GLASS MATERIAL MANUFACTURING APPARATUS, GLASS MATERIAL MANUFACTURING METHOD, AND MATERIAL PROFILE PREDICTION METHOD

Abstract

One aspect of the present disclosure makes it possible to predict the refractive index profile of a transparent glass base material which is obtained in the process of the production of a glass microparticle accumulated body by a VAD method. The glass base material production apparatus comprises a gas supply system, a burner and a signal processing device. The signal processing device is equipped with an imaging device for taking an image of a particle flow composed of glass microparticles and an arithmetic unit. The arithmetic unit extracts image data showing the state of at least flame or a particle flow from the image taken by the imaging device at one or a plurality of arbitrary points of time during a period from the start of the production of a glass microparticle accumulated body to the end of the production of the glass microparticle accumulated body, and regressively predicts the refractive index profile of a transparent glass base material, which is a desired variable, from explanatory variables including the data on the image.

Description

Technical Field

[0001] The present disclosure relates to a glass base material manufacturing apparatus, a glass base material manufacturing method, and a base material profile prediction method. This application claims priority based on Japanese Patent Application No. 2020-088009 filed on May 20, 2020, relies on its content, and incorporates it herein by reference in its entirety.

Background Art

[0002] Patent Document 1 discloses monitoring the deposition surface shape of a glass fine particle deposit (which finally constitutes a part of an optical fiber base material) by the VAD (Vaper-phase Axial Deposition) method, controlling the pulling speed so as to obtain a target shape, controlling at least any one of the concentration and flow rate of a gas containing a glass raw material or the like and the burner position together with the control of the pulling speed, and stopping the deposition operation of the glass fine particles when deviating from the target shape.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

[0004] The glass preform manufacturing apparatus according to an embodiment of the present disclosure is an apparatus for manufacturing a glass fine particle deposit by the VAD method. In order to achieve the above object, it includes a gas supply system, a burner, and a profile prediction system. The gas supply system separately supplies a glass raw material gas and a flame generation gas (fuel gas). The burner generates glass fine particles from the glass raw material gas in a flame obtained by burning the fuel gas supplied from the gas supply system, and sprays the glass fine particles in the flame onto the glass fine particle deposit. The profile prediction system outputs a prediction result of the refractive index profile of a transparent glass preform obtained by dehydrating and sintering the glass fine particle deposit at one or more arbitrary times during the period from the start to the end of the manufacture of the glass fine particle deposit. In particular, the profile prediction system includes an imaging device and a calculation unit. The imaging device images a flame sprayed from the burner onto the glass fine particle deposit or a particle flow composed of glass fine particles generated in the flame. The calculation unit performs image processing for extracting at least image data representing the state of the flame or the particle flow from the image obtained by the imaging device. Further, the calculation unit performs regression prediction of the refractive index profile of the transparent glass preform, which is the target variable, from explanatory variables including at least the image data.

Brief Description of the Drawings

[0005]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Embodiments for Carrying Out the Invention

[0006] [Problems to be Solved by the Present Disclosure] As a result of studying the above-described prior art, the inventors have found the following problems. That is, in the glass base material manufacturing method disclosed in Patent Document 1, although the deposition shape of the glass fine particle deposit by the VAD method is monitored, it is not possible to predict the refractive index profile of the transparent glass base material obtained by dehydrating and sintering the glass fine particle deposit. Conventionally, since the refractive index profile of the transparent glass base material obtained from the glass fine particle deposit including the portion to be the core of the optical fiber as the final product greatly affects the optical characteristics of the optical fiber, profile measurement is performed before the next process is executed.

[0007] FIG. 1 is a diagram showing operations from a conventional glass base material manufacturing method (deposition process) to profile measurement. In the conventional glass base material manufacturing method, in the deposition process, a core burner 17 is arranged on the center side of the glass fine particle deposit 14, and a cladding burner 18 is arranged on the outer peripheral side, and glass fine particles are sprayed from each of them. Then, the transparent glass base material 140 is obtained by heating the glass fine particle deposit 14 with the heater 141. The obtained transparent glass base material 140 corresponds to the core of the optical fiber as the final product and the portion corresponding to the optical cladding around the core, and becomes a part that directly affects the optical characteristics of the optical fiber. At a plurality of locations along the longitudinal direction of the obtained transparent glass base material, the refractive index profile in the radial direction is measured (profile measurement). As a result of this profile measurement, the transparent glass base material determined to be qualified proceeds to the next process for manufacturing. On the other hand, if the transparent glass base material to be measured is determined to be defective as a result of the profile measurement, the manufacturing conditions are adjusted in the deposition process (feedback). However, usually, several days have passed from the start time of manufacturing the glass fine particle deposit 14 to the end time of the profile measurement. If it is determined that the transparent glass base material 140 for which the profile has been measured is defective, all of the glass products manufactured during these several days, that is, all or part of the glass fine particle deposit 14 and the transparent glass base material 140 manufactured before the manufacturing conditions were adjusted may be determined to be defective products. Thus, it has been difficult to improve the manufacturing yield in the glass base material manufacturing method according to the above-described prior art.

[0008] The present disclosure has been made to solve the problems described above, and aims to provide a glass base material manufacturing apparatus, a glass base material manufacturing method, and a base material profile prediction method that make it possible to predict the refractive index profile of a transparent glass base material, including the portion that is to become the core of the final product, the optical fiber, at the pre-sintering stage of manufacturing the glass particulate deposit.

[0009] [Effects of this disclosure] According to various embodiments of the present disclosure, it is possible to predict the profile of the transparent glass preform over the entire length of the transparent glass preform. Furthermore, by predicting the profile during the manufacture of such a soot glass deposit, it is possible to change the manufacturing conditions, and it is possible to effectively suppress the occurrence of characteristic defects due to structural defects in the final product, that is, the optical fiber.

[0010] [Description of the embodiments of the present disclosure] First, the contents of the embodiments of the present disclosure will be individually listed and described.

[0011] (1) A glass base material manufacturing apparatus according to an embodiment of the present disclosure is an apparatus for manufacturing a soot glass deposit (including a portion to become the core of an optical fiber, which is a final product) by the VAD method, and in one aspect thereof, includes a gas supply system, a burner, and a profile prediction system. The gas supply system separately supplies a glass raw material gas and a flame generating gas (fuel gas). The burner generates glass fine particles from the glass raw material gas in a flame obtained by burning the fuel gas supplied from the gas supply system, and sprays the glass fine particles generated in the flame onto the soot glass deposit. The profile prediction system outputs a prediction result of a refractive index profile of a transparent glass base material obtained by dehydrating and sintering the soot glass deposit at one or more arbitrary points in time during the period from the start of manufacturing the soot glass deposit to the end of manufacturing the soot glass deposit.

[0012] In particular, the profile prediction system includes an imaging device and a computing unit. The imaging device captures an image of a flame sprayed from a burner onto the glass microparticle deposit or a particle flow composed of glass microparticles generated within the flame. The computing unit performs image processing to extract image data representing at least the state of the flame or the particle flow from the image obtained by the imaging device. Further, the computing unit performs regression prediction of the refractive index profile of the transparent glass base material, which is the target variable, from the explanatory variables including the image data. Note that the image data representing the state of the particle flow includes contour information of the flame or the particle flow identified in the image capturing the flame or the particle flow, luminance information of light reaching from the flame or the particle flow, and the like.

[0013] As described above, the prediction of the refractive index profile of the transparent glass base material is performed at one or more arbitrary points in time during the period from the start to the end of the production of the glass microparticle deposit. Therefore, it becomes possible to predict the profile of the transparent glass base material over the entire length of the transparent glass base material. Further, by predicting the profile during the production of such a glass microparticle deposit, it becomes possible to change the production conditions, and it becomes possible to effectively suppress characteristic defects caused by structural defects of the optical fiber, which is the final product.

[0014] More specifically, the glass base material manufacturing apparatus and the like of the present disclosure enable prediction of the refractive index profile of the transparent glass base material, which is the main part of the finally obtained optical fiber base material (the part including the part to be the core of the optical fiber), obtained by dehydrating and sintering the glass microparticle deposit, during the production of the glass microparticle deposit. This means that the production conditions of the transparent glass base material can be adjusted before profile measurement, and as a result, it becomes possible to reduce the number of defective products of the transparent glass base material. Further, it becomes possible to reduce the skill level of process engineers and shorten the feedback time for adjusting production conditions.

[0015] Furthermore, according to the glass preform manufacturing apparatus and the like of the present disclosure, since the refractive index profile of the glass preform can be controlled to a desired profile shape over the entire length of the preform, the characteristics of the obtained transparent glass preform can be stabilized to desired characteristics. Also, by manufacturing an optical fiber, which is the final product, from an optical fiber preform including such a transparent glass preform, the optical characteristics of the optical fiber can be stabilized to desired characteristics.

[0016] (2) The glass preform manufacturing method of the present disclosure is a method for manufacturing a glass fine particle deposit by the VAD method, and is realized by the above-described glass preform manufacturing apparatus. Specifically, as one aspect of the glass preform manufacturing method, it includes a gas supply step, a deposition step, and a prediction step. In the gas supply step, a glass raw material gas and a fuel gas are individually supplied to a burner. In the deposition step, glass fine particles are generated from the glass raw material gas in a flame obtained by combustion of the fuel gas supplied to the burner, and the glass fine particles generated in the flame are sprayed onto the glass fine particle deposit. In the prediction step, a refractive index profile of a transparent glass preform obtained by dehydration and sintering of the glass fine particle deposit is predicted at one or more arbitrary points in time during the period from the start to the end of the deposition step. In particular, the prediction step includes an imaging step and a calculation step. In the imaging step, a flame sprayed from the burner onto the glass fine particle deposit or a particle flow composed of glass fine particles generated in the flame is imaged. In the calculation step, at least image data representing the state of the flame or the particle flow is extracted from the image obtained in the imaging step. Further, a refractive index profile of a transparent glass preform as a target variable is regression-predicted from explanatory variables including the extracted image data. Also by the glass preform manufacturing method, the same effects as those of the above-described glass preform manufacturing apparatus are achieved.

[0017] (3) The base material profile prediction method of the present disclosure is a method applicable to the above-described glass base material manufacturing apparatus and glass base material manufacturing method, and predicts the refractive index profile of a transparent glass base material obtained by dehydration and sintering of a glass fine particle deposit manufactured by the VAD method at one or more arbitrary points in time during the period from the start to the end of the manufacture of the glass fine particle deposit. Specifically, as one aspect, the base material profile prediction method includes an imaging step, an image processing step, and a calculation step. In the imaging step, a flame sprayed from a burner onto a glass fine particle deposit or a particle flow composed of glass fine particles generated within the flame is imaged. This imaging of the flame or particle flow (imaging step) is performed at an arbitrary point in time when glass fine particles are being generated from the glass raw material gas supplied to the burner within the flame obtained by combustion of the fuel gas supplied to the burner and the glass fine particles generated within the flame are being sprayed onto the glass fine particle deposit. In the image processing step, image data representing the state of the flame or particle flow is extracted from the image obtained in the imaging step. In the calculation step, the refractive index profile of the transparent glass base material as the target variable is regression-predicted from explanatory variables including at least the image data extracted in the image processing step. The base material profile prediction method also exhibits the same effects as the above-described glass base material manufacturing apparatus.

[0018] (4) As one aspect of the present disclosure, the explanatory variable preferably includes at least the contour data of the flame or the particle flow within the flame. As will be described later as an example, this is because a high correlation can be confirmed between the contour data of the particle flow and the shape of the refractive index profile of the transparent glass base material through basic analysis (the same applies to the contour data of the flame). Further, as one aspect of the present disclosure, the explanatory variable further preferably includes at least any one of the luminance distribution data of the flame or the particle flow, the data obtained by quantifying the installation position and installation angle of the burner, the flow rate data of the glass raw material gas input into the burner, the flow rate data of the fuel gas, the temperature in the heating furnace during dehydration and sintering (sintering temperature), and the gas flow rate supplied into the heating furnace during the dehydration and sintering. Since a high correlation can be confirmed between these data and the refractive index profile of the transparent glass base material, including these data as explanatory variables enables more accurate profile prediction.

[0019] (5) As one aspect of the present disclosure, the objective variable preferably includes the refractive index profile of the transparent glass base material, or one or more types of data characterizing the refractive index profile of the transparent glass base material. Note that this refractive index profile is the distribution of the specific refractive index difference along the radial direction (the direction orthogonal to the base material central axis) of the transparent glass base material. In this case, it becomes possible to visually display the prediction result.

[0020] (6) As one aspect of the present disclosure, in the arithmetic unit or arithmetic process, the refractive index profile or one or more types of data characterizing the refractive index profile are set as target variables, and for each target variable, a learning model is constructed in advance using regression analysis including at least one of decision tree regression, random forest (RF), gradient boosting, multiple regression, and Lasso regression, and the constructed learning model is used to predict the target variable (regression prediction). This learning model is a prediction model constructed using known data on the correlation between the explanatory variables and the target variables. Thus, as regression prediction, it is preferable that regression analysis including at least one of decision tree regression, random forest, gradient boosting, multiple regression, and Lasso regression is performed for each target variable. By applying regression analysis suitable for prediction for each target variable, it becomes possible to predict the profile with high accuracy.

[0021] (7) As one aspect of the present disclosure, the glass base material manufacturing apparatus having the above-described structure may further include a filter disposed between the imaging device and the space sandwiched between the glass fine particle deposit and the burner. This filter allows light of a predetermined wavelength from the flame or particle flow to pass through. For example, when imaging thermal radiation light from the flame or particle flow, the load of image processing is reduced by removing light of unnecessary wavelengths.

[0022] As described above, each aspect listed in this [Description of Embodiments of the Present Disclosure] column is applicable to each of the remaining all aspects or to all combinations of these remaining aspects.

[0023] [Details of Embodiments of the Present Disclosure] Hereinafter, the specific structures of the glass base material manufacturing apparatus, the glass base material manufacturing method, and the base material profile prediction method of the present disclosure will be described in detail with reference to the accompanying drawings. Note that the present invention is not limited to these examples, and is shown by the claims, and is intended to include all modifications within the meaning and scope equivalent to the claims. Also, in the description of the drawings, the same elements are denoted by the same reference numerals and redundant descriptions are omitted.

[0024] FIG. 2 is a diagram showing a configuration example of a glass preform manufacturing apparatus (an apparatus for realizing a glass preform manufacturing method according to an embodiment of the present disclosure). The basic configuration of the glass preform manufacturing apparatus 10 shown in FIG. 2 is substantially the same as the apparatus configuration shown in Patent Document 1 above, but is not limited to the apparatus configuration shown in Patent Document 1 above, and an apparatus configuration similar to a general glass preform manufacturing apparatus is applicable. The glass preform manufacturing apparatus 10 according to the present embodiment is different from the apparatus configuration shown in Patent Document 1 above and the apparatus configuration of a general glass preform manufacturing apparatus in that it includes a profile prediction system in addition to the basic configuration as described above. That is, the glass preform manufacturing apparatus 10 according to the present embodiment mainly includes a configuration for performing a deposition process (a reaction vessel, a gas supply system, a burner, a drive unit, etc.), a configuration for controlling the deposition process (a control unit, an image analysis unit, a drive unit, etc.), and a configuration for predicting the refractive index profile of a transparent glass preform obtained after dehydration and sintering from the deposition state of glass fine particles (a profile prediction system including an arithmetic unit, etc.).

[0025] The glass preform manufacturing apparatus 10 according to the present embodiment includes a reaction vessel 11 for manufacturing a glass fine particle deposit 14. An exhaust duct 29 is provided in the reaction vessel 11, and a part of a support rod 12 having a starting glass rod for depositing glass fine particles attached to one end thereof is located in the reaction vessel 11. As an example, a glass rod made of quartz glass with a diameter of 25 mm and a length of 400 mm is applied as the starting glass rod.

[0026] The other end of the support rod 12 is supported by the lifting and rotating device 15. This lifting and rotating device 15 rotates the support rod 12 along the arrow S1 in Fig. 2 and raises and lowers the support rod 12 along the direction indicated by the arrow S2. The operation of the lifting and rotating device 15 is controlled by a driving unit (lifting speed control unit) 20 that forms part of the control unit (control device) 16. The control unit 16 individually performs gas flow rate control for the gas supply system 19, position control of the core burner stage 24 and the angle adjustment mechanism 24a provided on the core burner stage 24, and position control of the cladding burner stage 25 and the angle adjustment mechanism 25a provided on the cladding burner stage 25 according to the control conditions input from the outside.

[0027] In the example of Fig. 2, the core burner 17 has five pipes (arranged concentrically) with different outer diameters. The glass raw material gas (SiCl4, GeCl4, and O2) supplied from the gas supply system 19 is introduced into the pipe with the smallest diameter (the innermost pipe). Burner seal gas (N2) is introduced into the space on its outer periphery, fuel gas (H2) is introduced into the space on the further outer periphery, burner seal gas (N2) is introduced into the further outer periphery, and combustion-supporting gas (O2) is introduced into the further outer periphery. In the flame of the core burner 17, glass fine particles (SiO2) and a refractive index adjusting dopant (GeO2) are generated by the hydrolysis reaction and combustion reaction of the glass raw material gas shown below, and the glass fine particles generated in the flame are sprayed from the core burner 17 onto the glass fine particle deposit 14. The burner seal gas is a gas introduced to prevent thermal degradation of the burner tip and deposition of glass fine particles on the burner tip, and functions to separate the glass raw material gas, fuel gas, and combustion-supporting gas near the pipe end of the core burner 17. SiCl4 + 2H2O -> SiO2 + 4HCl GeCl4 + O2 -> GeO2 + 2Cl2

[0028] Incidentally, the structure of the cladding burner 18 is also substantially the same as that of the core burner 17 described above, but the types of raw materials of the dopants for refractive index adjustment contained in the glass raw material gas supplied from the gas supply system 19 are different. For example, when fluorine (F) is added as a dopant for refractive index adjustment to the cladding portion, the glass raw material gas contains CF4 together with SiCl4. However, when the refractive index of the cladding portion is not adjusted, the raw material of the dopant for refractive index adjustment may not be contained in the glass raw material gas.

[0029] In the glass substrate manufacturing apparatus 10, the deposition surface shape of a part of the glass fine particle deposit 14 (particularly, the periphery of the portion that should become the core of the optical fiber) is monitored by a measurement camera (CCD camera) 21. The signal processing unit (image processing unit) 22 outputs the video data generated based on the electrical signal from the measurement camera 21 to an image analysis unit (deposition shape measurement unit) 23 that constitutes a part of the control unit 16. The image analysis unit 23 divides the video data (moving image) into two-dimensional images (still images), and extracts the deposition surface shape from the obtained two-dimensional images (still images). Then, the image analysis unit 23 performs drive control on the drive unit 20 (outputs a corrected drive control signal to the drive unit 20) so that the extracted deposition surface shape becomes the target shape. Further, the image analysis unit 23 calculates the correction amounts of the concentration and flow rate of the gas containing the glass raw material and the like, and the correction amount of the burner position so that the deposition surface shape becomes the target shape. The control unit 16 controls the gas supply system 19, the core burner stage 24, and the cladding burner stage 25 according to the correction amounts obtained by the image analysis unit 23.

[0030] The glass substrate manufacturing apparatus 10 according to the present embodiment further includes a profile prediction system that outputs the prediction result of the refractive index profile of the transparent glass substrate 140 (see FIG. 1) obtained by dehydration and sintering of the glass fine particle deposit 14 at one or more arbitrary times during the period from the start to the end of the production of the glass fine particle deposit 14. In one example, profile prediction (the base material profile prediction method of the present disclosure) is executed at 36,000 points (at 1-second intervals) spaced apart from each other along the longitudinal direction of the glass fine particle deposit 14.

[0031] As shown in FIG. 2, the profile prediction system includes a filter 102, a CCD camera (imaging device) 103, a signal processing unit 104, a calculation unit 105, and an output unit 106 such as a display. Note that at least the calculation unit 105 and the output unit 106 can be configured by a personal computer (hereinafter referred to as “PC”) 110. The filter 102 is disposed between the CCD camera 103 and the space sandwiched between the glass particle deposit 14 and the core burner 17, and allows light of a predetermined wavelength (for example, a part of the thermal radiation light from the particle flow) from the particle flow in the burner flame to pass through. In particular, the load of image processing is reduced by removing light of unnecessary wavelengths. As an example, the sampling interval by the CCD camera 103 is about 0.1 second to 1 second. The shutter speed is between 0.1 ms and 1000 ms.

[0032] Furthermore, in the profile prediction system, the arithmetic unit 105 performs image processing to extract image data representing at least the state of a flame or a particle flow from the two-dimensional image obtained by the imaging device. As an example, the image processing in the arithmetic unit 105 is performed using image analysis software. Specifically, after the two-dimensional image from the signal processing unit 104 is adjusted in luminance, the contour of the flame or the particle flow is clarified. Further, the arithmetic unit 105 performs regression prediction of the refractive index profile of the transparent glass base material 140 as the target variable from the explanatory variables including at least the data in which the contour of the flame or the particle flow is coordinated. More specifically, the contour data of the flame or the particle flow, the burner installation position (the installation position of the burner along the burner X-axis, the burner installation angle), the flow rate of the glass raw material gas (including the raw material of the dopant for refractive index adjustment), the flow rate of the fuel gas (H2), the flow rate of the combustion-supporting gas (O2), the dehydration and sintering conditions (temperature, gas flow rate) when the glass fine particle deposit 14 is vitrified into transparent glass, etc. are set as explanatory variables, and the data characterizing the refractive index profile is set as the target variable. The dehydration and sintering process is a process of vitrifying the glass fine particle deposit 14 into transparent glass in a heating furnace. While supplying at least one type of gas selected from, for example, nitrogen, argon, helium, chlorine, etc. into the core tube in which the glass fine particle deposit 14 is housed, the glass fine particle deposit 14 is heated by a heater disposed outside the core tube, thereby performing dehydration and sintering (vitrification) of the glass fine particle deposit 14. The "gas flow rate" as the dehydration and sintering condition means the flow rate of the above gas supplied into the core tube in this dehydration and sintering process. After determining the explanatory variables and the target variable, the arithmetic unit 105 models in advance the correlation between the contour of the flame or the particle flow and the refractive index profile using decision tree regression, random forest (RF), gradient boosting, multiple regression, lasso regression, etc. During the production of the glass fine particle deposit 14, the arithmetic unit 105 uses the learning model (prediction model) constructed in this way to predict the refractive index profile of the transparent glass base material 140. Note that the image data representing the state of the flame or the particle flow includes the contour data of the flame or the particle flow specified by the two-dimensional image of the particle flow, or the luminance information of the thermal radiation light from the particles.

[0033] FIG. 3 is a diagram for explaining the position of the burner with respect to the glass particle deposit 14 during the deposition of glass particles. As shown in FIG. 3, the glass particle deposit 14 and the core burner 17 are arranged such that the base material central axis AX1 and the burner central axis (substantially the central axis of the pipe) AX2 intersect. Such a relative positional relationship is the same for the positional relationship between the glass particle deposit 14 and the cladding burner 18. The core burner stage 24 moves the burner central axis AX2 along the burner X-axis. Further, the angle adjustment mechanism 24a adjusts the angle of the core burner with respect to the core burner stage 24 (burner angle θ).

[0034] FIG. 4 is a diagram for explaining, as an example, the contour data of the particle flow (which may be luminance distribution data) and the basic analysis that serves as the basis for determining the burner position as explanatory variables. Note that FIG. 4 shows the contour change (from pattern 1 to pattern 3) of the particle flow (the group of glass particles in the flame sprayed from the core burner 17 onto the glass particle deposit 14) adjusted by the positional relationship between the glass particle deposit 14 and the core burner 17, and the refractive index profile of the transparent glass base material 140 obtained by dehydrating and sintering the glass particle deposit 14 (the refractive index profile along the diameter direction orthogonal to the central axis AX0 of the transparent glass base material 140).

[0035] In FIG. 4, in the right column of pattern 1, a scissor-type refractive index profile is shown as the schematic shape of the refractive index profile of the transparent glass base material 140 obtained after dehydration and sintering. In pattern 1, as shown in the left column, the core burner 17 is arranged such that the burner central axis AX2 of the core burner 17 is shifted upward (a position closer to the starting glass rod 13 than X0) with respect to the origin X0 of the burner X-axis. By arranging the core burner 17 in such a position with respect to the glass particle deposit 14, a scissor-type refractive index profile having a refractive index peak at a position away from the central axis AX0 of the transparent glass base material 140 after dehydration and sintering is likely to be obtained.

[0036] In the right column of Pattern 2, a mountain-shaped refractive index profile is shown as a schematic shape of the refractive index profile of the transparent glass base material 140 obtained after dehydration and sintering. In Pattern 2, as shown in the left column, the core burner 17 is arranged such that the burner central axis AX2 of the core burner 17 is shifted downward (a position farther from the starting glass rod 13 than the origin X0 of the burner X-axis) with respect to the origin X0 of the burner X-axis. By arranging the core burner 17 at such a position with respect to the glass particle deposit 14, a mountain-shaped refractive index profile having a refractive index peak at the core center that coincides with the central axis AX0 of the transparent glass base material 140 after dehydration and sintering is easily obtained.

[0037] In the right column of Pattern 3, a trapezoidal refractive index profile is shown as a schematic shape of the refractive index profile of the transparent glass base material 140 obtained after dehydration and sintering. In Pattern 3, as shown in the left column, the core burner 17 is arranged such that the burner central axis AX2 of the core burner 17 intersects the origin X0 of the burner X-axis. By arranging the core burner 17 at such a position with respect to the glass particle deposit 14, a trapezoidal refractive index profile with small refractive index fluctuations in the peripheral region centered on the central axis AX0 of the transparent glass base material 140 after dehydration and sintering is easily obtained.

[0038] FIG. 5 is a diagram showing an example of a schematic configuration of the PC 110 including the arithmetic unit 105. In the arithmetic unit 105, an image signal from the signal processing unit 104 is captured, and image processing for extracting explanatory variables (such as contour data of a flame or a particle flow, luminance distribution data of a flame or a particle flow, etc.) is performed. The explanatory variables are determined through the basic analysis as shown in FIG. 4. Examples of the explanatory variable candidates include contour data of a flame or a particle flow, luminance distribution data of a flame or a particle flow, the installation position of the core burner 17 (burner X-axis position, burner angle θ, etc.), the flow rate of the glass raw material gas, the flow rate of the fuel gas for flame generation, etc. As other information, the shape of the deposition surface, the conditions of dehydration and sintering (including temperature, gas flow rate, etc.) can also be candidates for explanatory variables. The above contour data is obtained by performing image processing in advance, while other information is input from the outside as manufacturing condition data. In the present embodiment, based on the results of the basic analysis shown in FIGS. 4 and 9, the contour data of the particle flow existing in the flame, the luminance distribution data of the particle flow, and the burner installation position (burner X-axis position and burner angle) are determined as explanatory variables.

[0039] The arithmetic unit 105 utilizes the explanatory variables and objective variables obtained in past manufacturing to construct a correlation model between the explanatory variables and the objective variables using a learning model, has a memory storing the constructed learning model, and performs regression prediction using the learning model. For each objective variable, a randomly selected regression analysis (in the example of FIG. 5, regression analysis 1 to regression analysis 3) is performed. As the regression analysis, for example, decision tree regression, random forest, gradient boosting, multiple regression, Lasso regression, etc. are applicable.

[0040] Generally, the random forest is an analysis method that repeatedly performs a process of randomly selecting learning data to construct decision trees, and performs classification and regression based on the majority vote or average value of the estimation results of each decision tree. In particular, since the random forest uses a plurality of learning models (decision trees), it is called ensemble learning.

[0041] Gradient boosting is an analytical method that first performs decision tree analysis, then repeats the process of constructing a decision tree for the error between the predicted value of the constructed decision tree model and the true value multiple times. Like random forest, it is an ensemble learning method, but while random forest creates decision trees in parallel, gradient boosting constructs decision trees in series.

[0042] Multiple regression analysis is an analytical method for predicting one objective variable using multiple explanatory variables (numeric values). For example, when one objective variable is y and n (an integer equal to or greater than 1) explanatory variables are xi (where i is an integer from 1 to n), the multiple regression analysis is given by the following formula (1). y=a1×x1+a2×x2+…+an×xn+b…(1) Here, ai (i is an integer from 1 to n) is the regression coefficient, and b is the intercept. A learning model is constructed by determining the regression coefficients ai and intercept b using multiple learning data with known objective variables y and explanatory variables xi.

[0043] Lasso regression analysis is an analytical model that adds "L1 regularization" to linear regression, such as the multiple regression analysis mentioned above. In Lasso regression analysis, the regression coefficients for data that are unlikely to affect the prediction are brought closer to zero, so it is essentially a regression analysis that selects only important explanatory variables.

[0044] In this embodiment, the response variable obtained by regression prediction is the refractive index profile of the transparent glass base material 140 or data characterizing the refractive index profile. The output unit 106 includes a monitor or the like that reproduces the refractive index profile that is regression predicted.

[0045] Fig. 6 is a diagram for explaining the image processing (contour data extraction) in the above-mentioned calculation unit 105. Fig. 7 is a table showing numerical data obtained by the image processing in Fig. 6.

[0046] The calculation unit 105 captures the video data (moving image) output from the signal processing unit 104, divides this data into n (an integer of 1 or more) two-dimensional still images Gi (i is an integer from 1 to n), and for each two-dimensional still image, extracts the contour data (image data) of the particle flow in the flame sprayed from the core burner 17 to the glass microparticle deposit 14. Specifically, as shown in the upper part of FIG. 6, the two-dimensional still image Gi is an image obtained by imaging the imaging area RA shown in FIG. 2 with the CCD camera 103, and the necessary area is cut out. The contour data of the particle flow is extracted after adjusting the luminance of the cut-out area. The extracted contour data is composed of an upper contour Fu and a lower contour Fd as shown in the upper part of FIG. 6. Finally, each contour Fu and Fd is subjected to a smoothing process. Note that the upper part of FIG. 6 shows the smoothing process of the upper contour Fu located within the area GS. When the above-mentioned contour data extraction is completed for the n two-dimensional still images, n contour coordinates are obtained. The lower part of FIG. 6 graphs the n contour coordinates (the horizontal axis is denoted as the "contour X-axis" and the vertical axis is denoted as the "contour Y-axis"). The table in FIG. 7 summarizes the results of averaging the n contour coordinates for each base material sample.

[0047] In the graph shown in the lower part of FIG. 6, the area GA including the n upper contours Fu is an area defined by the coordinates Y011 to Y072 (hereinafter referred to as "explanatory variable candidates") on the contour Y-axis. For each of the explanatory variable candidates Y011 to Y072, the average contour coordinate value of the n upper contours Fu on the contour X-axis is calculated. On the other hand, the area GB including the n lower contours Fd is an area defined by the coordinates X006 to X066 (hereinafter referred to as "explanatory variable candidates") on the contour X-axis. For each of the explanatory variable candidates X006 to X066, the average contour coordinate value of the n lower contours Fd on the contour Y-axis is also calculated. In the example of FIG. 7, as explanatory variables for each base material sample, the average contour coordinate values of each of the 7 candidates (Y011, Y021,..., Y071) extracted from the explanatory variable candidates Y011 to Y072 that define the area GA, and the average contour coordinate values of each of the 7 candidates (X006, X016,..., X066) extracted from the explanatory variable candidates X006 to X066 that define the area GB are set.

[0048] Furthermore, in the image processing in the above-described arithmetic unit 105, as shown in FIG. 8, luminance distribution data is extracted for each two-dimensional still image as an explanatory variable. Note that FIG. 8 is a diagram for explaining the image processing (luminance distribution data extraction) in the arithmetic unit 105.

[0049] Similar to the above-described contour data extraction operation, the arithmetic unit 105 captures the video data (moving image) output from the signal processing unit 104, divides this data into n (an integer of 1 or more) two-dimensional still images Gi (i is an integer from 1 to n), and for each two-dimensional still image, extracts the luminance distribution of the particle flow generated in the flame sprayed from the core burner 17 to the glass microparticle deposit 14. Specifically, as shown in the upper part of FIG. 8, the two-dimensional still image Gi is an image obtained by imaging the imaging area RA shown in FIG. 2 with the CCD camera 103, and a necessary area is cut out. For example, the luminance distribution of the particle flow in the two-dimensional still image Gi is defined as the luminance BT(CPx) at each point (70 luminance measurement points in the example of FIG. 8) on the line segment FL connecting the start point CP1 and the end point CP70 on the two-dimensional still image Gi. Here, x is an arbitrary integer from 1 to 70. By averaging the luminance BT(CPx) obtained from each of the n two-dimensional still images at the same luminance measurement point, the average luminance ABT(CPx) is obtained.

[0050] The obtained average luminance ABT(CPx) indicates the average luminance distribution at a total of 70 locations from CP1 to CP70 on the line segment FL. Furthermore, the luminance distribution data prepared as an explanatory variable is composed of, for example, 40 average luminance values ABT(CPx) arbitrarily selected from among CP1 to CP70. Note that the number of luminance measurement points selected from the luminance measurement points on the line segment FL is not limited to 40. For example, by evaluating each of the luminance distributions composed of 5, 10, 20, and 40 locations, the optimal number of components (number of luminance measurement points) of the luminance distribution data may be determined.

[0051] Next, FIG. 9 is a diagram for explaining the basic analysis that serves as the basis for determining the burner position as an explanatory variable. As shown in FIG. 5, in the present embodiment, not only the contour data of the particle flow but also the burner installation position (burner X-axis position and burner angle θ) of the core burner 17 are used as explanatory variables. In the upper part of FIG. 9, for each base material, a trend graph of the contour Y-axis value at which the contour X-axis value of the upper contour Fu of the particle flow is minimized is shown. Group A is a base material group in which the position of the core burner 17 is moved along the burner X-axis with respect to Group 0, and Group B is a base material group in which the angle of the core burner 17 is further adjusted. In the lower part of FIG. 9, it can be seen that for Group 0 in which the core burner 17 is in the standard position, in Group A, the flame contour moves upward and the contour Y-axis value at which the contour X-axis value of the upper contour Fu is minimized changes. Therefore, in the present embodiment, the burner installation position (burner X-axis position and burner angle θ) is also added as an explanatory variable.

[0052] FIG. 10 is a diagram for explaining the objective variable. The present embodiment predicts the refractive index profile of the transparent glass base material 140 obtained by sintering the glass fine particle deposit 14 on which glass fine particles are deposited by the VAD method while dehydrating. As an example, four types of data characterizing the refractive index profile to be predicted are the objective variables.

[0053] The four types of data characterizing the refractive index profile in FIG. 10 are explained. The objective variable 1 (feature "A") represents the difference between the maximum specific refractive index difference and the specific refractive index difference of the central axis A0. The objective variable 2 (feature "B") represents the distance from the central axis A0 to the radial position at which the maximum specific refractive index difference is obtained. The objective variable 3 (feature "C / D") indicates the slope of the profile. The feature "C" is the specific refractive index difference that is half of the maximum specific refractive index difference, the feature "D" is the distance from the radial position at which the maximum specific refractive index difference is obtained to the radial position at which the specific refractive index difference is half of the maximum specific refractive index difference, and the ratio "C / D" of them is the objective variable 3. The objective variable 4 (feature "E") represents the maximum specific refractive index difference.

[0054] Figure 11 shows a part of the data (explanatory variables and objective variables) used to construct the learning model in this embodiment. In this embodiment, data was acquired from 71 base material samples having a "scissors type" refractive index profile, 14 base material samples having a "mountain type" refractive index profile, and 11 base material samples having a "trapezoidal type" refractive index profile. Note that in this embodiment, a learning model is constructed using all data (explanatory variables and objective variables) excluding the test target, and the error between the measured value of the test target and the predicted value of the learning model is evaluated by RMSE (Root Mean Square Error). RMSE: Root Mean Square Error ((1 / n)×Σ(true value - predicted value) 2 ) 1 / 2 n: Number of data

[0055] Figure 11 shows the explanatory variables (measured values) and objective variables (measured values) of each of the total 96 base material samples classified as described above. Note that in the upper part of Figure 11, data of four types of objective variables (A, B, C / D, E) are shown for each base material sample. In the lower part of Figure 11, as explanatory variables, the average contour coordinates at each of the seven locations in region GB shown in Figure 7, the average contour coordinates at each of the seven locations in region GA shown in Figure 8, the luminance data at 40 locations shown in Figure 8, the burner X-axis position, and the burner angle, which are composed of a total of 56 types of data, and these 56 types of data constituting the explanatory variables are shown for each base material sample.

[0056] Figure 12 shows the analysis method with the highest prediction accuracy for each objective variable in the prediction of the base material profile of the present disclosure, and its accuracy (RMSE). Further, Figure 13 is a graph showing the correlation between the predicted value and the measured value of the objective variable.

[0057] As shown in FIG. 12, the arithmetic unit 105 predicts the target variable 1 (feature “A”) using the random forest that provides the highest accuracy for the regression prediction of the target variable 1. The contour data of the particle flow in the flame, the luminance distribution data of the particle flow, and the burner position (burner X-axis position and burner angle) were determined as explanatory variables through the basic analysis shown in FIGS. 4 and 9. Further, the arithmetic unit 105 predicts the target variable 2 (feature “B”) and the target variable 3 (feature “C / D”) using gradient boosting that provides the highest accuracy for the regression prediction of the target variable 2 and the target variable 3. Furthermore, the arithmetic unit 105 predicts the target variable 4 (feature “E”) using Lasso regression that provides the highest accuracy for the regression prediction of the target variable 4.

[0058] As can be seen from FIG. 13, for any of the target variables 1 to 4, a significant correlation can be confirmed between the predicted value and the measured value of the target variable.

[0059] FIG. 14 shows the base material profile predicted by the base material profile prediction method mentioned in paragraphs “0053” to “0059”. The solid line is the measured refractive index profile, and the broken line is the predicted profile. It can be seen that in both Pattern 2 “mountain type” and Pattern 1 “scissors type”, regression prediction is performed with high accuracy.

[0060] FIG. 15 shows the result of regression prediction of all points (500 points in the radial direction) of the refractive index profile using only the random forest. The solid line is the measured profile, and the broken line is the predicted profile. It can be seen that the base material profile can be predicted with high accuracy even with only one type of learning model.

Explanation of reference numerals

[0061] 10…Glass base material manufacturing apparatus, 11…Reaction vessel, 12…Support rod, 13…Starting glass rod, 14…Glass particle deposit, 15…Lifting and rotating device, 16…Control unit, 17…Core burner, 18…Clad burner, 19…Gas supply system, 20…Drive unit, 21…Measurement camera, 22…Signal processing unit, 23…Image analysis unit, 24a, 25a…Angle adjustment mechanism, 24, 25…Stage, 29…Exhaust duct, RA…Imaging area, 102…Filter, 103…CCD camera (imaging device), 104…Signal processing unit, 105…Calculation unit, 106…Output unit, 110…Personal computer (PC).

Claims

1. A glass base material manufacturing apparatus for manufacturing a glass fine particle deposit by the VAD method, A gas supply system for separately supplying a glass raw material gas and a fuel gas, A burner that generates glass fine particles from the glass raw material gas in a flame obtained by combustion of the fuel gas supplied from the gas supply system, and blows the glass fine particles in the flame onto the glass fine particle deposit, A profile prediction system that outputs a prediction result of the refractive index profile of a transparent glass base material obtained by dehydration and sintering of the glass fine particle deposit at one or more arbitrary times during the period from the start to the end of the production of the glass fine particle deposit, comprising The profile prediction system An imaging device for imaging the flame generated by the burner or a particle flow composed of the glass fine particles generated in the flame, Extracts at least image data representing the state of the flame or the particle flow from the image obtained by the imaging device, and regression-predicts the refractive index profile of the transparent glass base material as the target variable from the explanatory variables including the image data, having A glass base material manufacturing apparatus.

2. The explanatory variables include at least the contour data of the flame or the contour data of the particle flow in the flame, The glass base material manufacturing apparatus according to claim 1.

3. The explanatory variables further include at least any one of the luminance distribution data of the flame or the particle flow, the data obtained by quantifying the installation position and installation angle of the burner, the flow rate data of the glass raw material gas input to the burner, the flow rate data of the fuel gas, the temperature in the heating furnace during dehydration and sintering, and the gas flow rate supplied into the heating furnace during dehydration and sintering, The glass base material manufacturing apparatus according to claim 2.

4. The target variable includes data characterizing the refractive index profile of the transparent glass base material, The glass base material manufacturing apparatus according to any one of claims 1 to 3.

5. For each of the target variables, the calculation unit uses regression analysis including one or more of decision tree regression, random forest, gradient boosting, multiple regression, and Lasso regression to perform regression prediction on the refractive index profile of the transparent glass base material. The glass base material manufacturing apparatus according to any one of claims 1 to 4.

6. A filter disposed between the imaging device and the space sandwiched between the glass particle deposit and the burner, the filter passing light of a predetermined wavelength from the particle flow, further comprising. The glass base material manufacturing apparatus according to any one of claims 1 to 5.

7. A method for manufacturing a glass base material by producing a glass particle deposit by the VAD method and dehydrating and sintering the glass particle deposit in a heating furnace, A gas supply step of separately supplying a glass raw material gas and a fuel gas to a burner, A deposition step of generating glass particles from the glass raw material gas in a flame obtained by combustion of the fuel gas supplied to the burner and spraying the glass particles in the flame onto the glass particle deposit, A prediction step of predicting the refractive index profile of the transparent glass base material obtained by dehydration and sintering of the glass particle deposit at one or more arbitrary times during the period from the start to the end of the deposition step, Comprising, The prediction step is, An imaging step of imaging the flame generated by the burner or a particle flow composed of glass particles generated in the flame, Extract at least image data representing the state of the flame or the particle flow from the image obtained in the imaging step, and perform a regression prediction of the refractive index profile of the transparent glass base material, which is the target variable, from the explanatory variables including the image data. including A method for manufacturing a glass base material.

8. The explanatory variables include at least the contour data of the flame or the contour data of the particle flow in the flame. The method for manufacturing a glass base material according to claim 7.

9. The explanatory variables further include at least any one of the luminance distribution data of the flame or the particle flow, the data obtained by quantifying the installation position and installation angle of the burner, the flow rate data of the glass raw material gas input to the burner, the flow rate data of the fuel gas, the temperature in the heating furnace during dehydration and sintering, and the gas flow rate supplied into the heating furnace during dehydration and sintering. The method for manufacturing a glass base material according to claim 8.

10. The target variable includes data characterizing the refractive index profile of the transparent glass base material. The method for manufacturing a glass base material according to any one of claims 7 to 9.

11. The calculation step uses regression analysis including one or more of decision tree regression, random forest, gradient boosting, multiple regression, and Lasso regression for each target variable to perform regression prediction of the refractive index profile of the transparent glass base material. The method for manufacturing a glass base material according to any one of claims 7 to 10.

12. A base material profile prediction method for predicting the refractive index profile of a transparent glass base material obtained by dehydration and sintering of a glass microparticle deposit manufactured by the VAD method at one or more arbitrary time points during the period from the start to the end of the manufacture of the glass microparticle deposit. While generating glass fine particles from the glass raw material gas supplied to the burner within the flame obtained by combustion of the fuel gas supplied to the burner, and spraying the glass fine particles in the flame onto the glass fine particle deposit at any given time, an imaging step of imaging the flame generated by the burner or a particle flow composed of the glass fine particles generated within the flame; An image processing step of extracting image data representing the state of the flame or the particle flow from the image obtained in the imaging step; An arithmetic step of regression-predicting the refractive index profile of the transparent glass base material as a target variable from explanatory variables including at least the image data extracted in the image processing step; Including A base material profile prediction method.

13. The explanatory variables include at least the contour data of the flame or the contour data of the particle flow within the flame. The base material profile prediction method according to claim 12.

14. The explanatory variables further include at least any one of the luminance distribution data of the flame or the particle flow, the data obtained by quantifying the installation position and installation angle of the burner, the flow rate data of the glass raw material gas input to the burner, the flow rate data of the fuel gas, the temperature inside the heating furnace during dehydration and sintering, and the gas flow rate supplied into the heating furnace during dehydration and sintering. The base material profile prediction method according to claim 13.

15. The target variable includes data characterizing the refractive index profile of the transparent glass base material. The base material profile prediction method according to any one of claims 12 to 14.

16. The arithmetic step uses regression analysis including one or more of decision tree regression, random forest, gradient boosting, multiple regression, and Lasso regression for each target variable to regression-predict the refractive index profile of the transparent glass base material. The base material profile prediction method according to any one of claims 12 to 15.

Images