AU2020101671A4 - Method and System for Predicting Quality of Polycrystalline Silicon Ingot Based on Ingredient Data - Google Patents
Method and System for Predicting Quality of Polycrystalline Silicon Ingot Based on Ingredient Data Download PDFInfo
- Publication number
- AU2020101671A4 AU2020101671A4 AU2020101671A AU2020101671A AU2020101671A4 AU 2020101671 A4 AU2020101671 A4 AU 2020101671A4 AU 2020101671 A AU2020101671 A AU 2020101671A AU 2020101671 A AU2020101671 A AU 2020101671A AU 2020101671 A4 AU2020101671 A4 AU 2020101671A4
- Authority
- AU
- Australia
- Prior art keywords
- data
- quality
- ingredient
- polycrystalline silicon
- support vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- C—CHEMISTRY; METALLURGY
- C30—CRYSTAL GROWTH
- C30B—SINGLE-CRYSTAL GROWTH; UNIDIRECTIONAL SOLIDIFICATION OF EUTECTIC MATERIAL OR UNIDIRECTIONAL DEMIXING OF EUTECTOID MATERIAL; REFINING BY ZONE-MELTING OF MATERIAL; PRODUCTION OF A HOMOGENEOUS POLYCRYSTALLINE MATERIAL WITH DEFINED STRUCTURE; SINGLE CRYSTALS OR HOMOGENEOUS POLYCRYSTALLINE MATERIAL WITH DEFINED STRUCTURE; AFTER-TREATMENT OF SINGLE CRYSTALS OR A HOMOGENEOUS POLYCRYSTALLINE MATERIAL WITH DEFINED STRUCTURE; APPARATUS THEREFOR
- C30B28/00—Production of homogeneous polycrystalline material with defined structure
- C30B28/04—Production of homogeneous polycrystalline material with defined structure from liquids
- C30B28/06—Production of homogeneous polycrystalline material with defined structure from liquids by normal freezing or freezing under temperature gradient
-
- C—CHEMISTRY; METALLURGY
- C30—CRYSTAL GROWTH
- C30B—SINGLE-CRYSTAL GROWTH; UNIDIRECTIONAL SOLIDIFICATION OF EUTECTIC MATERIAL OR UNIDIRECTIONAL DEMIXING OF EUTECTOID MATERIAL; REFINING BY ZONE-MELTING OF MATERIAL; PRODUCTION OF A HOMOGENEOUS POLYCRYSTALLINE MATERIAL WITH DEFINED STRUCTURE; SINGLE CRYSTALS OR HOMOGENEOUS POLYCRYSTALLINE MATERIAL WITH DEFINED STRUCTURE; AFTER-TREATMENT OF SINGLE CRYSTALS OR A HOMOGENEOUS POLYCRYSTALLINE MATERIAL WITH DEFINED STRUCTURE; APPARATUS THEREFOR
- C30B29/00—Single crystals or homogeneous polycrystalline material with defined structure characterised by the material or by their shape
- C30B29/02—Elements
- C30B29/06—Silicon
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Chemical & Material Sciences (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Evolutionary Computation (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Materials Engineering (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- General Business, Economics & Management (AREA)
- Organic Chemistry (AREA)
- Metallurgy (AREA)
- Databases & Information Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Crystallography & Structural Chemistry (AREA)
- Educational Administration (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Silicon Compounds (AREA)
- Crystals, And After-Treatments Of Crystals (AREA)
Abstract
The present invention relates to a method and system for predicting quality of a polycrystalline
silicon ingot based on ingredient data. The method includes acquiring ingredient data of a
polycrystalline silicon ingot; preprocessing the ingredient data, wherein the preprocessing includes
cleaning data and deleting data with missing values; reducing dimensions of the preprocessed
ingredient data by using a diffusion mapping algorithm; building a support vector data description
model with the dimension-reduced ingredient data in a set proportion; and performing quality
prediction classification of the dimension-reduced ingredient data by using the trained support vector
data description model, to obtain product quality prediction classification results. The method and
system for predicting quality of a polycrystalline silicon ingot based on ingredient data according to
the present invention can predict the quality of polycrystalline silicon ingots quickly, accurately and
at low cost.
13
Description
The present invention relates to the field of quality prediction of a polycrystalline silicon ingot, and in particular, to a method and system for predicting quality of a polycrystalline silicon ingot based on ingredient data.
Polycrystalline silicon is a form of monatomic silicon. When molten monatomic silicon solidifies under supercooling conditions, silicon atoms are arranged into many crystal nuclei in the form of a diamond lattice. If these crystal nuclei grow into crystal grains with different crystal orientations, these crystal grains will combine and crystallize into polysilicon. A complete cycle of polycrystalline silicon ingot casting mainly includes: spraying and feeding, inspection before ingot casting, ingot casting operation process, squaring, silicon wafer processing, silicon wafer testing, etc.
In the process of polycrystalline silicon ingot casting, its state can be described according to the actual measured physical quantities or mathematical statistics of system variables or parameters, while a process state, namely a system operation state, can reflect a quality state of products. Therefore, quality of a polycrystalline silicon ingot prediction aims to improve the equipment productivity by effectively monitoring and controlling key points of ingot production without increasing investment. The product quality is predicted before the completion of all polysilicon production processes, so as to perform intervention and correction in advance when product quality prediction results are abnormal.
At present, the prediction of quality of a polycrystalline silicon ingot is mainly based on semi melting or full-melting process test or numerical simulation by a computer.
A working process of the former includes: feeding, vacuumizing and heating, aerating and melting, crystal growth, annealing, cooling, discharging, etc. This process is the same as the production process commonly used in industry at present, and is relatively mature.
In the process of numerical simulation, the latter mainly uses professional crystal production simulation software CGSim. The software can analyze and calculate a series of parameters such as temperature field, flow field and thermal stress inside the crystal in a furnace during the crystal growth process. A basic operation process includes: creating a geometric model, defining physical parameters of materials, setting heater power and boundary conditions of the model, adjusting crystal characteristics, setting a solver and determining convergence.
The process test is a complete process to simulate the actual production process. This needs to simulate all ingredients, and the experimental cost is very high. For ingredients from many different manufacturers and batches, many different tests need to be performed, and a single test result is not universal. Besides, the process test has high requirements for the accuracy of the whole test process. In addition, due to a series of inevitable factors such as high temperature and sealing in a polysilicon growth furnace, the in-situ test measurement is not only high in cost, but also difficult to realize concretely. Therefore, this method is not practical. In the process of numerical simulation by a computer, it is necessary to know an experimental device and process in advance, design the geometric model for the experiment, define the heater power, and set the boundary conditions of the model. The process is high in complexity and requires a long time.
In the actual industrial production of polycrystalline silicon ingots, the process design and environment are certain and unchangeable. Therefore, the key influencing factor on product quality lies in ingredient data. Different ingredient data directly leads to different quality of products. In the actual production process, the data size of abnormal product data is relatively small. This leads to the imbalance of data categories, and makes existing two-category classification methods prone to category shift and difficult to use effectively.
At present, conventional binary classifiers, such as a support vector machine (SVM), have been widely used in image quality monitoring and other fields. However, when the data is unbalanced, the binary classifiers can only distinguish normal data serving as amajority class, while for abnonnal data as a minority class, the recognition rate of a SVM classifier is very low, which does not meet the needs of actual industrial production.
Therefore, the existing methods cannot predict the quality of polycrystalline silicon ingots quickly, accurately and at low cost.
The present invention provides a method and system for predicting quality of a polycrystalline silicon ingot based on ingredient data, which can predict the quality of polycrystalline silicon ingots quickly, accurately and at low cost.
To achieve the above purpose, the present invention provides the following technical solutions.
A method for predicting quality of a polycrystalline silicon ingot based on ingredient data includes: acquiring ingredient data of a polycrystalline silicon ingot, where the ingredient data includes quality of primary polycrystalline bulk, quality of broken polycrystal, quality of fragments, quality of self-produced purified ingot cores, quality of purchased purified ingot cores, quality of top skin, quality of leftover materials, quality of tailings and minority carrier lifetime value; preprocessing the ingredient data, where the preprocessing includes cleaning data and deleting data with missing values; reducing dimensions of the preprocessed ingredient data by using a diffusion mapping algorithm; building a support vector data description model with the dimension-reduced ingredient data in a set proportion; and performing product quality prediction classification of the dimension-reduced ingredient data by using the trained support vector data description model, to obtain product quality prediction classification results, where the product quality prediction classification results include polycrystalline silicon ingots with acceptable quality and polycrystalline silicon ingots with unacceptable quality; and the trained support vector data description model takes the dimension reduced ingredient data as inputs, and takes the product quality prediction classification results as outputs.
Optionally, before the reducing dimensions of the preprocessed ingredient data by using a diffusion mapping algorithm, the method further includes:
dividing the preprocessed data to obtain numerical data and character data.
Optionally, after the building a support vector data description model with the dimension reduced ingredient data in a set proportion, the method further includes:
optimizing the support vector data description model by using a Monte Carlo algorithm to obtain the trained support vector data description model.
A system for predicting quality of a polycrystalline silicon ingot based on ingredient data includes:
an ingredient data acquisition module, configured to acquire ingredient data of a polycrystalline silicon ingot; where the ingredient data includes quality of primary polycrystalline bulk, quality of broken polycrystal, quality of fragments, quality of self-produced purified ingot cores, quality of purchased purified ingot cores, quality of top skin, quality of leftover materials, quality of tailings and minority carrier lifetime value;
a preprocessing module, configured to preprocess the ingredient data; where the preprocessing includes cleaning data and deleting data with missing values; a dimension reduction processing module, configured to reduce dimensions of the preprocessed ingredient data by using a diffusion mapping algorithm; a model building module, configured to build a support vector data description model with the dimension-reduced ingredient data in a set proportion; and a product quality prediction classification module, configured to perform quality prediction classification of the dimension-reduced ingredient data by using the trained support vector data description model, to obtain product quality prediction classification results; where the product quality prediction classification results include polycrystalline silicon ingots with acceptable quality and polycrystalline silicon ingots with unacceptable quality; and the trained support vector data description model takes the dimension-reduced ingredient data as inputs, and takes the product quality prediction classification results as outputs.
Optionally, the system further includes:
a data dividing module, configured to divide the preprocessed data to obtain numerical data and character data.
Optionally, the system further includes:
a module for determining a trained support vector data description model, configured to optimize the support vector data description model by using a Monte Carlo algorithm to obtain the trained support vector data description model.
According to specific examples of the present invention, the present invention has the following technical effects.
In the method and system for predicting quality of a polycrystalline silicon ingot based on ingredient data according to the present invention, dimensions of ingredient data features are reduced through a diffusion mapping algorithm, so as to achieve attribute reduction and remove redundancy features. Then, through a trained support vector data description model, the quality of polycrystalline silicon ingots is predicted without prior knowledge. That is, polycrystalline silicon ingots with both acceptable and unacceptable quality can be determined quickly, accurately and at low cost.
To describe the technical solutions in the examples of the present invention or in the prior art more clearly, the following briefly describes the accompanying drawings required for the examples. Apparently, the accompanying drawings in the following description show merely some examples of the present invention, and a person of ordinary skill in the art may still derive other accompanying drawings from these accompanying drawings without creative efforts.
FIG. 1 is a schematic flowchart of a method for predicting quality of a polycrystalline silicon ingot based on ingredient data according to the present invention; and
FIG. 2 is a schematic structural diagram of a system for predicting quality of a polycrystalline silicon ingot based on ingredient data according to the present invention.
The following clearly and completely describes the technical solutions in the examples of the present invention with reference to accompanying drawings in the examples of the present invention. Apparently, the described examples are merely some rather than all of the examples of the present invention. All other examples obtained by a person of ordinary skill in the art based on the examples of the present invention without creative efforts shall fall within the protection scope of the present invention.
The present invention provides a method and system for predicting quality of a polycrystalline silicon ingot based on ingredient data, which can predict the quality of polycrystalline silicon ingots quickly, accurately and at low cost.
In order to make the foregoing objectives, features, and advantages of the present invention more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific examples.
FIG. 1 is a schematic flowchart of a method for predicting quality of a polycrystalline silicon ingot based on ingredient data according to the present invention. As shown in FIG. 1, the method for predicting quality of a polycrystalline silicon ingot based on ingredient data according to the present invention includes the following steps.
S101: acquire ingredient data of a polycrystalline silicon ingot, where the ingredient data includes quality of primary polycrystalline bulk, quality of broken polycrystal, quality of fragments, quality of self-produced purified ingot cores, quality of purchased purified ingot cores, quality of top skin, quality of leftover materials, quality of tailings and minority carrier lifetime value.
S102: preprocess the ingredient data, where the preprocessing includes cleaning data to delete data with missing values.
Divide the preprocessed data to obtain numerical data and character data. The numerical data is ingredient quality data in the ingredient data; and the character data is data about labeled batch numbers in the ingredient data.
S103: reduce dimensions of the ingredient data by using a diffusion mapping algorithm. The diffusion mapping algorithm achieves dimension reduction by keeping a diffusion distance in the diffusion process as far as possible, aiming at defining a global relationship through a local relationship of sample points.
A specific process of the diffusion mapping algorithm includes:
constructing a neighbor graph, where for a data set X=x, 2 ,.. XA}used in the present invention, xi E RD, and i=1,2,..., N. An improved k-nearest neighbor standard is adopted in this algorithm. For the existing character data and numerical data, a Euclidean distance is used to calculate a distance between character data features, and a Mahalanobis distance is used to calculate a distance between numerical data features. If two points X and xi are neighbor points, a connecting line of the two points forms an edge to reflect a local relationship between points. Formulas of the Euclidean distance and Mahalanobis distance are shown as follows respectively, where S is a corresponding covariance matrix:
2 - (x,r-x)(x x DM(x,x, 1 )
A sparse weight matrix A is constructed. For and xi given, a Gaussian kernel function is used to define a similarity matrix between two sample points, namely:
D(x. - x A = A(xi,xj)= 2pu21)) 0 Au<
where U is a variance of a Gaussiankernel, is a critical value of the sparse matrix, which is generally 5, and an element value less than 0 is set to 0. An element A in the sparse weight matrix reflects a similarity between the sample points i and xj and satisfies the conditions of non negativity and symmetry.
A diffusion kernel matrix K"' is constructed. By a weighted graph Laplacian normalization method, matrix elements are obtained by the following formula:
K,'= K(x,,x.)= k
k. x where K represens a one-step transition probability from the sample point j to another and ' point X j, and KdK _(Ks)) )enotes a transition probability in the case of t-step random walk.
Dimensions are reduced on the basis of the diffusion distance. Based on the foregoing definition, the diffusion distance D (x,,x ) is constructed as follows. 2 (K' - K3k) k 99(XA)
The diffusion distance is kept unchanged, and the kernel matrix K' is subjected to feature decomposition to solve a characteristic value and a corresponding characteristic vector. Since A=1 is trivial, the corresponding characteristic vector 0 is removed, and the characteristic vector t), t)2 ' --' " corresponding to the remaining d maximum characteristic values ' .' ' Ad
is taken as a low-dimensional embedding result, to get the dimension-reduced data as X=[ 1 V 1 , .Z2V2 _'. Adtd]T
S104: build a support vector data description model with the dimension-reduced ingredient data in a set proportion. First normal class data is mapped into a high-dimensional space by a kernel function, and then a spherical description boundary containing as much normal data as possible is constructed in the high-dimensional space. For an unknown sample point, the distance between the unknown sample point and the sphere center is determined, and those located in the description boundary are regarded as normal data, otherwise, they are determined as abnormal data, thus realizing the prediction of abnormal data.
An optimization problem is constructed. Dimension-reduced normal ingredient data X={x, x 2,..., x},O<<N of polycrystalline silicon ingots with acceptable quality is used as
training data, and the optimization problem is constructed as shown in the following formula:
min F(R, a, R R2 +Clg,
s.t. I # (x,)- aII'-R2 +, j,>:0, i= 1, 2,...,l1
where R and a are respectively a radius and a sphere center of a hypersphere in the corresponding high-dimensional feature space, i is a slack variable, C>0 is a penalty parameter, and 0 is a mapping function. By solving the Lagrange dual problem, the foregoing optimization problem can be transformed into:
MaX a, (#(x,) -#Cx,)) -YYaa (#(x,).# (x,)) a =1 i=1 j=1
s.t. a =1,0 <a, K C, i=1,2..., i=1
A Lagrange multiplier ai can be obtained by solving the quadratic programming problem in the foregoing formula, so that the sphere center a and the radius R of a corresponding hypersphere can be solved, thereby obtaining information on the hypersphere.
Abnormal data is predicted through a decision function. A decision function is defined as:
2 f(x)=1|I#(x)-a11 -R
For an unknown target point x, the distance between the unknown target point and the sphere center is calculated. When f(x) 0, i.e., the target point is located in the spherical boundary, polycrystalline silicon ingots are defined as having acceptable quality; otherwise, the polycrystalline silicon ingots are defined as having unacceptable quality.
A specific process of obtaining a trained support vector data description model includes:
optimizing the support vector data description model by using a Monte Carlo algorithm to obtain the trained support vector data description model.
S105: perform product quality prediction classification of the dimension-reduced ingredient data by using the trained support vector data description model, to obtain product quality prediction classification results, where the product quality prediction classification results include polycrystalline silicon ingots with acceptable quality and polycrystalline silicon ingots with unacceptable quality; and the trained support vector data description model takes the dimension reduced ingredient data as inputs, and takes the product quality prediction classification results as outputs.
FIG. 2 is a schematic structural diagram of a system for predicting quality of a polycrystalline silicon ingot based on ingredient data according to the present invention. As shown in FIG. 2, the system for predicting quality of a polycrystalline silicon ingot based on ingredient data according to the present invention includes: an ingredient data acquisition module 201, a preprocessing module 202, a dimension reduction processing module 203, a model building module 204 and a product quality prediction classification module 205.
The ingredient data acquisition module 201 is configured to acquire ingredient data of a polycrystalline silicon ingot; where the ingredient data includes quality of primary polycrystalline bulk, quality of broken polycrystal, quality of fragments, quality of self-produced purified ingot cores, quality of purchased purified ingot cores, quality of top skin, quality of leftover materials, quality of tailings and minority carrier lifetime value.
The preprocessing module 202 is configured to preprocess the ingredient data, where the preprocessing includes cleaning data and deleting data with missing values.
The dimension reduction processing module 203 is configured to reduce dimensions of the preprocessed ingredient data by using a diffusion mapping algorithm.
The model building module 204 is configured to build a support vector data description model with the dimension-reduced ingredient data in a set proportion.
The product quality prediction classification module 205 is configured to perform quality prediction classification ofthe dimension-reduced ingredient data by using the trained support vector data description model, to obtain product quality prediction classification results, where the product quality prediction classification results include polycrystalline silicon ingots with acceptable quality and polycrystalline silicon ingots with unacceptable quality; and the trained support vector data description model takes the dimension-reduced ingredient data as inputs, and takes the product quality prediction classification results as outputs.
The system for predicting quality of a polycrystalline silicon ingot based on ingredient data according to the present invention further includes: a data dividing module and a module for determining a trained support vector data description model.
The data dividing module is configured to divide the preprocessed data to obtain numerical data and character data.
The module for determining a trained support vector data description model is configured to optimize the support vector data description model by using a Monte Carlo algorithm to obtain the trained support vector data description model.
Each example of the specification is described in a progressive manner, each example focuses on the difference from other examples, and the same and similar parts between the examples may refer to each other. Since the system disclosed in the example corresponds to the method disclosed in the example, the description is relatively simple. For relevant information, reference is made to the description of the method.
Specific examples are used for illustration of the principles and implementations of the present invention. The description of the examples is only used to help illustrate the method and its core ideas of the present invention. In addition, persons of ordinary skill in the art can make various modifications in terms of specific examples and scope of application in accordance with the teachings of the present invention. In conclusion, the content of this specification shall not be construed as a limitation to the present invention.
Claims (6)
1. A method for predicting quality of a polycrystalline silicon ingot based on ingredient data, comprising:
acquiring ingredient data of a polycrystalline silicon ingot, wherein the ingredient data comprises quality of primary polycrystalline bulk, quality of broken polycrystal, quality of fragments, quality of self-produced purified ingot cores, quality of purchased purified ingot cores, quality of top skin, quality of leftover materials, quality of tailings and minority carrier lifetime value;
preprocessing the ingredient data, wherein the preprocessing comprises cleaning data and deleting data with missing values;
reducing dimensions of the preprocessed ingredient data by using a diffusion mapping algorithm;
building a support vector data description model with the dimension-reduced ingredient data in a set proportion; and
performing product quality prediction classification of the dimension-reduced ingredient data by using the trained support vector data description model, to obtain product quality prediction classification results, wherein the product quality prediction classification results comprise polycrystalline silicon ingots with acceptable quality and polycrystalline silicon ingots with unacceptable quality; and the trained support vector data description model takes the dimension reduced ingredient data as inputs, and takes the product quality prediction classification results as outputs.
2. The method for predicting quality of a polycrystalline silicon ingot based on ingredient data according to claim 1, wherein the preprocessing comprises cleaning data and deleting data with missing values, and after the preprocessing, the method further comprises:
dividing the preprocessed data to obtain numerical data and character data.
3. The method for predicting quality of a polycrystalline silicon ingot based on ingredient data according to claim 1, wherein after the building a support vector data description model with the dimension-reduced ingredient data in a set proportion, the method further comprises:
optimizing the support vector data description model by using a Monte Carlo algorithm to obtain the trained support vector data description model.
4. A system for predicting quality of a polycrystalline silicon ingot based on ingredient data, comprising: an ingredient data acquisition module, configured to acquire ingredient data of a polycrystalline silicon ingot; wherein the ingredient data comprises quality of primary polycrystalline bulk, quality of broken polycrystal, quality of fragments, quality of self-produced purified ingot cores, quality of purchased purified ingot cores, quality of top skin, quality of leftover materials, quality of tailings and minority carrier lifetime value; a preprocessing module, configured to preprocess the ingredient data; wherein the preprocessing comprises cleaning data and deleting data with missing values; a dimension reduction processing module, configured to reduce dimensions of the preprocessed ingredient data by using a diffusion mapping algorithm; a model building module, configured to build a support vector data description model with the dimension-reduced ingredient data in a set proportion; and a product quality prediction classification module, configured to perform quality prediction classification of the dimension-reduced ingredient data by using the trained support vector data description model, to obtain product quality prediction classification results; wherein the product quality prediction classification results comprise polycrystalline silicon ingots with acceptable quality and polycrystalline silicon ingots with unacceptable quality; and the trained support vector data description model takes the dimension-reduced ingredient data as inputs, and takes the product quality prediction classification results as outputs.
5. The system for predicting quality of a polycrystalline silicon ingot based on ingredient data according to claim 4, further comprising:
a data dividing module, configured to divide the preprocessed data to obtain numerical data and character data.
6. The system for predicting quality of a polyrystalline silicon ingot based on ingredient data according to claim 4, further comprising:
a module for determining a trained support vector data description model, configured to optimize the support vector data description model by using a Monte Carlo algorithm to obtain the trained support vector data description model.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010303765.8 | 2020-04-17 | ||
| CN202010303765.8A CN111639065B (en) | 2020-04-17 | 2020-04-17 | Polycrystalline silicon ingot casting quality prediction method and system based on batching data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| AU2020101671A4 true AU2020101671A4 (en) | 2020-09-10 |
Family
ID=72331458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| AU2020101671A Active AU2020101671A4 (en) | 2020-04-17 | 2020-08-05 | Method and System for Predicting Quality of Polycrystalline Silicon Ingot Based on Ingredient Data |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111639065B (en) |
| AU (1) | AU2020101671A4 (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112365789A (en) * | 2020-11-09 | 2021-02-12 | 中国科学院城市环境研究所 | Dynamic demonstration method and system for concentration and components of water-soluble ions of aerosol |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101216436A (en) * | 2008-01-03 | 2008-07-09 | 东华大学 | Fabric flaw automatic detection method based on Support Vector data description theory |
| CN101609510B (en) * | 2009-07-15 | 2012-01-11 | 北京交通大学 | Dimensionality reduction method of image and video |
| KR102015397B1 (en) * | 2013-06-28 | 2019-10-21 | 엘지디스플레이 주식회사 | Organic light emitting display device and method for driving the same |
| CN104091078B (en) * | 2014-07-14 | 2017-12-12 | 西北工业大学 | Product Multi-information acquisition indicating failure means to save the situation based on D S evidence theories |
| CN106446931A (en) * | 2016-08-30 | 2017-02-22 | 苏州大学 | Feature extraction and classification method and system based on support vector data description |
| CN108010029B (en) * | 2017-12-27 | 2020-11-03 | 江南大学 | Fabric defect detection method based on deep learning and support vector data description |
| JP2019185384A (en) * | 2018-04-10 | 2019-10-24 | キヤノン株式会社 | Image authentication device, image authentication method, computer program and storage medium |
| CN110297907B (en) * | 2019-06-28 | 2022-03-08 | 谭浩 | Method for generating interview report, computer-readable storage medium and terminal device |
-
2020
- 2020-04-17 CN CN202010303765.8A patent/CN111639065B/en active Active
- 2020-08-05 AU AU2020101671A patent/AU2020101671A4/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111639065B (en) | 2022-10-11 |
| CN111639065A (en) | 2020-09-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107092923B (en) | The electric melting magnesium furnace process monitoring method of method is locally linear embedding into based on improvement supervision core | |
| JP4584295B2 (en) | Two-stage virtual measurement method | |
| CN106054840B (en) | A kind of online managing and control system of whole process product quality | |
| AU2020101671A4 (en) | Method and System for Predicting Quality of Polycrystalline Silicon Ingot Based on Ingredient Data | |
| CN114897227B (en) | Multi-steel mechanical property forecasting method based on improved random forest algorithm | |
| CN107067100A (en) | Wind power anomalous data identification method and device for identifying | |
| CN110175682A (en) | A kind of optimization core pivot element analysis fault monitoring method based on Chaos-Particle Swarm Optimization | |
| CN111652518B (en) | Method and apparatus for analyzing process data and computer readable medium | |
| US20230359183A1 (en) | Method, system and non-transitory computer-readable medium for reducing work-in-process | |
| CN115358297A (en) | Injection molding machine abnormity detection method and system based on improved MKECA method | |
| CN102621953B (en) | Automatic online quality monitoring and prediction model updating method for rubber hardness | |
| CN114004044A (en) | Quick identification method for thermal error of machine tool spindle based on temperature sensitive point | |
| TW202212020A (en) | Quality designing method and electrical device | |
| Chang | Collaborative decision making algorithm for selection of optimal wire saw in photovoltaic wafer manufacture | |
| CN116401570B (en) | Intelligent processing system for printing quality monitoring big data | |
| CN116579654A (en) | Online intelligent quality monitoring method and system for IF steel | |
| Dong et al. | Just-in-time learning-based soft sensor for mechanical properties of strip steel via multi-block weighted semisupervised models | |
| CN108447737B (en) | A kind of relay base quality optimization system based on simplex search | |
| CN115905802A (en) | Semiconductor wafer test yield analysis method based on thermodynamic diagram | |
| JP2002323924A (en) | Method and device for detecting defective device, program, and method for manufacturing product | |
| CN114118844A (en) | Automobile parts quality analysis system | |
| TWI689888B (en) | Method for determining abnormal equipment in semiconductor processing system and program product | |
| CN117831659B (en) | Method and device for online detection of quality of wide and thick plates, electronic equipment and storage medium | |
| CN113189967B (en) | Control performance diagnosis method for semiconductor process batch control system | |
| CN115687899B (en) | Hybrid feature selection method based on high-dimensional spinning data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FGI | Letters patent sealed or granted (innovation patent) |