Process Parameters Of Calcining Based On Decision Tree Research And Analysis

Effectively improving the quality of calcined petroleum coke is the goal of the company's prebaked anode calcination stage.The time-based dimensions of the rotary kiln will form data in the form of hourly reports,class reports,daily reports,etc.These data are large and cannot be intuitively found in the data.With the rapid development of network communication technology and data storage technology,data mining technology has emerged.Through this technology,data information can be ed into knowledge that can be understood by management personnel,so that efficient decision-making can be made for decision-making in enterprise production.guide.Regression is an important predictive data method and one of the main purposes of data mining.There is a certain relationship between the process parameters recorded in the rotary kiln and the physical and chemical indexes of the calcined petroleum coke.The physical and chemical indexes of the calcined petroleum coke can be predicted through the process parameter data.Find out the relationship between data and build a mapping model between independent variable and dependent variable,which is of great significance for enterprises to achieve profit targets.In this paper,the method of decision tree regression is used to excavate the data relationship between rotary kiln process parameters and calcined petroleum coke physical indicators to provide guidance for management personnel to make decisions on production operations.The main work of this dissertation is the processing of data,summarized as follows:1.The process of data preprocessing of the data set before inputting the data tuple into the model learning:detecting abnormal points,paying attention to the impact of the noise data on the training results of the training set and the experimental set;using the average of the features to the missing data The value is filled.2.This paper proposes the optimization of the random forest algorithm by adjusting the temperature of the calcining zone temperature(%)of the rotary kiln to improve the priority of the feature.The process parameters of the rotary kiln during calcination and the physical and chemical characteristics of the calcined petroleum coke The mapping and the excellent regression results of the characteristic data of calcined petroleum coke physical indicators.3.Using CART algorithm and random forest algorithm to predict the physical and chemical indicators of calcined petroleum coke,and evaluate the learner.Taking process parameters as independent variables,physical and chemical indicators of calcined petroleum coke form a mapping for strain variables,and regression prediction of process parameter values by means of fitting methods,and the prediction results and actual values are evaluated,and it is found that using random forest algorithm,the overall Predictive accuracy reaches 92%,and good test results are achieved...