Ball mill load parameter soft measurement method based on signal decomposition and Gaussian process

Abstract

The invention provides a ball mill load parameter soft measurement method based on signal decomposition and a Gaussian process, and relates to the technical field of ball mill load parameter measurement. The method comprises an off-line modeling stage and an on-line measurement stage. In the offline modeling stage, ball mill vibration sound signals are obtained through a large number of experiments, and the vibration signals and corresponding load parameters serve as a sample set; using adaptive noise complete ensemble empirical mode decomposition CEEMDAN to decompose the vibration sound signal and the vibration signal into a series of intrinsic mode functions IMF and residual errors; then establishing a Gaussian mixture model, classifying low-dimensional spectrum features of an intrinsic mode function (IMF), and determining a total residual error of a sample; and finally, establishing a Gaussian process regression model based on the classified overall low-dimensional spectrum features of each training sample to realize soft measurement of the load parameters of the ball mill. In the online measurement stage, the Gaussian process regression model in the offline modeling stage is used for carrying out soft measurement on ball mill parameters, and ball mill load parameters and confidence intervals are output.

Description

technical field [0001] The invention relates to the technical field of ball mill load parameter measurement, in particular to a ball mill load parameter soft measurement method based on signal decomposition and Gaussian process. Background technique [0002] Wet ball mill is a ball mill widely used in the domestic steel production process. The grinding process is used as the pre-process of the beneficiation process. The ball mill is used to grind the crushed ore into raw materials of suitable particle size. The grinding process is the bottleneck of the whole beneficiation process, and it is also the main energy consumption link in the beneficiation process. In addition to power consumption, the steel ball consumption of the ball mill is also a factor that cannot be ignored. Studies have shown that the steel ball consumption of the ball mill is usually proportional to the energy consumption of the ball mill. The load parameters such as the material level in the barrel of the...

AI Technical Summary

Problems solved by technology

The existing ball mill soft measurement methods are commonly used: feature extraction based on principal component analysis and other methods, and then approximated by regression models such as support vector machines, forward neural networks, and extreme learning machines. Poor, easy to cause misjudgment; Based on time-domain decomposition such as empirical mode decomposition, the aforementioned regression model approximation method is used after decomposition. This type of method has the problems of decomposition signal mode aliasing and difficulty in modal function classification. Misclassification of function gets anomalous measurements

Images