Cement grinding mill system power consumption index prediction method based on XGBoost

Abstract

The invention discloses a cement grinding mill system power consumption index prediction method based on XGBoost. The cement grinding mill system power consumption index prediction method comprises the following steps: selecting eight variables related to the power consumption of the cement mill; collecting required variable data by adopting an OPC technology; removing abnormal data by adopting anartificial experience removal method and criterion; constructing an XGBoost model input and output layer; initializing a weight parameter according to the sample data; training a first tree accordingto the weight; updating the weight parameter according to the target function after the training is finished; performing a new round of decision tree training; when the weight sum of the samples is smaller than a set threshold value or the number of iterations reaches a set value, stopping tree building; completing the training of the XGBoost model; and substituting the industrial field sample data set into the trained model to complete the online prediction of the power consumption index of the cement grinding mill system, training the sample data through XGBoost, and inputting the variabledata of the actual cement production field into the trained model to realize the online prediction of the power consumption index of the cement grinding mill.

Description

technical field [0001] The invention relates to an XGBoost-based method for predicting the power consumption index of a cement mill system, and belongs to the field of predicting the power consumption index of a cement mill system. Background technique [0002] According to the relevant data of the cement industry, my country's cement output ranks among the top in the world, and the cement mill system has been widely used in contemporary cement production. Realizing the online prediction of the power consumption index of the cement mill system is conducive to guiding the scheduling of various parameters in the cement mill grinding process Optimization is conducive to reducing the power consumption of the cement mill production process, so as to achieve the purpose of saving energy, reducing emissions and improving production efficiency. However, due to the hysteresis in the grinding process of the cement mill system and the strong coupling of multiple industrial variables, it...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved in the present invention is to provide a method for predicting the power consumption index of the cement mill system based on XGBoost (eXtreme GradientBoosting), which not only solves the problem of the hysteresis of the cement mill system, but also solves the strong coupling between multiple variables in the cement mill system At the same time, compared with the neural network model, XGBoost has a fast calculation speed. During training, all CPU cores can be used to parallelize tree building, which greatly improves the training rate. For missing value features in industrial data, it can automatically learn missing values ​​by enumeration. Value splitting direction, adding L1 and L2 regularization items to the algorithm to prevent overfitting and enhance generalization ability

Images