Data-driven feature-based PCBA (printed circuit board assembly) maintenance work hour prediction method

Abstract

The invention discloses a data-driven feature-based PCBA (Printed Circuit Board Assembly) maintenance working hour prediction method. The method comprises the following steps: S1, determining a maintenance working mode of a PCBA; S2, establishing an evaluation index system to obtain weights of different evaluation indexes; S3, acquiring a plurality of job records, endowing scores for evaluation indexes in the plurality of job records, selecting one job record as a reference job, taking the rest job records as a sample job set, and then calculating a similarity coefficient; and S4, obtaining a PCBA board maintenance work time prediction model. According to the method, the concept of similarity is introduced into the PCBA board maintenance field, the weight of the man-hour influence factors is determined in a questionnaire survey mode, and the man-hour influence factors of each maintenance operation are scored. By calculating the similarity coefficient of the maintenance operation, the function relationship between the man-hour and the similarity coefficient is determined, and the maintenance man-hour can be predicted through the function relationship as long as the similarity coefficient of the maintenance operation with unknown man-hour is obtained.

Description

Technical field [0001] The present invention relates to the field of work-based prediction techniques, and more particularly to a data-driven-based feature-based PCBA board maintenance work time prediction method. Background technique [0002] Currently, there are many types of products that require substantures, but there are similarities in different models. In the process of product testing, there will always be some products that cannot be tested because of improper operation, machine failure, etc., and these products are determined after the DEBUG process determines the fault location and repairs. If it is a PCBA board failure problem, it will be sent to the returning room for repairs; if it is other problems, the product will return to the production line maintenance. In addition to the PCBA board directly sent from the production line, there will be a rework order from the customer product. [0003] After PQ analysis of the production line product and the repair product, t...

AI Technical Summary

Problems solved by technology

For different maintenance operations, the following three situations may occur: 1. The PCBA board to be repaired is the same, and the components to be replaced are different.

2. The repaired PCBA boards are different, but the replaced components are the same

3. The PCBA board that needs to be repaired is different from the components that need to be replaced

[0006] At present, in the research of man-hour quota in the field of welding, there is no research that combines fuzzy theory with BP neural network to improve the prediction ability of the prediction model

Images