Xianzhong Chen

Prediction of the hot metal silicon content in blast furnace based on extreme learning machine

Silicon content in hot metal is an important indicator for the thermal condition inside the blast furnace in the iron-making process. The operators often refer the silicon content and its change trend for the guidance of next production. In this paper, we establish the neural network model for the prediction of silicon content in hot metal based on extreme learning machine (ELM) algorithm. Considering the imbalanced operating data, weighted ELM (W-ELM) algorithm is employed to make prediction for the change trend of silicon content. The outliers hidden in the real production data often tend to undermine the accuracy of prediction model. First, an outlier detection method based on W-ELM model is proposed from a statistical view. Then we modified the ordinary ELM and W-ELM algorithms in order to reduce the interference of outliers, and proposed two enhanced ELM frameworks respectively for regression and classification applications. In the simulation part, the real operating data is employed to verify the better performance of the proposed algorithm.

Blast furnace stockline measurement using radar

Abstract This paper presents a synergistic approach to stockline depth tracking within a blast furnace. Frequency modulated continuous wave (FMCW) radar can be used to measure the depth and surface profile of the burden surface; however, the radar signal is easily disturbed by radar anomalies during the process of continuous measurement. Data from the rotating chute and the charging signal provide information on the contextual relevance of these anomalies. An improved Kalman filter and anomaly detection model were developed to increase measurement accuracy by utilising contextual information. The approach was validated on production blast furnaces. The root mean squared (RMS) error in the measured depth was reduced by 17% when the proposed approach is used. The results suggest that this approach successfully adapts to changes in the pattern and characteristics of the burden surface.

Orthogonal Waveform Design Based on the Modified Chaos Genetic Algorithm for MIMO Radar

In view of the traditional genetic algorithm easily fall into local optimum in the late iterations, this paper puts forward an improved chaos genetic algorithm coded orthogonal signal design method which combines the chaos theory and genetic algorithm for MIMO radar. In order to prevent and overcome the ‘premature’ phenomenon in the process of optimization, the traversal features of the chaos optimization is introduced to the genetic algorithm, which reduces the autocorrelation peak side lobe and cross-correlation peak. Simulation results show that the proposed algorithm is feasible and effective.