Prebaked Anode Formulation Data Evaluation And Analysis

As a supporting industry in the electrolytic aluminum industry,prebaked anodes are inseparable from the progress of electrolytic aluminum technology.In the prebaked anode carbon block production process,calcination,roasting,molding and assembly processes are required.The main research stage in this dissertation is the roasting stage,which aims to analyze the relationship between the proportions of calcined petroleum coke with different particle sizes and the physicochemical indexes of roasting blocks generated after roasting by using data mining theory,and establishes and optimizes the prebaked anode recipe for roasting.Block prediction assessment model to assess the pros and cons of the formula,to provide reference for the optimization of pre-baked anode formula,to reduce,waste of resources,increase production efficiency,and contribute to the development of aluminum electrolysis industry.This paper studies the related technical and theoretical knowledge of prebaked anode roasting process,including prebaked anode production process,quality standard of roasting block quality assessment and relevant data mining knowledge.The main data mining algorithms used in this project include K-means clustering algorithm,BP neural network and KNN classification algorithm.First of all,this paper analyzes and evaluates the quality parameters of the calcined block,and tests whether the calcined block is qualified according to the enterprise standard,and uses ash,bulk density,resistivity,true density,compressive strength,C02 reactivity,and thermal expansion coefficient as dimensions.The-means clustering algorithm classifies the roasting blocks and calculates the class basis points.Finally,the roasting block quality is evaluated using the roasting block class-weighted evaluation method proposed in this paper.Secondly,a predictive assessment model of BBP anode formula baking block was established to explore the intrinsic relationship between anode formulation and baking block quality scores.Then,the traditional KNN classification algorithm is optimized and integrated with the BP neural network,and an optimized model BBPK model for the BBP model is proposed.In the BBPK model,KNN was used to classify the physical and chemical indicators of the roasting blocks output by the neural network and the roasting blocks were subjected to category-weighted scoring to assess the quality of the roasting blocks fired under the new recipe.The BBP model and the BBPK model were used to carry out the experiments and the results of the comparative experiments were analyzed.It was demonstrated that the optimization of the model for predicting and evaluating the roasted anode blocks was significant.Finally,using software engineering design ideas,using SSM architecture,building a lightweight software development model,designing pre-baked anode formulations to evaluate and analyze Web system applications,predicting the evaluation of recipe data,visualizing predictive assessment data,and providing enterprise production decisions help.