L9 orthogonal array. In each set of experiments, the processing parameters of the four processes are set to different combinations, including cases where each parameter deviates from the set value individually, deviates from each other simultaneously, and deviates from all four parameters simultaneously. Each set of experiments processes 30 bearings, and then uses a bearing performance comprehensive testing bench to perform comprehensive performance testing on each bearing. This testing bench can simulate the actual working conditions of the bearing, apply the rated load, and simultaneously measure the dynamic vibration acceleration, rotational accuracy, temperature rise, and fatigue life conversion factor of the bearing. The testing bench normalizes the measured values ​​of these four indicators according to industry standards, and then calculates a dimensionless value between 0 and 1 by taking the complement after weighted geometric mean. This value is called the overall process comprehensive performance offset. The closer the overall process comprehensive performance offset is to 0, the closer the bearing's comprehensive performance is to the ideal state. The closer it is to 1, the more severe the performance degradation.

[0036] The relationship construction module based on the experimental data above calculates the main effect coefficient of a single process, the interaction coupling coefficient between pairs of processes, the full coupling coefficient of four processes, and the nonlinear saturation factor of each process. Taking the main effect coefficient of a single process as an example, the coefficient for inner ring grinding is 0.35, indicating that when the inner ring grinding depth deviates from the set value to the boundary of the entire tolerance range, this process alone will increase the overall performance deviation of the entire process by 0.35. The main effect coefficients for outer ring grinding, raceway grinding, and ultra-finishing are 0.30, 0.25, and 0.10, respectively. Among the pairwise interaction coupling coefficients, the coupling coefficient between inner ring grinding and outer ring grinding is 1.90, indicating that when these two processes deviate in the same direction at the same time, the performance degradation effect they produce together is much greater than the simple sum of their individual effects. The full coupling coefficient of four processes is calculated to be negative, indicating that when the four processes reach the tolerance boundary at the same time, the nonlinear saturation effect slows down the performance degradation rate. After all coefficients are calculated, the relationship construction module solidifies them into executable code and stores it in the system for subsequent real-time calculation and calling.

[0037] After the production line officially starts, the real-time data acquisition and calculation module begins operation. When the first workpiece enters the inner ring grinding station, the laser displacement sensor installed at this station begins to continuously collect data. The sensor probe is aligned with the inner surface of the workpiece, recording the initial inner diameter value before grinding and measuring it again after grinding. The actual grinding depth is calculated by the difference. The measurement result is sent to the system in digital signal form via a data acquisition card, along with a timestamp indicating the completion of this process. After the inner ring grinding is completed, the workpiece automatically flows to the outer ring grinding station, which is also equipped with a laser displacement sensor. The actual value of the outer ring grinding depth is measured in the same way and packaged with the timestamp before being sent to the outer ring grinding station. The workpiece continues to the raceway grinding station. After grinding at this station, the workpiece is not removed from the machine. Immediately, a roundness measuring head installed at the same station performs a 360-degree scan. The roundness measuring head collects more than a thousand contour points every revolution. The built-in Fourier transform algorithm extracts the harmonic amplitudes of wavenumbers 2 to 15. The root mean square of these amplitudes is calculated as the actual value of the raceway roundness deviation, and a timestamp is sent to the system. Finally, the workpiece enters the ultra-precision machining station. After machining, the contact roughness automatic measurement unit automatically extends a probe to scan multiple positions on the workpiece surface. The average value is taken as the actual value of the surface roughness, and a final timestamp is sent.

[0038] After receiving the actual processing parameters for the four processes, the real-time acquisition and calculation module first checks whether the timestamps of the four processing actions on the same workpiece are all within a preset 30-second time window. If the maximum difference between the four timestamps does not exceed 30 seconds, these parameters are considered to belong to the same workpiece. If the difference exceeds 30 seconds, the system will issue an alarm to prompt the operator to check for workpiece stagnation or sensor triggering abnormalities. For the four parameters belonging to the same workpiece, the real-time acquisition and calculation module calculates the normalized deviation for each process. Taking inner ring grinding as an example, assuming the actual measured grinding depth is 0.154 mm, the set value is 0.15 mm, and the tolerance range half-width is 0.01 mm. The normalized deviation is 0.4; similarly, the actual grinding depth of the outer ring is 0.177 mm, and the normalized deviation is -0.3; the actual roundness deviation of the raceway is 1.7 μm, and the normalized deviation is 0.4; the actual surface roughness is 0.045 μm, and the normalized deviation is 0.25; the real-time acquisition and calculation module substitutes these four normalized deviations into the pre-stored relational formula in the relational formula construction module, and calculates the single-process main effect contribution, pairwise interaction contribution, and full coupling term contribution in turn, and finally outputs the overall performance offset of the current workpiece. In this example, the calculated overall performance offset is 0.854; this value is passed to the comparison and judgment module.

[0039] The comparison and judgment module pre-stores an allowable performance deviation threshold. The method for determining this allowable performance deviation threshold is as follows: During the production line debugging phase, the operator randomly selects two hundred bearings from the trial production batch and measures the overall performance deviation of each bearing throughout the entire process. At the same time, each bearing is judged to be qualified according to national standards. After sorting the deviations of all qualified bearings from smallest to largest, the 95th percentile is taken as the qualified threshold. Assuming the qualified threshold is 0.62, the allowable performance deviation threshold is set to 80% of the qualified threshold, i.e., 0.496. The comparison and judgment module compares the deviation calculated for the current workpiece, 0.854, with the threshold of 0.496. The judgment result is that it exceeds the threshold, thus triggering the adjustment execution module.

[0040] The adjustment execution module first calculates the individual deviation contribution value of each process. For the inner ring grinding process, the contribution value is equal to the single-process main effect term of the process plus all pairwise interactive coupling terms involving inner ring grinding. In specific calculation, the adjustment execution module extracts the pre-stored main effect coefficients and interaction coefficients from the relational construction module, and performs product and summation operations in combination with the current normalized deviation. The calculated contribution value is 0.169 for inner ring grinding, 0.495 for outer ring grinding, 0.207 for raceway grinding, and 0.157 for ultra-finishing. The adjustment execution module sorts these contribution values ​​from largest to smallest to obtain the adjustment priority sequence: outer ring grinding is ranked first, raceway grinding is ranked second, inner ring grinding is ranked third, and ultra-finishing is ranked last.

[0041] The adjustment execution module first adjusts the outer ring grinding process. The current normalized deviation of the outer ring grinding is -0.3, indicating that the actual grinding depth is less than the set value. Reducing the deviation brings the actual value closer to the set value, which means reducing the set machining parameters. The system adjusts according to a preset step size, which is 8% of the half-width of the tolerance range within the outer ring grinding process. The half-width is 0.01 mm, therefore the step size is 0.0008 mm. The adjustment execution module reduces the set machining parameters for the outer ring grinding from the initial 0.18 mm to 0.1792 mm. After adjustment, the adjustment execution module recalculates the predicted value of the overall process performance offset using the associated constraint relationship. The prediction result shows that the offset has decreased to 0.658, but is still higher than the threshold of 0.4. 96; Therefore, the adjustment execution module continued to adjust the same process in the same direction, reducing the set machining parameters by 0.0008 mm to 0.1784 mm again. The recalculated offset prediction value was 0.512, which was still higher than the threshold. The adjustment execution module made a third adjustment, reducing the set parameters to 0.1776 mm. The recalculated offset prediction value was 0.458, which was lower than the threshold. At this point, the adjustment of the outer ring grinding process was stopped. The adjusted set value of 0.1776 mm was still within the tolerance range of 0.17 mm to 0.19 mm in the process, so the parameter was valid. Since adjusting only the outer ring grinding process had already reduced the overall performance offset to below the allowable threshold, the adjustment execution module no longer adjusted the subsequent raceway grinding, inner ring grinding, and ultra-precision machining processes.

[0042] The adjustment execution module writes the finalized machining parameters for each process into the parameter table of the CNC system; the inner ring grinding remains at the initial 0.15 mm, the outer ring grinding is updated to 0.1776 mm, the raceway grinding remains at 1.5 micrometers, and the ultra-precision machining remains at 0.04 micrometers; subsequent bearings in the same batch are all controlled according to this set of parameters; after each bearing is machined, the system repeats the above complete process of real-time acquisition, calculation, comparison, and adjustment, thereby achieving piece-by-piece adaptive optimization of the machining parameters for each bearing.

[0043] After continuously processing one hundred bearings, the operators compiled production data. Before adjustment, six of the randomly selected one hundred bearings had a final overall performance deviation exceeding the acceptable threshold of 0.62, resulting in a failure rate of 6%. After adopting the system of this invention, when processing the same one hundred bearings, only one had a deviation exceeding the acceptable threshold, reducing the failure rate to 1%. Furthermore, because the system triggers adjustment when the deviation reaches 80% of the acceptable threshold, there has never been a situation where the deviation significantly exceeds the acceptable threshold, effectively preventing the generation of batches of defective products. The operators also observed that the system converged fastest when prioritizing the outer ring grinding process, which is consistent with the high main effect coefficient and interactive coupling coefficient of this process in the correlation constraint relationship, verifying the effectiveness of the priority strategy based on the ranking of individual deviation contribution values.

[0044] In this embodiment, if the overall performance deviation is still not lower than the threshold after a certain process is adjusted to the tolerance boundary, for example, if the deviation still exceeds the standard after the outer ring grinding setting parameter is adjusted to the tolerance boundary of 0.17 mm, the adjustment execution module will proceed to the next process according to the sequence, that is, start adjusting the raceway grinding process, and the adjustment method is exactly the same; the adjustment execution module will repeat this process until the overall performance deviation is lower than the threshold or all adjustable processes have reached the boundary; if the deviation still exceeds the standard after all processes are adjusted to the boundary, the system will issue an alarm to prompt the operator to check the upstream cause, such as excessive wear of the grinding wheel or abnormal coolant.

[0045] The bearing grinding parameter adaptive control system provided by this invention calculates the overall performance deviation of the entire process in real time, provided that the actual parameters of each process do not exceed the tolerance range of their respective processes. It determines the adjustment priority based on the individual deviation contribution value of each process and adjusts the set processing parameters of each process in a stepwise manner according to the priority order, so that the overall performance deviation is restored to below the allowable threshold. Throughout the process, the system does not need to rely on offline sampling results or require manual intervention in adjustment decisions, thus realizing true adaptive control of bearing grinding parameters.

[0046] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An adaptive control method for bearing grinding parameters, characterized in that, Includes the following steps: S1. Obtain the set processing parameters and corresponding tolerance ranges of each process in the bearing grinding process, wherein the processes include inner ring grinding, outer ring grinding, raceway grinding and ultra-precision machining in sequence; S2. Establish the correlation constraint relationship between the entire process processing parameters and bearing performance. The correlation constraint relationship is used to characterize the comprehensive influence of the main effect of a single process, the interactive coupling effect between processes, and the nonlinear coupling effect of the entire process on the overall performance of the bearing. S3. Under the condition that each process follows its own set processing parameters and the actual parameters do not exceed the tolerance range of each process, collect the actual processing parameters of each process in real time and calculate their normalized deviation, and substitute them into the correlation constraint relationship to calculate the current overall performance offset of the whole process. S4. Determine whether the overall performance deviation of the entire process exceeds the preset allowable performance deviation threshold: If it does not exceed the threshold, maintain the current processing parameters of each process and continue processing; if it exceeds the threshold, proceed to step S5. S5. Calculate the individual deviation contribution value of each process, where the individual deviation contribution value is the weighted deviation term corresponding to the process in the associated constraint relationship; sort the individual deviation contribution values ​​of each process from largest to smallest, and adjust each process in sequence according to the sorting: adjust the set processing parameters of the process in the direction of decreasing deviation by a preset step size, and recalculate the overall performance offset of the entire process; if it is still higher than the allowable performance offset threshold, continue to adjust the same process in the same direction until its set processing parameters reach the boundary of the tolerance range within the process; if it is still not lower than the threshold after reaching the boundary, proceed to the next process according to the sorting; repeat the above process until the overall performance offset of the entire process is lower than the allowable performance offset threshold; then control the subsequent bearing grinding process with the finally adjusted set processing parameters of each process.

2. The adaptive control method for bearing grinding parameters according to claim 1, characterized in that: The processing parameters set for each process in S1 include: grinding depth of the inner ring grinding process, grinding depth of the outer ring grinding process, raceway roundness deviation of the raceway grinding process, and surface roughness of the ultra-precision machining process.

3. The adaptive control method for bearing grinding parameters according to claim 1, characterized in that: The correlation constraint relationship in S2 includes: Let the actual machining parameters of each process, such as inner ring grinding, outer ring grinding, raceway grinding, and ultra-precision machining, be denoted as follows: The corresponding setting value is The tolerance range half-width within each process is Normalized deviation Then the association constraint relationship is: Where i and j represent process indices, and 1 ≤ i <j≤4, To determine the single-process main effect coefficients by combining orthogonal experiments and analysis of variance, based on the weighted proportion of the influence of each process parameter on the overall bearing performance. The coefficient represents the interaction coupling between each pair of processes. The coupling coefficient for the four processes is... It is a nonlinear saturation factor.

4. The adaptive control method for bearing grinding parameters according to claim 3, characterized in that: The calculation process is as follows: For each pair of processes i and j, two extreme value combination experiments are set up, keeping the deviation of the remaining processes at 0. In the first experiment, δ i and δ j If both are set to +1 or both are set to -1, the overall performance offset of the entire process under this combination is measured and recorded as ΔP. ij同 In the second experiment, δ i and δ j Let the tolerance boundary values ​​have opposite signs, i.e., one is +1 and the other is -1. The measured overall performance offset of the entire process is denoted as Δ. Pij异 ,but The calculation process for γ is as follows: the deviation δ of the four processes is... i Simultaneously, they are set to the same sign for their respective tolerance boundary values, both being +1 or -1. The overall performance offset ΔP of the entire process at this time is measured and then calculated using the formula... The solution yields, where This is the sum of the main effect terms of a single process. It is the sum of all pairwise interactive terms.

5. The adaptive control method for bearing grinding parameters according to claim 1, characterized in that: The calculation process of the overall performance offset in step S3 is as follows: During the sequential processing of each process, the actual processing parameters of each process are collected respectively, so that the collection timestamps of the four processes in this processing correspond one-to-one with the processing time of the corresponding process; all the actual processing parameters under the same timestamp are substituted into the correlation constraint relationship to solve for the overall performance offset of the entire process at that time. The overall performance offset of the entire process is a real-time scalar value that characterizes the degree of deviation of the bearing's overall performance.

6. The adaptive control method for bearing grinding parameters according to claim 1, characterized in that: The judgment logic for the overall performance deviation in step S4 is as follows: the preset allowable performance deviation threshold is 80% of the critical value of the bearing's overall performance qualification at the factory; after each calculation of the overall performance deviation, a threshold comparison is immediately performed; only when the value of the overall performance deviation is strictly greater than the allowable performance deviation threshold is it determined to be exceeded and the parameter adjustment process in step S5 is triggered; if it is equal to or less than the threshold, it is determined to be not exceeded and the current processing parameters remain unchanged.

7. The adaptive control method for bearing grinding parameters according to claim 1, characterized in that: In step S5, the calculation and sorting rules for the individual deviation contribution value of each process are as follows: extract the main effect term of the corresponding process and the sum of all pairwise interactive coupling terms containing the process from the correlation constraint relationship, and use it as the individual deviation contribution value of the process; arrange the individual deviation contribution values ​​of the four processes of inner ring grinding, outer ring grinding, raceway grinding and ultra-precision machining in descending order of value to generate a process adjustment priority sequence.

8. The adaptive control method for bearing grinding parameters according to claim 1, characterized in that: In step S5, the preset step size is 5% to 10% of the half-width of the tolerance range within the process of that process.

9. An adaptive control system for bearing grinding parameters, used to implement the adaptive control method for bearing grinding parameters as described in any one of claims 1 to 8, characterized in that, include: The data acquisition module is configured to acquire the set processing parameters for each process in bearing grinding and their corresponding tolerance ranges within the process. The relational construction module is configured to establish the correlation constraint relationship between the entire process processing parameters and the bearing performance. This correlation constraint relationship is used to characterize the comprehensive impact of the main effect of a single process, the interactive coupling effect between processes, and the nonlinear coupling effect of the entire process on the overall performance of the bearing. The real-time acquisition and calculation module is configured to acquire the actual processing parameters of each process in real time and calculate their normalized deviation under the condition that each process follows its own set processing parameters and the actual parameters do not exceed the tolerance range of each process. The normalized deviation is then substituted into the associated constraint relationship to solve for the current overall performance offset of the entire process. The comparison and judgment module is configured to determine whether the overall performance offset of the entire process exceeds a preset allowable performance offset threshold: if it does not exceed the threshold, the current processing parameters of each process are maintained and processing continues; if it exceeds the threshold, the adjustment execution module is triggered. The execution module is adjusted to calculate the individual deviation contribution value of each process. The individual deviation contribution value is the weighted deviation term corresponding to the process in the associated constraint relationship. The individual deviation contribution values ​​of each process are sorted from largest to smallest, and adjustments are performed on each process in sequence according to the sorting: the set processing parameters of the process are adjusted by a preset step size in the direction of decreasing deviation, and the overall performance offset of the entire process is recalculated. If it is still higher than the allowable performance offset threshold, the same process is adjusted in the same direction until its set processing parameters reach the boundary of the tolerance range within the process. If it is still not lower than the threshold after reaching the boundary, the process is moved to the next process according to the sorting. The above process is repeated until the overall performance offset of the entire process is lower than the allowable performance offset threshold. Then, the subsequent bearing grinding process is controlled by the finally adjusted set processing parameters of each process.

Images