Resin process prediction method and device based on product performance and process data model

Abstract

Disclosed is a resin process prediction device based on product performance and process data model which comprises the following steps: collecting a production process and process data; transforming characteristic value of the collected continuous change data; carrying out sampling inspection on batch products, and recording performance indexes and quality indexes of the batch products; respectively establishing neural network models for the production process data, the performance data and the quality data of the plurality of batches; evaluating and verifying the neural network model, and optimizing the neural network model; defining a search interval of a performance parameter, and searching data closest to the performance index in the search interval as a data set; according to the original parameter data in the data set, calculating a value interval of each process parameter; and searching from the lower limit of the interval, and calculating product performance indexes through the optimized neural network model to obtain a process parameter value interval. According to the invention, the problems of difficult optimization of the current production process and low efficiency of the formula experiment process are solved, and the product quality and the performance stability are improved.

Description

technical field [0001] The invention relates to a resin process prediction method and device based on product performance and process data models, and belongs to the technical field of process data mining. Background technique [0002] Prepreg is an intermediate material for the manufacture of composite materials and their parts. It constitutes the basic unit of composite materials. The uniformity and stability of its quality is an important link to ensure the quality and reliability of composite materials and their parts. [0003] Carbon fiber prepreg is made of synthetic resin, carbon fiber woven fabric, epoxy resin / phenolic resin / cyanate resin, and release paper. It is usually produced to order according to the performance requirements of the customer for the material, because The customer's performance requirements for the product are based on the actual working conditions, so it is necessary to prepare the resin required for the production of carbon fiber cloth accordin...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above problems, the present invention proposes a resin process prediction method and device based on product performance and process data model, which can solve the problems of difficult optimization of current production process and low efficiency of formula experiment process

Images