Method for predicting molten iron quality index interval of blast furnace ironmaking system under chance constraint
By introducing the LSTM and PSO algorithms to construct an interval prediction model, the uncertainty problem of predicting molten iron quality indicators in the blast furnace ironmaking system is solved, and more accurate and reliable prediction results are achieved, which is suitable for complex industrial processes.
Patent Information
- Application Number
- CN202510735168.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-04
- Publication Date
- 2025-09-12
AI Technical Summary
Traditional methods for predicting molten iron quality indicators in blast furnace ironmaking systems are unable to fully reflect the uncertainty within the model, resulting in inaccurate and unreliable prediction results. In particular, when faced with raw material fluctuations, data measurement errors and external interference, time series prediction methods are unable to provide an effective uncertainty range.
The long short-term memory neural network (LSTM) and particle swarm optimization (PSO) algorithm are used to construct an interval prediction model. By calculating the probability density function and cumulative distribution function, the upper and lower bound difference range under a specific confidence level is determined. The particle swarm optimization algorithm is then used to find the optimal model parameters to minimize the width of the prediction interval.
It significantly improves the prediction accuracy and reliability of the blast furnace ironmaking system, can quantify the uncertainty of the prediction, enhances the robustness of the model in the face of random interference, and provides more flexible and secure decision support.