# US 2024/0002964 A1 · Method and system for determining converter tapping quantity

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2024/0002964 A1
- Document ID: US20240002964A1
- Status: Published application
- Kind: A1
- Application number: 18/308,784
- Filed: 2023-04-28
- Published: 2024-01-04
- Earliest claimed: 2022-07-01
- Assignees: University of Science and Technology Beijing
- Applicants: University of Science and Technology Beijing
- Inventors: Yanping BAO; Ruixuan ZHENG; Lihua ZHAO
- Web page: https://patentgrep.com/patent/US20240002964A1

## CPC

- C21C 5/30: Regulating or controlling the blowing
- G06N 3/048: Activation functions
- G06N 3/08: Learning methods
- G06N 5/00: Computing arrangements using knowledge-based models
- C21C 2300/06: Modeling of the process, e.g. for control purposes; CII
- Y02P 90/30: Computing systems specially adapted for manufacturing
- C21C: PROCESSING OF PIG-IRON, e.g. REFINING, MANUFACTURE OF WROUGHT-IRON OR STEEL; TREATMENT IN MOLTEN STATE OF FERROUS ALLOYS
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- Y02P: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS

## Abstract

The invention relates to a method and a system for determining the steel-tapping quantity of a converter, which consider that the working environment of the steel-making process of the converter is severe, the measurement is difficult and the interference of other factors is large, and provide a data-driven prediction model based on data, combine a Principal Component Analysis (PCA) with a RBF neural network, find the relation and the internal relation among variables by carrying out mathematical analysis on the related internal structure of the original variables, can quickly and accurately realize the prediction of the steel-tapping quantity of the converter, improve the component hit rate and the product stability in the steel-making process of the converter, are beneficial to realizing the control of narrow regions of steel-making components, save the alloying cost and have good application prospects in the field of ferrous metallurgy.

## Claims

### Claim 1

A method for determining the tapping amount of a converter, characterized in that it comprises the following steps:
S1. Collect converter production data sets and establish a prediction model database; The converter production data set includes: blowing cycle, oxygen supply time, total oxygen content, whether to blow carbon or not;
S2. Conduct data screening and elutriation on the collected converter production data set, and conduct normalization preprocessing on the screened and elutriated data;
S3. Determine the process parameter variables that affect the converter tapping quantity as input variables to the model;
The method for determining the input variables of the model is to analyze the correlation between the process parameter variables in step S1 and the converter tapping amount through a bivariate correlation analysis method;
S4. Perform dimensionality reduction processing on converter production process data using principal component analysis;
The processing steps of the principal component analysis method are as follows:
1) Data standardization processing, and the obtained standardization matrix;
2) Calculate the correlation coefficient matrix of the standardization matrix;
3) Solving the eigenvalues and eigenvectors of the correlation coefficient matrix;
4) Calculate the variance contribution rate and cumulative variance contribution rate of the principal component variables;
Select the first m principal component variables with a cumulative variance contribution rate greater than 80% as the target variable after dimensionality reduction;
Establish a RBF neural network prediction model for converter tapping capacity;
S6. Predictive model training and testing;
S7. Collect real-time data of on-site smelting process;
S8. Perform principal component analysis (PCA) dimensionality reduction processing on the collected real-time data;
S9. The prediction model of converter tapping quantity based on RBF neural network is applied to predict the converter tapping quantity;
S10. Ferroalloy is added during the tapping process based on the prediction results of converter tapping volume. After tapping, the data is stored in the prediction model database, and the prediction model is regularly updated.

### Claim 2 (depends on 1)

The method for determining the tapping amount of a converter according to claim 1, characterized in that in step S2.
Perform data screening and elutriation on the collected converter production dataset, specifically including:
Delete duplicate data, delete abnormal points and extreme abnormal points, and judge abnormal data according to formula (1):
{
x
i
>
U
+
1.5
IQR
❘
x
i
<
L
+
1.5
IQR
Abnormal
⁢
points
IQR
=
U
-
L
x
i
>
U
+
3
⁢
IQR
❘
x
i
<
L
+
3
⁢
IQR
Extreme
⁢
anomaly
⁢
points
(
1
)
Where, xi refers to the ith data after the data to be tested is sorted from small to large, U refers to the upper quartile, L refers to the lower quartile, and IQR refers to the interquartile distance.

### Claim 3 (depends on 1)

The method for determining the tapping amount of a converter according to claim 1, characterized in that in step S2, the data after screening and elutriation are preprocessed, specifically comprising:
Normalize the dataset to [−1, 1]. The specific processing method is shown in Equation (2):
x
ij
′
=
x
ij
-
mean
(
x
ij
)
max
⁡
(
x
ij
)
-
min
⁡
(
x
ij
)
(
2
)
Where xij is the raw data of the production data sample; x′ij is the normalized data; i represents the ith sample point; j represents the jth input variable; min(xij), max(xij) and mean(xij) are the minimum, maximum, and average values in the original data sample, respectively.

### Claim 4 (depends on 1)

The method for determining converter tapping capacity according to claim 1, characterized in that in step S4, the principal component analysis method is a classical multivariate statistical technique that reduces data dimensions by summarizing the characteristics of the data, and the processed data is converted into a new set of principal component variables, the variables being uncorrelated and orderly.

### Claim 5 (depends on 1)

The method for determining the tapping amount of a converter according to claim 1, characterized in that in step S5, the RBF neural network is as shown in formula (3):
y=Σk=1Kwkθk(x) (3)
Where, x=(x1, x2, . . . , xM)T represents the input variable of the network, M is the number of input variables; y represents the output of the network, w=(w1, w2, . . . , wM)T represents the connection weights of the hidden layer neurons and the output layer; B k (x) is the output value of the k th hidden layer neuron.

### Claim 6 (depends on 1)

The method for determining converter tapping quantity according to claim 1, which is characterized in that in step S5, the RBF neural network uses Gaussian function as the radial basis function, and uses the orthogonal least square method to select neurons to determine the RBF neural network to determine the hidden layer of the RBF neural network.

### Claim 7 (depends on 1)

A converter tapping quantity determination system implementing the method for determining the converter tapping quantity according claim 1.

### Claim 8 (depends on 1)

An information data processing terminal for realizing the method for determining the tapping amount of the converter according to claim 1.
