Classification and Prediction Method of 3D Material Surface of Blast Furnace

Due to the lack of blast furnace data, unknown mechanism and inaccurate model, it is a typical black box system, and the blast furnace operator has long relied on experience to operate the fabric and then control the gas flow distribution, which makes the furnace condition usually in a reasonable-deteriorating-reasonable iterative state. It is difficult to realize the optimization of blast furnace operation. In this paper, in order to solve the above problems, we propose a method for classifying and judging blast furnace 3D material surface based on deep learning 3D point cloud algorithm DGCNN, which classifies blast furnace material surface into four categories based on the data collected by phased array radar used at the blast furnace site, and uses deep learning neural network to learn so that it can accurately determine whether the 3D material surface is abnormal and analyze the abnormal areas and causes, and uses a variety of methods for blast furnace 3D material surface This paper proposes a new data enhancement method to meet the actual situation in the field by using various data enhancement methods for blast furnace three-dimensional material surface to improve the accuracy. This paper proposes a method for classifying and determining whether the material surface is abnormal or not using deep learning neural network, which has the advantage of fast and accurate.