Research And Application Of Data Correction Method For Copper Smelting Process

Process data plays an important role in copper smelting production.Its accuracy is related to the production safety and quality of copper smelting process,and directly affects the long-term business decision of copper smelting enterprises.However,process data will be inevitably interfered by random errors and significant errors in the process of acquisition and recording,which makes it unable to satisfy the constraints of material balance or element balance.Data correction technique integrates redundant data preprocessing,data reconciliation,and gross error detection technique to discover,trace,and eliminate errors,so that the balance constraints can be satisfied to improve the quality of process data.Data correction technique provides a solid foundation for effective business accounting of copper smelting enterprises.Based on a domestic copper smelting company,this paper analyzes the data form and balance relationship of the company's copper smelting process.On one hand,its redundant data needs to be pre-processed before reconciliation to improve the validity and consistency of the data.On the other hand,the data reconciliation model of copper smelting process is nonlinear with a large number of variables.General solving methods are not applicable or with poor performance.In order to solve the pre-processing problems of redundant data such as matching,regulating and abnormal tracing,this paper proposes segmented adaptive dynamic time warping(DTW)algorithm,which is consisted with two parts: adaptive DTW algorithm and segmented subsequence matching algorithm.The adaptive DTW algorithm adaptively slides and aligns sequences with a variable-length window on the basis origin of DTW algorithm.Based on the adaptive DTW algorithm,the segmented subsequence matching algorithm performs sliding and matching segmentally on the main sequence to efficiently complete subsequence matching.The results of experiment show that the proposed method can not only accurately and adaptively align the sequences,but also greatly reduce the time consumption of subsequence matching,which is highly robust and practical for redundant data matching.Since the optimization problem corresponding to the data reconciliation model of copper smelting is nonlinear with a large number of variables,this paper proposes multiplier conversion algorithm.The method transforms the original high-dimensional optimization problem into a low-dimensional problem by Lagrange multiplier transformation,and the original variables are restored and bounded by improved method.The converted problem is solved by improved differential evolution(DE)algorithm and its variable dimension is directly related to the number of constraints in original problem.The results of experiment show that the proposed method is stable,robust and fast,which can effectively deal with the nonlinear data reconciliation problem of copper smelting.Finally,the paper describes the design and implementation of the data correction system in the SAP platform of a domestic copper smelting company,which is based on the methods proposed in this paper.With the data correction system,the copper smelting company is able to accurately and efficiently solve the data correction problem,which provides good theoretical and data support for its balance accounting and business decision-making.