Optical glass drop information prediction method and device, electronic device and storage medium

Abstract

The optical glass drop information prediction method and device, electronic device, and storage medium provided in this application relate to the technical field of optical glass drop information prediction. In this application, first, sample process data and performance data labels are obtained. Second, based on at least one associated data that affects the performance of the optical glass drop, multiple sample process data are classified to obtain at least one data classification set. Then, for each data classification set, a target drop information prediction model is trained based on each sample process data and corresponding performance data label in the data classification set. Finally, the target drop information prediction model is used to predict the target optical glass drop description data and output the target optical glass drop prediction data. Based on the above, the problem of high production costs of optical glass drops in the prior art can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of optical glass drop information prediction, and in particular to an optical glass drop information prediction method and device, electronic equipment, and storage medium. Background Art

[0002] In the field of optoelectronic materials, small-sized lenses made of optical glass with special optical properties are widely used in optical instruments or mechanical systems. With the rapid development of industries such as smartphones, computers, new energy vehicles, and drones with imaging functions in recent years, the market demand for small-sized lenses has also increased sharply. Optical glass droplets, as their preforms, can be used to obtain small-sized lenses with ideal optical properties through rapid and precise molding. However, the inventors have found that in the current production method of optical glass droplets, it is necessary to test the process at each stage through experiments and adjust the process parameters through trial and error to achieve the droplet size and shape requirements given by the user. In addition, the process adjustment involves parameters such as the temperature of each point in the droplet tube, the nozzle type of the droplet furnace, the nozzle length, the mold type, the cooling air temperature, intensity, time, and the droplet cycle. As a result, there is a problem of high production costs. Summary of the Invention

[0003] In view of this, the purpose of the present application is to provide a method and device for predicting optical glass gob information, an electronic device and a storage medium, so as to improve the problem of high production cost of optical glass gob in the prior art.

[0004] To achieve the above objectives, this application adopts the following technical solutions:

[0005] A method for predicting optical glass drop information, comprising:

[0006] Acquire sample process data corresponding to the optical glass gob and performance data tags corresponding to the optical glass gob, wherein the sample process data is a plurality of data, each of which is used to describe the production process of the corresponding optical glass gob, and the performance data tags is a plurality of data, each of which is used to reflect the actual performance of the corresponding optical glass gob;

[0007] Classifying the plurality of sample process data based on at least one associated data having an impact on the performance of the optical glass gob to obtain at least one data classification set corresponding to the plurality of sample process data, wherein each of the data classification sets includes at least one sample process data;

[0008] For each data classification set in the at least one data classification set, training a target drip information prediction model corresponding to the data classification set based on each sample process data in the data classification set and a performance data label corresponding to the sample process data, wherein the target drip information prediction model is a random forest model;

[0009] The target optical glass drop description data to be predicted is acquired, and the target optical glass drop description data is predicted using the target drop information prediction model, and target optical glass drop prediction data corresponding to the target optical glass drop description data is output.

[0010] In a preferred embodiment of the present application, in the above-mentioned optical glass drop information prediction method, the steps of obtaining target optical glass drop description data to be predicted, predicting the target optical glass drop description data using the target drop information prediction model, and outputting target optical glass drop prediction data corresponding to the target optical glass drop description data include:

[0011] Acquiring target optical glass drop description data to be predicted, wherein the target optical glass drop description data is performance data;

[0012] Predicting the target optical glass drop description data based on each target drop information prediction model, and outputting each target optical glass drop prediction data corresponding to the target optical glass drop description data; or

[0013] Based on a target drop information prediction model that matches the target optical glass drop description data, the target optical glass drop description data is predicted, and target optical glass drop prediction data corresponding to the target optical glass drop description data is output, wherein the target optical glass drop prediction data belongs to process data.

[0014] In a preferred embodiment of the present application, in the above-mentioned optical glass drop information prediction method, the step of predicting the target optical glass drop description data based on a target drop information prediction model that matches the target optical glass drop description data, and outputting target optical glass drop prediction data corresponding to the target optical glass drop description data, includes:

[0015] For each of the plurality of sample process data, which includes a plurality of process parameters, determining a plurality of candidate process parameter combinations, wherein each candidate process parameter combination includes a plurality of process parameters, and the number of process parameters included in every two candidate process parameter combinations is the same;

[0016] For each candidate process parameter combination, predict the candidate process parameter combination based on a target drop information prediction model that matches the target optical glass drop description data, and output predicted performance data corresponding to the candidate process parameter combination;

[0017] For each piece of predicted performance data, determining a first error corresponding to each performance parameter in the predicted performance data and a second error corresponding to the predicted performance data, wherein the first error is used to reflect the difference between the performance parameter in the predicted performance data and the performance parameter in the target optical glass gob description data, and the second error is used to reflect the deviation balance between the performance parameter in the predicted performance data and the performance parameter in the target optical glass gob description data;

[0018] Selecting, from among the predicted performance data, predicted performance data whose first errors corresponding to the respective performance parameters are less than a predetermined first threshold and whose second errors are less than a predetermined second threshold from among the predicted performance data as candidate predicted performance data;

[0019] For each candidate predicted performance data, summing the first errors of the performance parameters in the candidate predicted performance data to obtain a summed first error, and performing a weighted summation calculation on the summed first error and a second error corresponding to the candidate predicted performance data to obtain a target error corresponding to the candidate predicted performance data, wherein a weighting coefficient corresponding to the summed first error is greater than a weighting coefficient corresponding to the second error;

[0020] Among the candidate prediction performance data, the candidate prediction performance data with the minimum corresponding target error is screened out, and the candidate process parameters corresponding to the candidate prediction performance data are combined as the target optical glass drop prediction data corresponding to the target optical glass drop description data.

[0021] In a preferred embodiment of the present application, in the above optical glass drop information prediction method, the step of determining a plurality of candidate process parameter combinations for each of the plurality of sample process data including a plurality of process parameters comprises:

[0022] Determining the parameter magnitude of each process parameter based on each of the plurality of sample process data including a plurality of process parameters, and determining the value range corresponding to each process parameter based on the parameter magnitude of each process parameter;

[0023] For each process parameter, a random value processing is performed in the value range corresponding to the process parameter at the current stage to obtain the current random value of the process parameter;

[0024] Predicting a parameter combination formed by current random values ​​of each process parameter based on a target drop information prediction model that matches the target optical glass drop description data, outputting current random predicted performance data, and determining a random first error between the current random predicted performance data and the target optical glass drop description data;

[0025] determining whether a random first error between current random predicted performance data and the target optical glass gob description data is less than a random first error between the random predicted performance data and the target optical glass gob description data in a previous stage;

[0026] When a random first error between the current random predicted performance data and the target optical glass gob description data is less than a random first error between the random predicted performance data and the target optical glass gob description data in a previous stage, a parameter combination formed by the current random values ​​of each process parameter is used as a candidate process parameter combination;

[0027] When the random first error between the current random predicted performance data and the target optical glass drop description data is not less than the random first error between the random predicted performance data in the previous stage and the target optical glass drop description data, the acceptance probability of the current random predicted performance data is determined based on the random first error, and based on the acceptance probability, it is determined whether to use the parameter combination formed by the current process parameter random values ​​of each process parameter as a candidate process parameter combination.

[0028] In a preferred embodiment of the present application, in the above-mentioned optical glass drop information prediction method, the steps of obtaining target optical glass drop description data to be predicted, predicting the target optical glass drop description data using the target drop information prediction model, and outputting target optical glass drop prediction data corresponding to the target optical glass drop description data include:

[0029] Acquire target optical glass drop description data to be predicted, wherein the target optical glass drop description data belongs to process data;

[0030] Predicting the target optical glass drop description data based on each target drop information prediction model, and outputting each target optical glass drop prediction data corresponding to the target optical glass drop description data; or

[0031] Based on a target drop information prediction model that matches the target optical glass drop description data, the target optical glass drop description data is predicted, and target optical glass drop prediction data corresponding to the target optical glass drop description data is output, wherein the target optical glass drop prediction data belongs to performance data.

[0032] In a preferred embodiment of the present application, in the above optical glass gob information prediction method, the step of obtaining sample process data corresponding to the optical glass gob and performance data labels corresponding to the optical glass gob includes:

[0033] Obtaining sample process data corresponding to the optical glass drop material, wherein the sample process data includes at least one parameter of a nozzle length of a production device, a nozzle thickness of a production device, a mold curvature of a production device, a mold height of a production device, a mold thickness of a production device, temperatures at various locations of the production device, a cooling air temperature, a cooling air direction, a cooling air intensity, a cooling air time, and a drop material cycle;

[0034] A performance data tag corresponding to the optical glass gob is obtained, wherein the performance data tag includes at least one parameter of the optical glass gob, including weight, center thickness, outer diameter, roundness, and upper and lower diameters.

[0035] In a preferred embodiment of the present application, in the above-mentioned optical glass gob information prediction method, the step of classifying the plurality of sample process data based on at least one associated data having an impact on the performance of the optical glass gob to obtain at least one data classification set corresponding to the plurality of sample process data includes:

[0036] For each optical glass gob, obtaining a production location of the optical glass gob, and using the production location of the optical glass gob and a cooling air direction of the optical glass gob as at least one piece of associated data that has an impact on the performance of the optical glass gob;

[0037] Based on at least one associated data that has an impact on the performance of the optical glass gob, a plurality of sample process data corresponding to the plurality of optical glass gobs are classified to obtain at least one data classification set corresponding to the plurality of sample process data.

[0038] The present application also provides an optical glass drop information prediction device, comprising:

[0039] a sample data acquisition module, configured to acquire sample process data corresponding to the optical glass gob and performance data tags corresponding to the optical glass gob, wherein the sample process data may be multiple, each of which is used to describe the production process of the corresponding optical glass gob; and the performance data tags may be multiple, each of which is used to reflect the actual performance of the corresponding optical glass gob;

[0040] a sample data classification module, configured to classify the plurality of sample process data based on at least one associated data having an impact on the performance of the optical glass gob, to obtain at least one data classification set corresponding to the plurality of sample process data, wherein each of the data classification sets includes at least one sample process data;

[0041] a prediction model training module, configured to train, for each data classification set in the at least one data classification set, a target drip information prediction model corresponding to the data classification set based on each sample process data in the data classification set and a performance data label corresponding to the sample process data, wherein the target drip information prediction model is a random forest model;

[0042] The drop information prediction module is used to obtain target optical glass drop description data to be predicted, and use the target drop information prediction model to predict the target optical glass drop description data, and output target optical glass drop prediction data corresponding to the target optical glass drop description data.

[0043] Based on the above, the present application further provides an electronic device, including:

[0044] Memory for storing computer programs;

[0045] The processor connected to the memory is used to execute the computer program stored in the memory to implement the above-mentioned optical glass drop material information prediction method.

[0046] On the basis of the above, the present application further provides a computer-readable storage medium, in which a computer program is stored. When the computer program is run, each step of the above-mentioned optical glass drop material information prediction method is executed.

[0047] The optical glass drop information prediction method and device, electronic device and storage medium provided in the present application first obtain sample process data and performance data labels; secondly, based on at least one associated data that has an impact on the performance of the optical glass drop, multiple sample process data are classified to obtain at least one data classification set; then, for each data classification set, based on each sample process data and the corresponding performance data label in the data classification set, a target drop information prediction model is trained to form; finally, the target optical glass drop description data is predicted using the target drop information prediction model, and the target optical glass drop prediction data is output. Based on the above content, after the target drop information prediction model is trained to form, the trained target drop information prediction model can be used to directly perform corresponding predictions. Compared with the technical means in the prior art that require experimental trial and error, the solution of the present application can reduce the cost caused by experimental trial and error, and therefore can improve the problem of high production cost of optical glass drops in the prior art. In addition, efficiency can be improved. Moreover, since the target drip information prediction model is trained by classification, the reliability of the target drip information prediction model can be higher, thereby ensuring the reliability of the prediction data. In this way, the problem of product mismatch caused by poor reliability of the prediction data can be further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings.

[0049] Figure 1 This is a structural block diagram of the electronic device provided in an embodiment of the present application.

[0050] Figure 2 A schematic flow chart of a method for predicting optical glass drop information provided in an embodiment of the present application.

[0051] Figure 3 A block diagram of an optical glass drop information prediction device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Generally, the components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0053] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for protection, but merely represents selected embodiments of the present application. All other embodiments obtained by persons of ordinary skill in the art based on the embodiments in the present application without creative work are within the scope of protection of the present application.

[0054] like Figure 1 As shown, an embodiment of the present application provides an electronic device, wherein the electronic device may include a memory, a processor, and an optical glass drop information prediction device.

[0055] Specifically, the memory and the processor are electrically connected directly or indirectly to enable data transmission or interaction. For example, the memory and the processor can be electrically connected via one or more communication buses or signal lines. The optical glass drop information prediction device includes at least one software function module stored in the memory in the form of software or firmware. The processor is configured to execute an executable computer program stored in the memory, such as the software function module and computer program included in the optical glass drop information prediction device, to implement the optical glass drop information prediction method provided in the embodiments of the present application.

[0056] Optionally, the memory may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0057] Furthermore, the processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a system on chip (SoC), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0058] I understand. Figure 1The structure shown is only for illustration, and the electronic device may also include Figure 1 More or fewer components than shown, or with Figure 1 The different configurations shown may, for example, further include a communication unit for exchanging information with other devices.

[0059] Combine Figure 2 The present application also provides an optical glass drop information prediction method applicable to the above electronic device. The steps defined in the process related to the optical glass drop information prediction method can be implemented by the electronic device.

[0060] The following will Figure 2 The specific process shown is explained in detail.

[0061] Step S110 , obtaining sample process data corresponding to the optical glass gob and performance data labels corresponding to the optical glass gob.

[0062] In an embodiment of the present application, the electronic device can obtain sample process data corresponding to the optical glass drop and a performance data label corresponding to the optical glass drop. There are multiple sample process data, each of which is used to describe the production process of the corresponding optical glass drop, and there are multiple performance data labels, each of which is used to reflect the actual performance of the corresponding optical glass drop. There is a one-to-one correspondence between the multiple performance data labels and the multiple sample process data, that is, one optical glass drop has one sample process data and one performance data label. For example, for optical glass drop A, its production process can be reflected by sample process data 1, and its actual performance can be reflected by performance data label 1. For optical glass drop B, its production process can be reflected by sample process data 2, and its actual performance can be reflected by performance data label 2.

[0063] Step S120 : classifying the plurality of sample process data based on at least one associated data that has an impact on the performance of the optical glass gob, and obtaining at least one data classification set corresponding to the plurality of sample process data.

[0064] In an embodiment of the present application, the electronic device can classify the plurality of sample process data based on at least one associated data that affects the performance of the optical glass gob, thereby obtaining at least one data classification set corresponding to the plurality of sample process data. In this way, sample process data (or optical glass gob) that are similar in terms of the associated data can be classified into one set. The associated data can be the sample process data itself, data other than the sample process data, or a portion of the sample process data. Each data classification set includes at least one sample process data.

[0065] Step S130 , for each data classification set in the at least one data classification set, based on each sample process data in the data classification set and the performance data label corresponding to the sample process data, training a target drip information prediction model corresponding to the data classification set.

[0066] In an embodiment of the present application, the electronic device can train a target drip information prediction model corresponding to each data classification set in the at least one data classification set based on each sample process data in the data classification set and the performance data label corresponding to the sample process data. The target drip information prediction model is a random forest model. Because a target drip information prediction model is trained based on sample process data from the same set, i.e., the training basis has a high degree of similarity in the dimensions of the associated data, the degree of interference with model training caused by differences in the associated data can be reduced, thereby protecting the model's fitting performance and enabling reliable predictions.

[0067] Step S140 , obtaining target optical glass drop description data to be predicted, and using the target drop information prediction model to predict the target optical glass drop description data, and outputting target optical glass drop prediction data corresponding to the target optical glass drop description data.

[0068] In an embodiment of the present application, after training and forming a target drop information prediction model, the electronic device can obtain target optical glass drop description data to be predicted, and use the target drop information prediction model to predict the target optical glass drop description data, outputting target optical glass drop prediction data corresponding to the target optical glass drop description data. It is understood that the prediction can be based on all target drop information prediction models or on a single matching target drop information prediction model. A matching target drop information prediction model refers to a matching target drop information prediction model (e.g., a high degree of similarity) on the dimension to which the associated data belongs.

[0069] Based on the above, after training and forming a target drop information prediction model, the trained target drop information prediction model can be used to directly perform corresponding predictions. Compared to the technical means required for trial and error in the prior art, the solution of the present application can reduce the costs associated with trial and error, thereby improving the high production costs of optical glass drop materials in the prior art. In addition, efficiency can be improved. Moreover, because the target drop information prediction model is trained based on classification, the reliability of the target drop information prediction model can be higher, ensuring the reliability of the prediction data. This can further improve the problem of product mismatch caused by poorly reliable prediction data, thereby avoiding product waste and other problems (and also reducing costs).

[0070] First, it should be noted that for step S110 , the specific method of obtaining the sample process data and performance data labels is not limited and can be selected according to actual needs.

[0071] For example, in an alternative embodiment, in order to ensure that a reliable target drip information prediction model can be trained based on the acquired sample process data and performance data labels, the above step S110 may further include the following content:

[0072] First, sample process data corresponding to the optical glass drop material can be obtained, wherein the sample process data includes at least one parameter of the nozzle length of the production equipment, the nozzle thickness of the production equipment, the mold curvature of the production equipment, the mold height of the production equipment, the mold thickness of the production equipment, the temperature at various locations of the production equipment, the cooling air temperature, the cooling air direction, the cooling air intensity, the cooling air time, and the drop material cycle. In other embodiments, other parameters may also be included;

[0073] Secondly, the performance data tag corresponding to the optical glass drop can be obtained, wherein the performance data tag includes at least one parameter of the optical glass drop, including weight, center thickness, outer diameter, roundness and upper and lower diameters. In other embodiments, other parameters may also be included.

[0074] It should be noted that after obtaining the sample process data corresponding to multiple optical glass drop materials, the sample process data can also be processed accordingly. For example, missing value checking and KNN nearest neighbor filling, 3σ (3 times the standard deviation) principle can be used to eliminate outliers, and data standardization can be used to reduce the dimensional influence between each feature.

[0075] Secondly, it should be noted that for step S120 , the specific method of classifying the plurality of sample process data based on at least one associated data having an impact on the performance of the optical glass gob is not limited and can be selected accordingly according to actual needs.

[0076] For example, in an alternative embodiment, in order to eliminate system errors caused by changes in different workstations and cooling air angles and ensure the fitting accuracy of the trained model, the above step S120 may further include the following:

[0077] First, for each optical glass gob, the production location of the optical glass gob can be obtained, and the production location of the optical glass gob and the cooling air direction of the optical glass gob are used as at least one piece of associated data that affects the performance of the optical glass gob; that is, there are two pieces of associated data, namely, the production location and the cooling air direction;

[0078] Secondly, based on at least one associated data that has an impact on the performance of the optical glass drop, multiple sample process data corresponding to the multiple optical glass drops are classified to obtain at least one data classification set corresponding to the multiple sample process data; exemplarily, classification can be performed based on the KNN algorithm, wherein the K value in the classification process can be 2-4; based on this, the sample process data in the same data classification set can be similar in the dimensions of production position and cooling air direction, so that the corresponding target drop information prediction model can eliminate corresponding interference.

[0079] Thirdly, it should be noted that for step S130, the specific method of training the target drip information prediction model corresponding to the data classification set based on each sample process data and the performance data label corresponding to the sample process data is not limited and can be selected according to actual needs.

[0080] For example, in an alternative embodiment, to ensure that the trained target drip information prediction model has high reliability, that is, to enable relatively reliable drip information prediction, during the training process, the data classification set can be divided into a training set and a test set at an 80%-20% ratio, and the set hyperparameters include the number of random forest decision trees, the maximum number of features, the maximum depth, the minimum number of sample splits, and the minimum number of sample leaves. Hyperparameter optimization can utilize a random grid search algorithm, with the mean squared error (MSE) selected as the evaluation metric. The optimal parameters are then used to train the random forest model and predict using the test set. Furthermore, the model performance can be evaluated using the evaluation metrics R2 (coefficient of determination) and RMSE. Specifically, the RMSE error for weight, center thickness, outer diameter, and roundness should not exceed 3% of the mean, with the upper and lower diameters determined by product requirements.

[0081] Fourthly, it should be noted that for step S140 , the specific method of using the target drop information prediction model to predict the target optical glass drop description data is not limited and can be selected according to actual needs.

[0082] For example, in an alternative embodiment, when performance data prediction is required based on process data, the above step S140 may further include the following:

[0083] First, description data of the target optical glass drop to be predicted can be obtained, wherein the description data of the target optical glass drop is process data, such as nozzle length of the production equipment, nozzle thickness of the production equipment, mold curvature of the production equipment, mold height of the production equipment, mold thickness of the production equipment, temperature at various locations of the production equipment, cooling air temperature, cooling air direction, cooling air intensity, cooling air time, and drop cycle, etc.;

[0084] Secondly, the target optical glass drop description data can be predicted based on each of the target drop information prediction models, and each target optical glass drop prediction data corresponding to the target optical glass drop description data can be output; or, the target optical glass drop description data can be predicted based on a target drop information prediction model that matches the target optical glass drop description data (such as the corresponding production position and cooling air direction are similar or consistent), and the target optical glass drop prediction data corresponding to the target optical glass drop description data can be output, wherein the target optical glass drop prediction data is performance data, such as the weight, middle thickness, outer diameter, roundness and upper and lower diameters of the optical glass drop.

[0085] For another example, in another alternative embodiment, when process data prediction is required based on performance data, the above-mentioned step S140 may further include step S141, step S142 and step S143, and the specific content of each step is described as follows.

[0086] Step S141 , obtaining description data of the target optical glass drop to be predicted.

[0087] In the embodiment of the present application, description data of the target optical glass gob to be predicted can be obtained, wherein the description data of the target optical glass gob is performance data, such as weight, center thickness, outer diameter, roundness, and upper and lower diameters of the optical glass gob.

[0088] Step S142 : predicting the target optical glass drop description data based on each target drop information prediction model, and outputting each target optical glass drop prediction data corresponding to the target optical glass drop description data.

[0089] In an embodiment of the present application, after obtaining target optical glass drop description data, the target optical glass drop description data can be predicted based on each target drop information prediction model, and each target optical glass drop prediction data corresponding to the target optical glass drop description data can be output. The target optical glass drop prediction data is process data, such as the nozzle length of the production equipment, the nozzle thickness of the production equipment, the mold curvature of the production equipment, the mold height of the production equipment, the mold thickness of the production equipment, the temperature at various locations of the production equipment, the cooling air temperature, the cooling air direction, the cooling air intensity, the cooling air time, and the drop cycle. In other words, one target drop information prediction model outputs one target optical glass drop prediction data. When multiple target drop information prediction models exist, multiple target optical glass drop prediction data can be output.

[0090] Step S143 , predicting the target optical glass drop description data based on a target drop information prediction model that matches the target optical glass drop description data, and outputting target optical glass drop prediction data corresponding to the target optical glass drop description data.

[0091] In an embodiment of the present application, after obtaining target optical glass gob description data, a target gob information prediction model that matches the target optical glass gob description data (e.g., a similar or consistent corresponding production location and cooling air direction) can be used to predict the target optical glass gob description data and output target optical glass gob prediction data corresponding to the target optical glass gob description data. The target optical glass gob prediction data is process data.

[0092] It is understandable that in the above-mentioned step S142 and step S143, the prediction method can be consistent, and the specific prediction method is not limited and can be selected according to actual needs.

[0093] For example, in an alternative embodiment, in order to reliably predict the target optical glass drop prediction data (i.e., reliably predict the process data based on the performance data), the above-mentioned step S143 may further include step S143a, step S143b, step S143c, step S143d, step S143e and step S143f, and the specific content of each step is described below.

[0094] Step S143a: for each of the plurality of sample process data including a plurality of process parameters, determine a plurality of candidate process parameter combinations.

[0095] In an embodiment of the present application, multiple candidate process parameter combinations can be determined for each of the multiple sample process data, each of which includes multiple process parameters. Each candidate process parameter combination includes multiple process parameters, and the number of process parameters included between every two candidate process parameter combinations is the same, such as the nozzle length of the production equipment, the nozzle thickness of the production equipment, the mold curvature of the production equipment, the mold height of the production equipment, the mold thickness of the production equipment, the temperature at various locations of the production equipment, the cooling air temperature, the cooling air direction, the cooling air intensity, the cooling air time, and the dripping cycle. However, the specific values ​​of at least some of the process parameters may be different.

[0096] Step S143b: For each candidate process parameter combination, based on a target drop information prediction model that matches the target optical glass drop description data, predict the candidate process parameter combination and output predicted performance data corresponding to the candidate process parameter combination.

[0097] In an embodiment of the present application, after determining multiple candidate process parameter combinations, each candidate process parameter combination can be predicted based on a target drop information prediction model that matches the target optical glass drop description data, and the predicted performance data corresponding to the candidate process parameter combination can be output, that is, the performance data is predicted based on the process data.

[0098] Step S143c: for each piece of predicted performance data, respectively determine a first error corresponding to each performance parameter in the predicted performance data and a second error corresponding to the predicted performance data.

[0099] In an embodiment of the present application, after outputting the predicted performance data, a first error corresponding to each performance parameter in the predicted performance data and a second error corresponding to the predicted performance data can be determined for each of the predicted performance data. The first error is used to reflect the difference between the performance parameter in the predicted performance data and the performance parameter in the target optical glass gob description data (e.g., the difference between the weight in the predicted performance data and the weight in the target optical glass gob description data, the difference between the middle thickness in the predicted performance data and the middle thickness in the target optical glass gob description data, etc.), and the second error is used to reflect the deviation balance between the performance parameter in the predicted performance data and the performance parameter in the target optical glass gob description data (e.g., first determining the difference between the weight in the predicted performance data and the weight in the target optical glass gob description data, the difference between the middle thickness in the predicted performance data and the middle thickness in the target optical glass gob description data, and performing variance calculation on the differences corresponding to the performance parameters to obtain the second error).

[0100] Step S143d: Filter out, from among the predicted performance data, the predicted performance data whose first errors corresponding to the performance parameters are smaller than a predetermined first threshold and whose corresponding second errors are smaller than a predetermined second threshold, as candidate predicted performance data.

[0101] In an embodiment of the present application, after determining the first error and the second error, the prediction performance data in which the first error corresponding to each performance parameter is less than a predetermined first threshold value (such as the percentage of the first error is less than 5%) and the corresponding second error is less than a predetermined second threshold value (such as the percentage of the second error is less than 10%) can be screened out from each of the prediction performance data as candidate prediction performance data, that is, some prediction performance data with relatively small errors are first screened out.

[0102] Step S143e, for each candidate prediction performance data, the first errors of the performance parameters in the candidate prediction performance data are summed to obtain the summed first error, and the summed first error and the second error corresponding to the candidate prediction performance data are weighted and summed to obtain the target error corresponding to the candidate prediction performance data.

[0103] In an embodiment of the present application, after candidate predicted performance data are screened, the first errors of the performance parameters in each candidate predicted performance data may be summed for each candidate predicted performance data to obtain a summed first error, and the summed first error and the second error corresponding to the candidate predicted performance data may be weighted and summed to obtain a target error corresponding to the candidate predicted performance data. The weighting coefficient corresponding to the summed first error is greater (and may be much greater) than the weighting coefficient corresponding to the second error. In other words, the target error may place greater emphasis on the first error, that is, on the difference from the required performance.

[0104] Step S143f, screening out the candidate prediction performance data whose corresponding target error has the minimum value among the candidate prediction performance data, and combining the candidate process parameters corresponding to the candidate prediction performance data as the target optical glass drop prediction data corresponding to the target optical glass drop description data.

[0105] In an embodiment of the present application, after obtaining the target error, the candidate prediction performance data with the minimum corresponding target error can be screened out from each of the candidate prediction performance data, and the candidate process parameters corresponding to the candidate prediction performance data can be combined as the target optical glass drop prediction data corresponding to the target optical glass drop description data.

[0106] It is understood that in the above step S143a, the specific method of determining multiple candidate process parameter combinations is not limited and can be selected according to actual needs. For example, in an alternative embodiment, in order to ensure the reliability of the determined candidate process parameter combinations, the above step S143a may further include the following:

[0107] First, based on the fact that each of the sample process data includes multiple process parameters, the parameter magnitude of each process parameter (i.e., the magnitude of the parameter, such as 1, 10, 100, etc.) can be determined, and based on the parameter magnitude of each process parameter, the value range corresponding to each process parameter (e.g., [0.8 ~1.2 ],in, Indicates the parameter magnitude of a process parameter);

[0108] Secondly, for each process parameter, a random value processing can be performed in the value range corresponding to the process parameter at the current stage to obtain the current random value of the process parameter. For example, 500 values ​​can be randomly selected in the value range;

[0109] Then, based on a target drop information prediction model that matches the target optical glass drop description data, a parameter combination formed by the current random values ​​of each process parameter (one combination per stage) can be predicted, and the current random predicted performance data can be output. In addition, a random first error between the current random predicted performance data and the target optical glass drop description data can be determined (i.e., the difference between the performance parameters in the random predicted performance data and the performance parameters in the target optical glass drop description data, i.e., the prediction error. It should be noted that the "random" in "random first error" is only used to distinguish it from the first error mentioned above, and refers to the error corresponding to the random predicted performance data, and does not have any other limiting effect).

[0110] Afterwards, it can be determined whether a random first error between the current random predicted performance data and the target optical glass gob description data is smaller than a random first error between the random predicted performance data and the target optical glass gob description data in a previous stage, that is, whether the current random first error is smaller than the random first error in the previous stage, that is, whether the random first error is decreasing;

[0111] Furthermore, when a random first error between the current random predicted performance data and the target optical glass gob description data is smaller than a random first error between the random predicted performance data and the target optical glass gob description data in a previous stage, a parameter combination formed by the current random value of each process parameter may be used as a candidate process parameter combination;

[0112] Finally, when the random first error between the current random predicted performance data and the target optical glass drop description data is not less than the random first error between the random predicted performance data in the previous stage and the target optical glass drop description data, the acceptance probability of the current random predicted performance data can be determined based on the random first error, and based on the acceptance probability, it can be determined whether to use the parameter combination formed by the current random value of each process parameter as a candidate process parameter combination. In this way, the determination of one stage can be completed and then enter the next stage until the candidate process parameter combinations that meet the requirements are determined, such as 1000; wherein, the acceptance probability , Represents the random first error, k is an adjustment parameter that can be selected according to actual needs. () is an exponential operation related to the natural logarithm base e.

[0113] Combine Figure 3 The present application also provides an optical glass drop information prediction device applicable to the above-mentioned electronic device. The optical glass drop information prediction device includes a sample data acquisition module, a sample data classification module, a prediction model training module, and a drop information prediction module.

[0114] In detail, the sample data acquisition module can be used to obtain sample process data corresponding to the optical glass gob and performance data labels corresponding to the optical glass gob, wherein the sample process data is multiple, each of which is used to describe the production process of the corresponding optical glass gob, and the performance data labels are multiple, each of which is used to reflect the actual performance of the corresponding optical glass gob. In the embodiment of the present application, the sample data acquisition module can be used to perform Figure 2 Regarding step S110 shown, the relevant contents of the sample data acquisition module can refer to the above description of step S110.

[0115] In detail, the sample data classification module can be used to classify the plurality of sample process data based on at least one associated data that has an impact on the performance of the optical glass gob, and obtain at least one data classification set corresponding to the plurality of sample process data, wherein each data classification set includes at least one sample process data. In the embodiment of the present application, the sample data classification module can be used to perform Figure 2 As shown in step S120, for the relevant content of the sample data classification module, reference can be made to the above description of step S120.

[0116] In detail, the prediction model training module can be used to train a target drip information prediction model corresponding to each data classification set in the at least one data classification set based on each sample process data in the data classification set and the performance data label corresponding to the sample process data, wherein the target drip information prediction model belongs to a random forest model. In the embodiment of the present application, the prediction model training module can be used to perform Figure 2 As shown in step S130, for the relevant content of the prediction model training module, reference can be made to the description of step S130 above.

[0117] In detail, the drop information prediction module can be used to obtain the target optical glass drop description data to be predicted, and use the target drop information prediction model to predict the target optical glass drop description data, and output the target optical glass drop prediction data corresponding to the target optical glass drop description data. In the embodiment of the present application, the drop information prediction module can be used to perform Figure 2 Regarding step S140 shown, the relevant contents of the drip information prediction module can refer to the above description of step S140.

[0118] In an embodiment of the present application, corresponding to the optical glass drop information prediction method applied to the electronic device, a computer-readable storage medium is also provided. The computer-readable storage medium stores a computer program that, when executed, executes each step of the optical glass drop information prediction method. The steps executed when the computer program is executed are not described in detail here; reference is made to the above explanation of the optical glass drop information prediction method.

[0119] In summary, the optical glass drop information prediction method and device, electronic device and storage medium provided by the present application first obtain sample process data and performance data labels; secondly, based on at least one associated data that has an impact on the performance of the optical glass drop, multiple sample process data are classified to obtain at least one data classification set; then, for each data classification set, based on each sample process data and the corresponding performance data label in the data classification set, a target drop information prediction model is trained to form; finally, the target optical glass drop description data is predicted using the target drop information prediction model, and the target optical glass drop prediction data is output. Based on the above content, after the target drop information prediction model is trained to form, the target drop information prediction model formed by training can be used to directly perform corresponding predictions. Compared with the technical means in the prior art that require experimental trial and error, the solution of the present application can reduce the cost caused by experimental trial and error. Therefore, it can improve the problem of high production cost of optical glass drops in the prior art. In addition, efficiency can be improved. Moreover, since the target drip information prediction model is trained by classification, the reliability of the target drip information prediction model can be higher, thereby ensuring the reliability of the prediction data. In this way, the problem of product mismatch caused by poor reliability of the prediction data can be further improved, thereby avoiding product waste and other problems (and also reducing costs).

[0120] In the several embodiments provided in the embodiments of the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device and method embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions and operations of the devices, methods and computer program products according to the multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, program segment or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0121] In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0122] If the functions are implemented in the form of software modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, electronic device, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk. It should be noted that, in this document, the terms "comprise," "include," or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article, or device. Without further constraints, an element defined by the phrase "comprises a..." does not preclude the existence of additional identical elements in the process, method, article or apparatus that includes the element.

[0123] The above description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A method for predicting optical glass drop information, characterized in that: include: Acquire sample process data corresponding to the optical glass gob and performance data tags corresponding to the optical glass gob, wherein the sample process data is a plurality of data, each of which is used to describe the production process of the corresponding optical glass gob, and the performance data tags is a plurality of data, each of which is used to reflect the actual performance of the corresponding optical glass gob; Classifying the plurality of sample process data based on at least one associated data having an impact on the performance of the optical glass gob to obtain at least one data classification set corresponding to the plurality of sample process data, wherein each of the data classification sets includes at least one sample process data; For each data classification set in the at least one data classification set, training a target drip information prediction model corresponding to the data classification set based on each sample process data in the data classification set and a performance data label corresponding to the sample process data, wherein the target drip information prediction model is a random forest model; Obtaining target optical glass drop description data to be predicted, and using the target drop information prediction model to predict the target optical glass drop description data, and outputting target optical glass drop prediction data corresponding to the target optical glass drop description data, including: Acquire target optical glass drop description data to be predicted, wherein the target optical glass drop description data belongs to process data; Predicting the target optical glass drop description data based on each target drop information prediction model, and outputting each target optical glass drop prediction data corresponding to the target optical glass drop description data; or Based on a target drop information prediction model that matches the target optical glass drop description data, the target optical glass drop description data is predicted, and target optical glass drop prediction data corresponding to the target optical glass drop description data is output, wherein the target optical glass drop prediction data belongs to performance data.

2. The optical glass drop information prediction method according to claim 1, characterized in that: The step of obtaining target optical glass drop description data to be predicted, predicting the target optical glass drop description data using the target drop information prediction model, and outputting target optical glass drop prediction data corresponding to the target optical glass drop description data includes: Acquiring target optical glass drop description data to be predicted, wherein the target optical glass drop description data is performance data; Predicting the target optical glass drop description data based on each target drop information prediction model, and outputting each target optical glass drop prediction data corresponding to the target optical glass drop description data; or Based on a target drop information prediction model that matches the target optical glass drop description data, the target optical glass drop description data is predicted, and target optical glass drop prediction data corresponding to the target optical glass drop description data is output, wherein the target optical glass drop prediction data belongs to process data.

3. The optical glass drop information prediction method according to claim 2, characterized in that: The step of predicting the target optical glass drop description data based on a target drop information prediction model that matches the target optical glass drop description data, and outputting target optical glass drop prediction data corresponding to the target optical glass drop description data, comprises: For each of the plurality of sample process data, which includes a plurality of process parameters, determining a plurality of candidate process parameter combinations, wherein each candidate process parameter combination includes a plurality of process parameters, and the number of process parameters included in every two candidate process parameter combinations is the same; For each candidate process parameter combination, predict the candidate process parameter combination based on a target drop information prediction model that matches the target optical glass drop description data, and output predicted performance data corresponding to the candidate process parameter combination; For each piece of predicted performance data, determining a first error corresponding to each performance parameter in the predicted performance data and a second error corresponding to the predicted performance data, wherein the first error is used to reflect the difference between the performance parameter in the predicted performance data and the performance parameter in the target optical glass gob description data, and the second error is used to reflect the deviation balance between the performance parameter in the predicted performance data and the performance parameter in the target optical glass gob description data; Screening out, from among the predicted performance data, predicted performance data whose first errors corresponding to the respective performance parameters are less than a predetermined first threshold and whose second errors are less than a predetermined second threshold, as candidate predicted performance data; For each candidate predicted performance data, summing the first errors of the performance parameters in the candidate predicted performance data to obtain a summed first error, and performing a weighted summation calculation on the summed first error and a second error corresponding to the candidate predicted performance data to obtain a target error corresponding to the candidate predicted performance data, wherein a weighting coefficient corresponding to the summed first error is greater than a weighting coefficient corresponding to the second error; Among the candidate prediction performance data, the candidate prediction performance data with the minimum corresponding target error is screened out, and the candidate process parameter combination corresponding to the candidate prediction performance data is used as the target optical glass drop prediction data corresponding to the target optical glass drop description data.

4. The optical glass drop information prediction method according to claim 3, characterized in that: The step of determining a plurality of candidate process parameter combinations for each of the plurality of sample process data including a plurality of process parameters comprises: Determining the parameter magnitude of each process parameter based on each of the plurality of sample process data including a plurality of process parameters, and determining the value range corresponding to each process parameter based on the parameter magnitude of each process parameter; For each process parameter, a random value processing is performed in the value range corresponding to the process parameter at the current stage to obtain the current random value of the process parameter; Predicting a parameter combination formed by current random values ​​of each process parameter based on a target drop information prediction model that matches the target optical glass drop description data, outputting current random predicted performance data, and determining a random first error between the current random predicted performance data and the target optical glass drop description data; determining whether a random first error between current random predicted performance data and the target optical glass gob description data is less than a random first error between the random predicted performance data and the target optical glass gob description data in a previous stage; When a random first error between the current random predicted performance data and the target optical glass gob description data is less than a random first error between the random predicted performance data and the target optical glass gob description data in a previous stage, a parameter combination formed by the current random values ​​of each process parameter is used as a candidate process parameter combination; When the random first error between the current random predicted performance data and the target optical glass drop description data is not less than the random first error between the random predicted performance data in the previous stage and the target optical glass drop description data, the acceptance probability of the current random predicted performance data is determined based on the random first error, and based on the acceptance probability, it is determined whether to use the parameter combination formed by the current process parameter random values ​​of each process parameter as a candidate process parameter combination.

5. The optical glass drop information prediction method according to any one of claims 1 to 4, characterized in that: The step of obtaining sample process data corresponding to the optical glass gob and performance data labels corresponding to the optical glass gob includes: Obtaining sample process data corresponding to the optical glass drop material, wherein the sample process data includes at least one parameter of a nozzle length of a production device, a nozzle thickness of a production device, a mold curvature of a production device, a mold height of a production device, a mold thickness of a production device, temperatures at various locations of the production device, a cooling air temperature, a cooling air direction, a cooling air intensity, a cooling air time, and a drop material cycle; A performance data tag corresponding to the optical glass gob is obtained, wherein the performance data tag includes at least one parameter of the optical glass gob, including weight, center thickness, outer diameter, roundness, and upper and lower diameters.

6. The optical glass drop information prediction method according to any one of claims 1 to 4, characterized in that: The step of classifying the plurality of sample process data based on at least one associated data having an impact on the performance of the optical glass gob to obtain at least one data classification set corresponding to the plurality of sample process data includes: For each optical glass gob, obtaining a production location of the optical glass gob, and using the production location of the optical glass gob and a cooling air direction of the optical glass gob as at least one piece of associated data that has an impact on the performance of the optical glass gob; Based on at least one associated data that has an impact on the performance of the optical glass gob, a plurality of sample process data corresponding to the plurality of optical glass gobs are classified to obtain at least one data classification set corresponding to the plurality of sample process data.

7. An optical glass drop information prediction device, characterized in that: include: a sample data acquisition module, configured to acquire sample process data corresponding to the optical glass gob and performance data tags corresponding to the optical glass gob, wherein the sample process data may be multiple, each of which is used to describe the production process of the corresponding optical glass gob; and the performance data tags may be multiple, each of which is used to reflect the actual performance of the corresponding optical glass gob; a sample data classification module, configured to classify the plurality of sample process data based on at least one associated data having an impact on the performance of the optical glass gob, to obtain at least one data classification set corresponding to the plurality of sample process data, wherein each of the data classification sets includes at least one sample process data; a prediction model training module, configured to train, for each data classification set in the at least one data classification set, a target drip information prediction model corresponding to the data classification set based on each sample process data in the data classification set and a performance data label corresponding to the sample process data, wherein the target drip information prediction model is a random forest model; The drop information prediction module is used to obtain target optical glass drop description data to be predicted, and use the target drop information prediction model to predict the target optical glass drop description data, and output target optical glass drop prediction data corresponding to the target optical glass drop description data, including: Acquire target optical glass drop description data to be predicted, wherein the target optical glass drop description data belongs to process data; Predicting the target optical glass drop description data based on each target drop information prediction model, and outputting each target optical glass drop prediction data corresponding to the target optical glass drop description data; or Based on a target drop information prediction model that matches the target optical glass drop description data, the target optical glass drop description data is predicted, and target optical glass drop prediction data corresponding to the target optical glass drop description data is output, wherein the target optical glass drop prediction data belongs to performance data.

8. An electronic device, characterized in that: include: memory for storing computer programs; A processor connected to the memory is used to execute the computer program stored in the memory to implement the optical glass drop information prediction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is run, the optical glass drop material information prediction method according to any one of claims 1 to 6 is executed.

Images