Method for calculating scrap steel recovery rate in steelmaking
Patent Information
- Application Number
- TW115111107
- Authority / Receiving Office
- TW · TW
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2026-03-24
- Publication Date
- 2026-11-11
- Estimated Expiration
- 2046-03-23
Smart Images

Figure TWG2TB001921127_001 
Figure TWG2TB001921127_002
Abstract
Claims
1. A method for calculating the scrap steel recycling rate in a steelmaking process, comprising: providing scrap steel raw materials and a steelmaking process to obtain the individual input masses (MScrap,i) of a plurality of scrap steels input into the steelmaking process, and a set of operating parameters of the steelmaking process; obtaining the actual steel output (Msteel) after the steelmaking process is completed; establishing a mass balance correlation based on the law of conservation of mass, wherein the mass balance correlation is based on the balance relationship between the individual metal recovery rates (Ri) of the plurality of scrap steels and the total mass of the individual input masses; using a data-driven machine learning model to correct and solve the mass balance correlation based on the set of operating parameters, to predict a deviation (bias) between the mass balance correlation and the actual steel output, thereby dynamically correcting the deviation and calculating the individual metal recovery rates of the plurality of scrap steels.
2. The method for calculating the scrap recycling rate in the steelmaking process as described in claim 1, wherein the mass balance correlation satisfies the following relationship: Wherein, i represents the i-th type of scrap steel, N is the total number of scrap steel types, and ∆Mbias is the bias predicted by the machine learning model.
3. The method for calculating the scrap recycling rate for steelmaking processes as described in claim 1, wherein the group of operating parameters for the steelmaking process includes at least one of electrical energy consumption, oxygen consumption, combustion aid usage, or carbon powder injection.
4. The method for calculating the scrap recycling rate in a steelmaking process as described in claim 1, wherein the group of operating parameters of the steelmaking process and the deviation are non-linearly related.
5. The method for calculating the scrap recycling rate in the steelmaking process as described in claim 4, wherein the machine learning model learns the nonlinear correlation to dynamically correct the deviation.
6. The method for calculating the recycling rate of scrap steel in the steelmaking process as described in claim 4, wherein the machine learning model learns the nonlinear correlation and dynamically corrects the deviation to calculate the individual metal recycling rate of the plurality of scrap steels.
7. The method for calculating the scrap recycling rate in the steelmaking process as described in claim 1, wherein the machine learning model is selected from a group of algorithms consisting of Gradient Boosting Decision Tree (GBDT), Support Vector Regression (SVR), ANN-like Neural Networks (ANN), or Random Forest.
8. The method for calculating the scrap recovery rate in the steelmaking process as described in claim 1, wherein the data-driven machine learning model utilizes production data from multiple historical furnace runs and, through iterative calculation and optimization algorithms, treats the individual metal recovery rate as the data-driven machine learning model.
9. The method for calculating the scrap steel recycling rate for steelmaking processes as described in claim 1, wherein the data-driven machine learning model is used for calculation after being driven by data, and the calculation includes calculating the individual metal recovery rate of the plurality of scrap steels and dynamically updating the deviation amount, at least one of the following:
10. The method for calculating the scrap recycling rate in a steelmaking process as described in claim 1, wherein a coefficient of determination between the predicted steel output and the actual steel output predicted by the data-driven machine learning model is not less than 0.8.
Citation Information
Patent Citations
Deformed steel bar life cycle environment influence evaluation method
CN115983673A
Multi-sensor fused intelligent steel scrap sorting system
CN120838708A
Scrap steel value accounting method based on multi-dimensional characteristics
CN120996839A
Iron and steel product carbon footprint management method and system based on life cycle assessment
WO2024244045A1
