Adjustment method of melt spinning process based on improved elm

Abstract

The invention relates to an improved ELM-based adjusting method for a melt spinning process. First, technology parameters and performance index data of existing fibers are collected and normalized; with the technology parameter data as an input sample and the performance index data as an output sample, an input layer weight and a hidden layer threshold value of ELM are optimized by an immune optimization algorithm to achieve improved ELM; then, based on the improved ELM, a database including technology parameters and performance indexes is built; and at last, the technology parameters corresponding to the pre-achieved performance indexes are researched from the database, and melt spinning is conducted according to the technology parameters. An antibody of the immune optimization algorithm is the input layer weight and the hidden layer threshold value and an antigen is an absolute error function Fv between an output value and a real value of the ELM. The ELM input layer weight and the hidden layer threshold value are optimized by the immune optimization algorithm, so mean square error and root-mean-square error of a sesquialter rate can be reduced; and technology adjusting precision can be enhanced.

Description

technical field [0001] The invention belongs to the field of intelligent optimization, and relates to an adjustment method for an improved ELM-based melt spinning process, in particular to an adjustment method for optimizing an improved ELM melt-spinning process based on an immune optimization algorithm. Background technique [0002] Due to its good performance, polyester fiber is widely used in textile products, industry, agriculture and other fields. The polyester fiber melt spinning process is the most critical link in the whole production process, which determines the quality of the primary fiber. At present, in the production line of polyester fiber production, the corresponding relationship between the main influencing factors in the melt spinning process and product performance indicators is often based on production experience, and the prediction of performance indicators is often carried out through manual sampling and measurement or analysis using instruments and e...

AI Technical Summary

Problems solved by technology

Using the immune optimization algorithm to optimize the input layer weight and hidden layer threshold of ELM can enhance the prediction accuracy of the model and improve the prediction speed of the model, but the immune algorithm often does not make full use of the prior knowledge of the research object. In data processing, artificial immune Generally, the system can only concentrate data samples (obtain feature points in the sample space), but cannot optimize the composition of the sample space (reasonably reduce the dimension of the sample space)

Images