Lens hardening process adaptive optimization method and system based on data driving

Abstract

The invention relates to the technical field of data processing, and provides a lens hardening process adaptive optimization method and system based on data driving. The method comprises the following steps: collecting a plurality of process parameters in a lens hardening process in real time; calling a hardening process prediction model, performing lens hardness prediction based on the plurality of process parameters, and comparing the predicted lens hardness parameter with a preset hardness value to generate a predicted hardness deviation; performing adaptive optimization on the plurality of process parameters according to the predicted hardness deviation to generate optimized process parameters; and the optimized process parameters are fed back to a production control end of the lens hardening process, and regulation and control of the lens hardening process parameters are executed. The technical problems of uneven hardness and low production efficiency caused by unstable process parameters in the lens hardening process are solved, and the technical effects of ensuring the consistency of the lens hardness and improving the production efficiency and the product quality through real-time prediction and adaptive optimization of the process parameters are achieved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a data-driven adaptive optimization method and system for lens hardening processes. Background Art

[0002] Lens hardening, an important method for strengthening the surfaces of materials like glass and plastic, is widely used in the production of optical lenses, display screens, and various functional lenses. This process increases the hardness of the lens surface through physical and chemical processes such as heating and cooling, thereby enhancing its wear resistance, scratch resistance, and service life. However, traditional lens hardening processes face several technical challenges, particularly regarding the stability and precision of process parameters. Currently, process control during lens hardening relies heavily on empirical rules and manual adjustments. Due to differences in material properties between batches and equipment precision limitations, uneven hardness distribution can easily occur, impacting product quality. Furthermore, slight variations in process parameters such as temperature, pressure, and time can lead to variations in hardness. These variations often go undetected and uncorrected during the production process, resulting in low production efficiency and increased scrap rates. Particularly in large-scale production, the lack of real-time feedback mechanisms and optimization methods results in inconsistent product quality, making it difficult to meet the stringent requirements for lens hardness uniformity and precision in the high-end market. Summary of the Invention

[0003] This application provides a data-driven adaptive optimization method and system for lens hardening processes, aiming to solve the technical problems of uneven hardness and low production efficiency caused by unstable process parameters in the lens hardening process.

[0004] The first aspect disclosed in the present application provides a data-driven adaptive optimization method for a lens hardening process, the method comprising: real-time collection of multiple process parameters during the lens hardening process; calling a hardening process prediction model, predicting the lens hardness based on the multiple process parameters, comparing the predicted lens hardness parameters with a predetermined hardness value, and generating a predicted hardness deviation; adaptively optimizing the multiple process parameters using the predicted hardness deviation to generate optimized process parameters; and feeding back the optimized process parameters to a production control end of the lens hardening process to execute regulation of the lens hardening process parameters.

[0005] Another aspect disclosed in the present application provides a data-driven adaptive optimization system for lens hardening processes, the system comprising: a data acquisition module for real-time acquisition of multiple process parameters during the lens hardening process; a hardness prediction module for calling a hardening process prediction model, predicting the lens hardness based on the multiple process parameters, comparing the predicted lens hardness parameters with a predetermined hardness value, and generating a predicted hardness deviation; an adaptive optimization module for adaptively optimizing the multiple process parameters using the predicted hardness deviation to generate optimized process parameters; a process parameter control module for feeding back the optimized process parameters to the production control end of the lens hardening process to execute control of the lens hardening process parameters.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] The above-mentioned data-driven adaptive optimization method for lens hardening processes first collects various process parameters (such as temperature, pressure, time, etc.) during the lens hardening process in real time and uses a hardening process prediction model to predict the hardness of the lens based on these parameters. Subsequently, the predicted hardness value is compared with the predetermined target hardness value to calculate the hardness deviation. After that, based on this deviation, the relevant process parameters are automatically adjusted to optimize the hardness level and generate new process parameters. Finally, these optimized process parameters are fed back to the production control end for corresponding adjustment and execution, thereby ensuring precise control of the lens hardening process and stable product quality.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0010] Figure 1 Schematic diagram of a flow chart of a data-driven adaptive optimization method for a lens hardening process in one embodiment.

[0011] Figure 2 This is a diagram of the architecture of a data-driven adaptive optimization system for lens hardening processes in one embodiment.

[0012] Description of the accompanying drawings: data acquisition module 11, hardness prediction module 12, adaptive optimization module 13, process parameter control module 14. DETAILED DESCRIPTION

[0013] The embodiments of the present application solve the technical problems of uneven hardness and low production efficiency caused by unstable process parameters in the lens hardening process by providing a data-driven adaptive optimization method and system for the lens hardening process.

[0014] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0015] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0016] Example 1, as Figure 1 As shown, the present application provides a data-driven adaptive optimization method for lens hardening process, the method comprising:

[0017] Real-time collection of multiple process parameters during lens hardening.

[0018] In an embodiment of the present application, during the lens hardening process, in order to achieve precise control and optimization of process parameters, a variety of high-precision sensors are first installed at key locations of the lens hardening production line, such as hardening tanks, curing furnaces, conveying devices, etc. These sensors include but are not limited to temperature sensors, pressure sensors, time recording devices, solution concentration detectors, and pH meters, etc., among which the temperature sensor monitors the temperature changes of the solution in the hardening tank in real time to ensure that the lens hardening process is always within the optimal range required by the process; the pressure sensor accurately measures the pressure applied to the lens surface during the hardening process to ensure the stability and uniformity of the pressure; the time recording device accurately records the residence time of each lens in the hardening process to avoid the influence of time deviation on the hardness; the solution concentration detector continuously detects the concentration of the hardening solution, and can provide timely feedback when the concentration changes due to evaporation or chemical reaction; the pH meter monitors the pH of the solution in real time to ensure that it is within the appropriate range to maintain the stability of the hardening reaction. These sensors transmit the collected data back to the system to form multiple process parameters, such as temperature, pressure, solution concentration, pH value, etc. These process parameters not only provide a basis for subsequent process optimization and prediction, but also provide real-time feedback for quality control of the entire production process, ensuring that each link is carried out under optimal process conditions, thereby improving production efficiency, reducing scrap rate, and ensuring the stability and consistency of product quality.

[0019] The hardening process prediction model is called to predict the hardness of the lens based on the multiple process parameters, and the predicted lens hardness parameters are compared with the predetermined hardness value to generate a predicted hardness deviation.

[0020] In one embodiment, during the lens hardening process, multiple process parameters are first input into a pre-trained hardening process prediction model. This hardening process prediction model is a machine learning model built based on historical data and regression algorithms. By learning the process parameters and hardness data of previous hardening processes, the model can predict the hardness value of the lens under the current hardening conditions. The hardening process prediction model uses the learned knowledge to calculate a predicted hardness value based on the received process parameters. This predicted value reflects the hardness that the lens surface may reach under the current parameters. After calculating the predicted hardness value, the predicted value is compared with the predetermined target hardness value to calculate a predicted hardness deviation. This predicted hardness deviation reflects the gap between the current process conditions and the expected process effect. The size of the deviation will directly affect subsequent process optimization and adjustment. If the predicted hardness deviation is large, an adaptive optimization process will be triggered to adjust the process parameters to ensure that the hardness of the final product meets the predetermined standards, thereby ensuring the quality and performance of the lens.

[0021] Furthermore, the present application provides a method for calling a hardening process prediction model, predicting the hardness of a lens based on the multiple process parameters, comparing the predicted lens hardness parameters with a predetermined hardness value, and generating a predicted hardness deviation, including:

[0022] Collect historical hardening process detection data sets, and train the hardening process prediction model based on regression learning; input the multiple process parameters into the hardening process prediction model to predict the lens hardness, and output the predicted lens hardness parameters; calculate the difference between the predicted lens hardness parameters and the predetermined hardness value to generate the predicted hardness deviation.

[0023] Preferably, first, a large amount of historical hardening process data is collected. These data include various process parameters (such as temperature, pressure, time, etc.) in the previous production process and the corresponding lens hardness values. Each set of data represents a complete hardening process, recording all the key parameters at the time and the final measured hardness value. By summarizing these data, a complete historical hardening process detection data set is formed for subsequent model training. Subsequently, a regression neural network is used to train the hardening process prediction model. The regression neural network is a machine learning method that uses the process parameters in the historical data as input features and the corresponding lens hardness values ​​as target outputs. The relationship between process parameters and hardness is learned through steps such as forward propagation, loss calculation, back propagation, and parameter optimization. A portion of the data that has not been used for training is used as a validation set to evaluate the trained model. When the evaluation result shows that the prediction accuracy meets the preset accuracy, the current regression neural network will be used as the hardening process prediction model. Afterwards, the real-time collected process parameters are input into a trained hardening process prediction model. Based on the learned mapping relationship between process parameters and hardness, the hardening process prediction model outputs a predicted lens hardness parameter. This predicted lens hardness parameter represents an estimated value for the lens surface hardness under the current process conditions. The predicted lens hardness parameter is then compared with a predetermined hardness value, which is an ideal hardness value determined based on product standards or design requirements. The difference between the predicted lens hardness parameter and the predetermined hardness value is calculated to obtain a predicted hardness deviation. The magnitude and direction of this predicted hardness deviation reflect the error in lens hardness under the current process conditions. A small deviation indicates that the current process parameters are relatively accurate and the lens hardness is close to the predetermined value. A large deviation indicates that the current process parameters need to be optimized and adjusted to ensure the quality of the final product. This process enables intelligent prediction and feedback of the lens hardening process, providing data support for subsequent adaptive optimization and ensuring that hardness control during the production process remains within the predetermined standard range.

[0024] Adaptive optimization of the plurality of process parameters is performed based on the predicted hardness deviation to generate optimized process parameters.

[0025] In one embodiment, after obtaining the predicted hardness deviation, it is determined whether the predicted hardness deviation exceeds a preset deviation threshold. If it is less than the deviation threshold, it indicates that the current process parameters basically meet the requirements and do not need to be adjusted. However, if the deviation exceeds the deviation threshold, it indicates that the process parameters need to be optimized. At this time, the relationship between the predicted hardness deviation and each process parameter is analyzed through historical data and statistical analysis methods to evaluate the weight of each process parameter on the hardness deviation. For example, changes in certain process parameters may have a greater impact on hardness, while changes in other parameters have a smaller impact on hardness. Subsequently, a weight is determined based on each parameter for adaptive adjustment. For example, some process parameters may require a large adjustment, while other parameters only require a slight adjustment. The optimization process gradually adjusts the process parameters until the predicted hardness deviation is close to the predetermined hardness value, thereby generating the final optimized process parameters. These optimized process parameters can effectively reduce the hardness deviation and ensure that the hardness of the lens reaches the predetermined target. These parameters will be used as control parameters in subsequent production processes and applied to actual production, thereby improving product quality consistency and production efficiency.

[0026] Furthermore, the present application provides a method for adaptively optimizing the plurality of process parameters based on the predicted hardness deviation to generate optimized process parameters, including:

[0027] Determine whether the predicted hardness deviation exceeds a preset deviation threshold; if so, analyze the deviation influence weights of the predicted hardness deviation and the multiple process parameters, and construct a variable optimization weight distribution; based on the variable optimization weight distribution, adaptively optimize the multiple process parameters to generate the optimized process parameters.

[0028] Preferably, after obtaining the predicted hardness deviation, the predicted hardness deviation is first compared with a preset deviation threshold, which is determined based on product standards and production requirements. If the predicted hardness deviation exceeds the preset deviation threshold, it indicates that there is a significant deviation in the current process parameters. At this time, based on historical data, a statistical analysis method is used to analyze the impact of each process parameter on the predicted deviation, identify the key factors that cause the deviation, such as temperature, pressure, time, etc., and assign a deviation impact weight to each parameter. This weight represents the contribution of the parameter to the hardness deviation. The greater the parameter's influence, the higher its weight, which means it will play a more important role in the optimization process. After analyzing the impact of each process parameter, these deviation impact weights are normalized to ensure that the sum of the weights of all process parameters is 1, thereby forming a variable optimization weight distribution. This variable optimization weight distribution reflects the degree of attention that different process parameters should receive during the optimization process. Subsequently, based on the constructed variable optimization weight distribution, multiple process parameters are adaptively optimized. Specifically, the value of each process parameter is adjusted according to the optimization weight of each parameter, and the hardness deviation is minimized through multiple iterations to approach the predetermined hardness value. After the adaptive optimization is completed, optimized process parameters will be generated. These optimized process parameters reflect how to adjust various process parameters under the current production conditions to ensure that the hardness of the lens meets the predetermined standards, thereby greatly improving the automation level of the production line and the stability of product quality, and reducing the scrap rate.

[0029] Furthermore, the present application provides a method for adaptively optimizing the plurality of process parameters based on the variable optimization weight distribution to generate the optimized process parameters, including:

[0030] Based on the variable optimization weight distribution and the predicted hardness deviation, the multiple process parameters are initialized and adjusted to generate a first adjustment parameter set and a second adjustment parameter set; the hardening process prediction model is called to perform hardness prediction on the first adjustment parameter set and the second adjustment parameter set to generate a first hardness prediction result and a second hardness prediction result; the first hardness prediction result and the second hardness prediction result are positioned as the hardness optimization direction and the temporary optimal adjustment parameters; the temporary optimal adjustment parameters are continued to be adjusted according to the hardness optimization direction, and iterated multiple times until the adjustment parameters that meet the convergence conditions are obtained to generate the optimized process parameters.

[0031] Optionally, when performing adaptive optimization, multiple process parameters will first be preliminarily adjusted based on the constructed variable optimization weight distribution and the current predicted hardness deviation. The purpose of the initialization adjustment is to quickly identify the process parameters that need to be adjusted significantly based on the hardness deviation and the influence weight of each process parameter, and to make preliminary adjustments. For those process parameters that have a greater impact on the hardness deviation (i.e., parameters with a larger weight), preliminary adjustments will be made according to a larger adjustment step size. For process parameters that have a smaller impact (i.e., parameters with a smaller weight), preliminary adjustments will be made according to a smaller adjustment step size. Through two random direction adjustments, a first adjustment parameter set and a second adjustment parameter set can be generated. These two sets of parameters will be compared and optimized in subsequent hardness predictions. Subsequently, the first adjustment parameter set and the second adjustment parameter set are respectively input into the hardening process prediction model to predict the lens hardness corresponding to each set of parameters, and a first hardness prediction result corresponding to the first adjustment parameter set and a second hardness prediction result corresponding to the second adjustment parameter set are generated. These two prediction results represent the estimated value of the lens hardness under different parameter adjustment conditions. The two predictions are then compared, and the difference between them determines the direction of hardness optimization and the temporary optimal adjustment parameters. The temporary optimal adjustment parameters are then adjusted according to the hardness optimization direction. After each adjustment, the hardening process prediction model is called to perform a hardness prediction and calculate the new hardness deviation. This process continues until the hardness deviation converges, that is, the deviation gradually decreases to reach the preset deviation threshold. Ultimately, after multiple iterations, a set of optimized process parameters is obtained. These optimized process parameters can effectively adjust the lens hardness to the predetermined hardness value, thereby ensuring that the lens hardness always meets the predetermined standard, improving the accuracy and efficiency of the production process, and reducing the scrap rate.

[0032] Furthermore, the present application provides that after generating the first hardness prediction result and the second hardness prediction result, the method further includes:

[0033] The differences between the first hardness prediction result, the second hardness prediction result and the predetermined hardness value are calculated respectively to generate a first difference and a second difference; if any difference between the first difference and the second difference meets the preset deviation threshold, the optimized process parameters are directly generated.

[0034] Optionally, after obtaining the first hardness prediction result and the second hardness prediction result, the first hardness prediction result and the second hardness prediction result are respectively compared with the predetermined hardness value to calculate a first difference and a second difference, wherein the first difference represents the error between the predicted hardness and the predetermined hardness under the first adjustment parameter set. The larger the difference, the more significant the effect of the currently adjusted process parameters on the hardness and the greater the deviation; the second difference reflects the error between the predicted hardness and the target hardness under the second adjustment parameter set. Subsequently, the calculated first difference and the second difference are compared with a preset deviation threshold. If either the first difference or the second difference is less than or equal to the preset deviation threshold, it means that the current process parameters are close enough to the target hardness value and no further adjustment is required. At this time, the currently adjusted process parameters are directly used as the optimized process parameters, thereby achieving efficient and accurate production process control.

[0035] Furthermore, the present application provides a method for analyzing the influence weights of the predicted hardness deviation and the deviations of the multiple process parameters and constructing a variable optimization weight distribution, including:

[0036] Based on the historical hardening process detection data set, the statistical analysis method is used to evaluate the influence of each process parameter on the hardness deviation, and determine the multiple influence weights of the multiple process parameters; the multiple influence weights are normalized so that the sum of the weights is 1, and the variable optimization weight distribution is generated.

[0037] Optionally, before optimization, a historical hardening process detection data set is obtained from a historical database. This historical hardening process detection data set contains process parameters (such as temperature, pressure, time, etc.) of different production batches and corresponding hardness measurement values. These data will be used to analyze the relationship between each process parameter and hardness deviation. Subsequently, the process parameters and corresponding hardness values ​​in each production cycle are extracted from the historical hardening process detection data set, including the specific value of each process parameter and the hardness measured after each production. The correlation coefficient between each process parameter and hardness is calculated by statistical analysis methods such as the Pearson correlation coefficient to determine the strength of the relationship between the change in the parameter and the hardness deviation. The larger the correlation coefficient, the greater the degree of influence of the process parameter on the hardness deviation. By using the correlation coefficient of each process parameter as the weight of the process parameter, multiple influence weights of multiple process parameters are obtained. The influence weight reflects the relative importance of each process parameter in the optimization process. Afterwards, by dividing each influence weight by the sum of the influence weights, the normalization processing of the multiple influence weights is completed so that the sum of the weights is 1, ensuring that the optimization adjustment of each process parameter is relatively balanced and avoiding excessive influence of a certain parameter on the process optimization process. The normalized multiple influence weights will be organized into a variable optimization weight distribution to reflect the relative importance of each process parameter in the optimization process. Process parameters with larger weights will be adjusted more during the optimization process, while parameters with smaller weights will be adjusted less or may not even be adjusted, thereby ensuring that hardness deviation is minimized and ideal production results are achieved.

[0038] Furthermore, the present application provides positioning the first hardness prediction result and the second hardness prediction result as a hardness optimization direction and a temporary optimal adjustment parameter, including:

[0039] Compare the first hardness prediction result and the second hardness prediction result, take the adjustment parameter with smaller hardness as the starting point and the adjustment parameter with larger hardness as the end point, analyze the change direction of each parameter from the starting point to the end point, and generate the hardness optimization direction; compare the first hardness prediction result and the second hardness prediction result, and generate the temporary optimal adjustment parameter with the adjustment parameter with larger hardness.

[0040] Optionally, after obtaining the first hardness prediction result and the second hardness prediction result, the two prediction results will be compared, and the prediction result with smaller hardness will be used as the starting point of the optimization process, and the prediction result with larger hardness will be used as the end point of the optimization process. Subsequently, the direction of change of the adjustment parameter from the starting point to the end point is analyzed, that is, the adjustment parameter values ​​corresponding to the two prediction results are compared to determine how to adjust the parameter from the adjustment parameter with smaller hardness (starting point) to the adjustment parameter with larger hardness (end point). For example, if it is found through comparison that a change in temperature from a small value to a large value causes the hardness to increase from a small value to a large value, then the increase in temperature is the direction of temperature change. By summarizing the change directions of these parameters, the hardness optimization direction is formed. In addition, the adjustment parameter corresponding to the prediction result with larger hardness will be selected as the temporary optimal adjustment parameter to further reduce the hardness deviation to approach the predetermined hardness value. This temporary optimal adjustment parameter will serve as the basis for subsequent optimization adjustments to ensure that hardness control can be performed more accurately and improve production efficiency.

[0041] The optimized process parameters are fed back to the production control end of the lens hardening process to perform regulation of the lens hardening process parameters.

[0042] In one embodiment, after the optimized process parameters are generated, these optimized process parameters will be transmitted to the production control end of the lens hardening process. After receiving the optimized process parameters, the production control end will adjust the hardening process according to these new parameters, specifically including adjusting the temperature, pressure, time, etc., to ensure that the hardening process is always in the optimal state, thereby improving production efficiency and reducing scrap rate.

[0043] In summary, the embodiments of the present application have at least the following technical effects:

[0044] The embodiment of the present application first collects multiple process parameters in real time during the lens hardening process; then, calls a hardening process prediction model to predict the lens hardness based on the multiple process parameters, compares the predicted lens hardness parameters with a predetermined hardness value, and generates a predicted hardness deviation; then, adaptively optimizes the multiple process parameters using the predicted hardness deviation to generate optimized process parameters; finally, feeds the optimized process parameters back to the production control end of the lens hardening process to execute the regulation of the lens hardening process parameters. These technical effects jointly solve the technical problems of uneven hardness and low production efficiency caused by unstable process parameters in the lens hardening process, and achieve the technical effect of ensuring lens hardness consistency and improving production efficiency and product quality through real-time prediction and adaptive optimization of process parameters.

[0045] The second embodiment is based on the same inventive concept as the data-driven lens hardening process adaptive optimization method in the above embodiment. Figure 2As shown, the present application provides a data-driven adaptive optimization system for lens hardening process, and the system includes: a data acquisition module 11: real-time acquisition of multiple process parameters in the lens hardening process; a hardness prediction module 12: calling the hardening process prediction model, predicting the lens hardness based on the multiple process parameters, comparing the predicted lens hardness parameters with the predetermined hardness value, and generating a predicted hardness deviation; an adaptive optimization module 13: adaptively optimizing the multiple process parameters with the predicted hardness deviation to generate optimized process parameters; a process parameter control module 14: feeding back the optimized process parameters to the production control end of the lens hardening process to execute the control of the lens hardening process parameters.

[0046] Furthermore, the hardness prediction module 12 is further configured to perform the following method:

[0047] Collect historical hardening process detection data sets, and train the hardening process prediction model based on regression learning; input the multiple process parameters into the hardening process prediction model to predict the lens hardness, and output the predicted lens hardness parameters; calculate the difference between the predicted lens hardness parameters and the predetermined hardness value to generate the predicted hardness deviation.

[0048] Furthermore, the adaptive optimization module 13 is further configured to perform the following method:

[0049] Determine whether the predicted hardness deviation exceeds a preset deviation threshold; if so, analyze the deviation influence weights of the predicted hardness deviation and the multiple process parameters, and construct a variable optimization weight distribution; based on the variable optimization weight distribution, adaptively optimize the multiple process parameters to generate the optimized process parameters.

[0050] Furthermore, the adaptive optimization module 13 is further configured to perform the following method:

[0051] Based on the variable optimization weight distribution and the predicted hardness deviation, the multiple process parameters are initialized and adjusted to generate a first adjustment parameter set and a second adjustment parameter set; the hardening process prediction model is called to perform hardness prediction on the first adjustment parameter set and the second adjustment parameter set to generate a first hardness prediction result and a second hardness prediction result; the first hardness prediction result and the second hardness prediction result are positioned as the hardness optimization direction and the temporary optimal adjustment parameters; the temporary optimal adjustment parameters are continued to be adjusted according to the hardness optimization direction, and iterated multiple times until the adjustment parameters that meet the convergence conditions are obtained to generate the optimized process parameters.

[0052] Furthermore, the adaptive optimization module 13 is further configured to perform the following method:

[0053] The differences between the first hardness prediction result, the second hardness prediction result and the predetermined hardness value are calculated respectively to generate a first difference and a second difference; if any difference between the first difference and the second difference meets the preset deviation threshold, the optimized process parameters are directly generated.

[0054] Furthermore, the adaptive optimization module 13 is further configured to perform the following method:

[0055] Based on the historical hardening process detection data set, the statistical analysis method is used to evaluate the influence of each process parameter on the hardness deviation, and determine the multiple influence weights of the multiple process parameters; the multiple influence weights are normalized so that the sum of the weights is 1, and the variable optimization weight distribution is generated.

[0056] Furthermore, the adaptive optimization module 13 is further configured to perform the following method:

[0057] Compare the first hardness prediction result and the second hardness prediction result, take the adjustment parameter with smaller hardness as the starting point and the adjustment parameter with larger hardness as the end point, analyze the change direction of each parameter from the starting point to the end point, and generate the hardness optimization direction; compare the first hardness prediction result and the second hardness prediction result, and generate the temporary optimal adjustment parameter with the adjustment parameter with larger hardness.

[0058] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0059] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

[0060] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A data-driven adaptive optimization method for lens hardening process, characterized in that: include: Real-time collection of multiple process parameters during lens hardening; calling a hardening process prediction model, predicting the hardness of the lens based on the plurality of process parameters, comparing the predicted lens hardness parameters with a predetermined hardness value, and generating a predicted hardness deviation; performing adaptive optimization of the plurality of process parameters based on the predicted hardness deviation to generate optimized process parameters; The optimized process parameters are fed back to the production control end of the lens hardening process to perform regulation of the lens hardening process parameters.

2. The data-driven adaptive optimization method for lens hardening process according to claim 1, characterized in that: Calling a hardening process prediction model, predicting the hardness of the lens based on the multiple process parameters, comparing the predicted lens hardness parameters with a predetermined hardness value, and generating a predicted hardness deviation, including: Collecting historical hardening process detection data sets and training the hardening process prediction model based on regression learning; Inputting the plurality of process parameters into the hardening process prediction model to predict the hardness of the lens, and outputting the predicted lens hardness parameters; The difference between the predicted lens hardness parameter and the predetermined hardness value is calculated to generate the predicted hardness deviation.

3. The data-driven adaptive optimization method for lens hardening process according to claim 1, characterized in that: Adaptively optimizing the plurality of process parameters based on the predicted hardness deviation to generate optimized process parameters includes: Determining whether the predicted hardness deviation exceeds a preset deviation threshold; If so, analyzing the influence weights of the predicted hardness deviation and the deviations of the plurality of process parameters, and constructing a variable optimization weight distribution; Based on the variable optimization weight distribution, the multiple process parameters are adaptively optimized to generate the optimized process parameters.

4. The data-driven adaptive optimization method for lens hardening process according to claim 3, characterized in that: Based on the variable optimization weight distribution, the plurality of process parameters are adaptively optimized to generate the optimized process parameters, including: Based on the variable optimization weight distribution and the predicted hardness deviation, initializing and adjusting the plurality of process parameters to generate a first adjustment parameter set and a second adjustment parameter set; Calling the hardening process prediction model to perform hardness prediction on the first adjustment parameter set and the second adjustment parameter set to generate a first hardness prediction result and a second hardness prediction result; Positioning the first hardness prediction result and the second hardness prediction result as hardness optimization directions and temporary optimal adjustment parameters; The temporary optimal adjustment parameters are continuously adjusted according to the hardness optimization direction, and the adjustment parameters are iterated multiple times until the adjustment parameters that meet the convergence conditions are obtained, thereby generating the optimized process parameters.

5. The data-driven adaptive optimization method for lens hardening process according to claim 4, characterized in that: After generating the first hardness prediction result and the second hardness prediction result, the method further includes: respectively calculating differences between the first hardness prediction result, the second hardness prediction result and the predetermined hardness value to generate a first difference value and a second difference value; If any difference between the first difference and the second difference meets the preset deviation threshold, the optimized process parameters are directly generated.

6. The data-driven adaptive optimization method for lens hardening process according to claim 3, characterized in that: Analyzing the influence weights of the predicted hardness deviation and the deviations of the plurality of process parameters, and constructing a variable optimization weight distribution, including: Based on a historical hardening process detection data set, using a statistical analysis method, evaluating the degree of influence of each process parameter on hardness deviation, and determining multiple influence weights of the multiple process parameters; The multiple influence weights are normalized so that the sum of the weights is 1, thereby generating the variable optimization weight distribution.

7. The data-driven adaptive optimization method for lens hardening process according to claim 4, characterized in that: Positioning the first hardness prediction result and the second hardness prediction result in a hardness optimization direction and a temporary optimal adjustment parameter includes: Comparing the first hardness prediction result and the second hardness prediction result, taking the adjustment parameter with smaller hardness as a starting point and the adjustment parameter with larger hardness as an end point, analyzing the change direction of each parameter from the starting point to the end point, and generating the hardness optimization direction; The first hardness prediction result and the second hardness prediction result are compared, and the temporary optimal adjustment parameter is generated using the adjustment parameter with greater hardness.

8. A data-driven adaptive optimization system for lens hardening processes, characterized in that: The system is used to execute the data-driven adaptive optimization method for lens hardening process according to any one of claims 1 to 7, comprising: Data acquisition module: real-time collection of multiple process parameters during lens hardening process; Hardness prediction module: calling the hardening process prediction model, predicting the hardness of the lens based on the multiple process parameters, comparing the predicted lens hardness parameters with the predetermined hardness value, and generating a predicted hardness deviation; Adaptive optimization module: performing adaptive optimization of the plurality of process parameters based on the predicted hardness deviation to generate optimized process parameters; Process parameter control module: feeds back the optimized process parameters to the production control end of the lens hardening process to execute the control of the lens hardening process parameters.

Images