A method for initial geometry-steady process coupling prediction and parameter regulation for precision cylindrical grinding roundness error

Abstract

The application discloses a kind of initial geometry-steady process coupling prediction and parameter regulation method for precision cylindrical grinding roundness error.The method measures initial roundness before grinding, roundness spectrum component and workpiece clamping eccentricity, collects grinding wheel speed, workpiece speed, grinding depth, feed amount, grinding wheel dressing state and cooling state, and constructs initial geometry-steady process coupling characteristics;Under the constraint of scale observable boundary, select the main information source of roundness, and preferentially use low-dimensional steady-state characteristics to train roundness error prediction model;Through cross-validation and uncertainty calibration, the prediction reliability is evaluated;Based on response surface and safety constraint, grinding wheel speed, workpiece speed, grinding depth and step grinding control suggestions are generated;The post-grinding measurement results are written back to update the parameter influence boundary of digital twin.The method can improve the stability of roundness error prediction and the engineering executability of parameter regulation.

Description

Technical Field

[0001] This invention relates to the fields of precision cylindrical grinding, geometric error prediction, process parameter optimization, digital twin and quality closed-loop control, and in particular to an initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding. Background Technology

[0002] The roundness error in cylindrical grinding is affected by both the initial geometric error before grinding and the grinding wheel speed, workpiece speed, grinding depth, feed method, and system stiffness. If only dynamic sensor signals are used to predict roundness, local transient disturbances are easily mistaken for the source of macroscopic geometric errors.

[0003] Existing roundness prediction models often incorporate roughness, vibration, and roundness into the same high-dimensional feature space, lacking information source boundary judgment for roundness indicators. This results in insufficient model interpretability, and parameter adjustment suggestions may not conform to equipment safety and process constraints.

[0004] Therefore, a method is needed to construct a roundness master model around the initial geometry and steady-state process parameters, and to generate parameter control suggestions by combining uncertainty calibration and safety constraints. Summary of the Invention

[0005] The technical problem to be solved by this invention is that the roundness error of cylindrical grinding is simultaneously affected by pre-grinding geometry inheritance, clamping eccentricity and steady-state process parameters, while dynamic sensing features can easily mask the dominant roundness factors and parameter control is constrained by safe rotation speed and surface quality. This invention provides an initial geometry-steady-state process coupled prediction and parameter control method.

[0006] To achieve the above objectives, the technical solution of this invention is as follows: an initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding, comprising: measuring the initial roundness, roundness spectral components, and clamping eccentricity before grinding; collecting the grinding wheel speed, workpiece speed, grinding depth, feed rate, grinding wheel dressing state, and cooling state; constructing initial geometry-steady-state process coupling features; selecting the main roundness information source based on the observable boundary of the scale; training the roundness error prediction model and outputting the predicted roundness error and uncertainty; generating parameter control suggestions under equipment safety constraints and process constraints; and writing the roundness measurement results after grinding back to the digital twin.

[0007] In the above-mentioned method for predicting and controlling the initial geometry-steady-state process coupling for roundness error in precision cylindrical grinding, the initial geometry-steady-state process coupling characteristics include the pre-grinding roundness peak and valley values, harmonic components, clamping eccentricity, grinding wheel linear speed, workpiece linear speed, speed ratio, unit grinding depth, and grinding wheel dressing interval.

[0008] In the above-mentioned method for initial geometry-steady-state process coupling prediction and parameter control of roundness error in precision cylindrical grinding, the scale observable boundary is determined by ablation comparison of the contributions of dynamic sensing features, steady-state process features and initial geometry features to roundness error prediction. When dynamic sensing features do not improve cross-validation performance, they are not written into the roundness master model.

[0009] In the above-mentioned method for initial geometry-steady-state process coupling prediction and parameter control of roundness error in precision cylindrical grinding, the roundness error prediction model adopts a regression model or a Gaussian process model, and outputs predicted values, confidence intervals, and ranking of major contribution features.

[0010] In the above-mentioned method for initial geometry-steady-state process coupling prediction and parameter control of roundness error in precision cylindrical grinding, the parameter control suggestions include increasing or decreasing the grinding wheel speed, adjusting the workpiece speed, limiting the grinding depth, adopting step grinding or pre-correction, and all suggestions meet the constraints of equipment safe speed, thermal load and surface quality.

[0011] In the above-mentioned method for initial geometry-steady-state process coupling prediction and parameter control of roundness error in precision cylindrical grinding, when the initial roundness before grinding exceeds the preset threshold, the system prioritizes outputting step-by-step grinding or pre-correction suggestions instead of directly increasing the depth of grinding per pass.

[0012] In the above-mentioned method for initial geometry-steady-state process coupling prediction and parameter control of roundness error in precision cylindrical grinding, the digital twin updates the parameter influence boundary, feature weight and uncertainty calibration coefficient according to the roundness error after grinding and the actual control results.

[0013] The technical solution adopted in this invention includes the following steps: initial geometric measurement, scale-observable boundary determination, roundness prediction, and safety control: S1. Measure the initial roundness, roundness spectrum components and clamping eccentricity before grinding, and collect steady-state process parameters and grinding wheel dressing status.

[0014] S2. Construct the initial geometry-steady-state process coupling features and normalize them according to workpiece batch, material, and grinding stage.

[0015] S3. Identify the observable boundary of the roundness error scale by ablation comparison to determine whether to introduce dynamic sensing features.

[0016] S4. Train the roundness error prediction model and output the predicted value of the roundness error after grinding, the confidence interval, and the key contribution features.

[0017] S5. Establish a response surface or local sensitivity model to analyze the influence of grinding wheel speed, workpiece speed and grinding depth on roundness error.

[0018] S6. Generate parameter control suggestions by combining equipment safety, heat load, surface quality and cycle time constraints.

[0019] S7. Write back the roundness measurement value after grinding and the control execution result to the digital twin, and update the feature weights and parameter influence boundaries.

[0020] By adopting the above-described technical solution, the present invention has the following beneficial effects: 1. It can distinguish between macroscopic geometric errors and local dynamic disturbances, avoiding the dominance of initial roundness by high-dimensional sensing features.

[0021] 2. By selecting information sources through scale-observable boundaries, the stability and interpretability of the roundness prediction model can be improved.

[0022] 3. By providing uncertainty outputs to indicate low-reliability operating conditions, quality fluctuations caused by incorrect parameter adjustments can be reduced.

[0023] 4. Parameter suggestions are generated through safety constraints to make the control results more suitable for on-site implementation.

[0024] 5. The digital twin is updated through post-grinding feedback, enabling the roundness model to adapt to changes in material, grinding wheel, and equipment conditions. Attached Figure Description

[0025] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The schematic structures and processes in the drawings are used to explain the invention and do not constitute an undue limitation on the scope of protection of the invention.

[0026] Figure 1 This is a schematic diagram of the overall process of an embodiment of the present invention; Figure 2 This is a schematic diagram of key modules or structural relationships in an embodiment of the present invention; The reference numerals in the figures include: Figure 1 In Chinese: S101 Pre-grinding geometric measurement; S102 Steady-state process acquisition; S103 Observable boundary judgment; S104 Roundness error prediction; S105 Safety control recommendations; S106 Post-grinding feedback update; Figure 2 In the middle: 10 is the input layer, including pre-grind roundness, clamping eccentricity, and steady-state process parameters; 20 is the core processing layer, including coupled feature construction, observable boundary judgment, roundness error regression, and safety constraint control; 30 is the output layer, including roundness prediction, parameter control suggestions, and uncertainty marking; 40 is the feedback layer, including the influence boundary of calibration parameters on post-grind measurement results. Figure 1 and Figure 2 The arrows in the diagram indicate the direction of data, control commands, or execution feedback. Detailed Implementation

[0027] The invention will be further described in conjunction with the accompanying drawings.

[0028] Please see Figure 1 , Figure 2 This invention provides an initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding, comprising: measuring the initial roundness, roundness spectral components, and clamping eccentricity before grinding; collecting data on grinding wheel speed, workpiece speed, grinding depth, feed rate, grinding wheel dressing state, and cooling state; constructing initial geometry-steady-state process coupling features; selecting the main roundness information source based on the observable boundary of the scale; training the roundness error prediction model and outputting the predicted roundness error and uncertainty; generating parameter control suggestions under equipment safety constraints and process constraints; and writing the roundness measurement results after grinding back to the digital twin.

[0029] Furthermore, the initial geometry-steady-state process coupling characteristics include pre-grinding roundness peak and valley values, harmonic components, clamping eccentricity, grinding wheel linear velocity, workpiece linear velocity, speed ratio, unit grinding depth, and grinding wheel dressing interval.

[0030] Furthermore, the scale observable boundary is determined by ablation comparison of the contributions of dynamic sensing features, steady-state process features, and initial geometric features to roundness error prediction. When dynamic sensing features do not improve cross-validation performance, they are not written into the roundness master model.

[0031] Furthermore, the roundness error prediction model adopts a regression model or a Gaussian process model, and outputs predicted values, confidence intervals, and ranking of major contributing features.

[0032] Furthermore, the parameter control recommendations include increasing or decreasing the grinding wheel speed, adjusting the workpiece speed, limiting the grinding depth, using step grinding or pre-correction, and all recommendations meet the equipment's safe speed, thermal load, and surface quality constraints.

[0033] Furthermore, when the initial roundness before grinding exceeds a preset threshold, the system prioritizes outputting step-by-step grinding or pre-correction suggestions instead of directly increasing the depth of grinding per pass.

[0034] Furthermore, the digital twin updates the parameter influence boundary, feature weights, and uncertainty calibration coefficients based on the post-grind roundness error and actual control results.

[0035] To facilitate understanding and explanation, the present invention will be further described below with reference to the accompanying drawings and specific embodiments: In one embodiment, before grinding, a roundness meter or online measuring device is used to obtain the workpiece roundness peak and valley values, harmonic components, and clamping eccentricity, while simultaneously recording the grinding wheel speed, workpiece speed, grinding depth, feed rate, grinding wheel dressing interval, and cooling status.

[0036] Before model training, the system constructs input combinations such as Initial-only, Process-only, Dynamic-only, Process+Initial, and All-features, and compares the prediction error and confidence interval width through cross-validation.

[0037] If the Process+Initial combination is better than the All-features combination, it indicates that the dynamic sensing features contribute little to roundness prediction. The system uses the initial geometry and steady-state process features as the main source of roundness information and uses the dynamic sensing features only as a reference for anomaly diagnosis.

[0038] During parameter adjustment, the system determines the adjustment direction based on the response surface. Within the equipment's safe range, the grinding wheel speed can be appropriately increased to enhance material removal continuity, while excessively high workpiece speeds and excessive grinding depths can be limited to reduce load fluctuations.

[0039] When the roundness error before grinding is large, the system will prioritize step grinding or pre-correction rather than increasing the grinding depth all at once, in order to reduce secondary errors caused by thermal deformation and elastic tool deflection.

[0040] Post-grinding measurement results are used to update the parameter influence boundaries in the digital twin. If prediction deviations occur multiple times under the same material or grinding wheel condition, the system recalibrates the model uncertainty and prompts for additional samples.

[0041] In a further embodiment, the roundness spectrum components are modeled according to low-order clamping eccentricity terms and high-order vibration-induced terms respectively; the low-order components are used to determine the workpiece clamping status, and the high-order components are used to assist in identifying abnormalities in the grinding wheel or support system.

[0042] The parameter control module can output multiple sets of alternative parameters and sort them according to roundness improvement, roughness risk, processing cycle time impact and equipment safety margin.

[0043] When post-grinding measurement results indicate that parameter adjustments have failed to improve roundness, the system marks the sample as an abnormal sample and prompts for checking the grinding wheel dressing, workpiece support, or measurement reference.

[0044] In a further embodiment, the roundness error is divided into two parts: inherited geometric error and process-induced error. The inherited geometric error is mainly characterized by pre-grind roundness, low-order components of the roundness spectrum, and clamping eccentricity, while the process-induced error is mainly characterized by speed ratio, unit grinding depth, grinding wheel dressing interval, and cooling state.

[0045] After the input combination comparison is completed, the system retains the dynamic sensing features that do not contribute enough in the anomaly diagnosis branch, instead of writing them into the roundness master model; the roundness master model outputs the predicted value, confidence interval, and contribution feature ranking for the parameter control module to call.

[0046] The parameter control module can output multiple sets of alternative parameters and sort them according to roundness improvement, roughness risk, processing cycle time impact and equipment safety margin. When the roundness target conflicts with the roughness target, a conservative control strategy is selected according to the preset quality priority.

[0047] When the initial roundness before grinding exceeds the threshold, the system prioritizes step grinding, pre-correction, or reclamping, rather than directly increasing the depth of grinding per pass, in order to reduce thermal deformation, elastic tool deflection, and surface quality fluctuations.

[0048] After the post-grinding measurement results are written back, if the predicted deviation occurs continuously under the same material batch, grinding wheel condition or clamping method, the system will recalibrate the uncertainty coefficient and prompt to check the grinding wheel dressing, workpiece support or measurement reference.

[0049] The foregoing has provided a detailed description of an initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding, as provided in the embodiments of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. It should be understood that the above embodiments are only for illustrating the present invention and are not intended to limit the scope of protection of the present invention. Those skilled in the art can make equivalent substitutions for specific equipment models, parameter ranges, data interfaces, and algorithm models without departing from the concept of the present invention. Any modifications, equivalent substitutions, improvements, or combinations made within the spirit and principles of the present invention should be included within the scope of protection of the present invention. In summary, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for initial geometry-steady-state process coupling prediction and parameter control of roundness error in precision cylindrical grinding, characterized in that, include: Measure the initial roundness, roundness spectrum components, and clamping eccentricity before grinding; Data such as grinding wheel speed, workpiece speed, grinding depth, feed rate, grinding wheel dressing status, and cooling status are collected; initial geometry-steady-state process coupling characteristics are constructed. Select the main source of roundness information based on the observable boundary of scale; train the roundness error prediction model and output the predicted roundness error and uncertainty; generate parameter adjustment suggestions under equipment safety constraints and process constraints; Write the roundness measurement results after grinding back into the digital twin.

2. The initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding according to claim 1, characterized in that, The initial geometry-steady-state process coupling characteristics include pre-grinding roundness peak and valley values, harmonic components, clamping eccentricity, grinding wheel linear velocity, workpiece linear velocity, speed ratio, unit grinding depth, and grinding wheel dressing interval.

3. The initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding according to claim 1, characterized in that, The observable boundary of the scale is determined by ablation comparison of the contributions of dynamic sensing features, steady-state process features and initial geometric features to the roundness error prediction. When dynamic sensing features do not improve cross-validation performance, they are not written into the roundness master model.

4. The initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding according to claim 1, characterized in that, The roundness error prediction model adopts a regression model or a Gaussian process model, and outputs predicted values, confidence intervals, and ranking of major contributing features.

5. The initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding according to claim 1, characterized in that, The parameter control recommendations include increasing or decreasing the grinding wheel speed, adjusting the workpiece speed, limiting the grinding depth, using step grinding or pre-correction, and all recommendations must meet the equipment's safe speed, thermal load, and surface quality constraints.

6. The initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding according to claim 1, characterized in that, When the initial roundness before grinding exceeds the preset threshold, the system prioritizes outputting step-by-step grinding or pre-correction suggestions instead of directly increasing the depth of grinding per pass.

7. The initial geometry-steady-state process coupling prediction and parameter control method for roundness error in precision cylindrical grinding according to claim 1, characterized in that, The digital twin updates the parameter influence boundary, feature weights, and uncertainty calibration coefficients based on the post-grind roundness error and actual control results.

Images