Multi-modal manufacturing data preprocessing method in intelligent factory

Abstract

The invention relates to a multi-modal manufacturing data preprocessing method in an intelligent factory, which comprises the following steps: S1, receiving manufacturing data acquired by an industrial data acquisition system, and forming a training set D by the manufacturing data; S2, clustering the training set D to obtain a clustering result, constructing a neural network model, and training the neural network model according to the training set D and the clustering result thereof to obtain a final multi-modal manufacturing data integration model; S3, classifying the manufacturing data set needing to be processed through the multi-modal manufacturing data integration model; S4, performing denoising processing on the manufacturing data set; S5, performing redundancy elimination processing on the manufacturing data set; and S6, identifying missing values in the manufacturing data set, and filling the missing values. The multi-modal manufacturing data are clustered according to the model cluster by constructing the multi-modal manufacturing data integration model, so that data processing of subsequent links is facilitated. A data multi-stage optimization cleaning strategy is used, so that the data quality is improved, and the data correctness, consistency, integrity and reliability are ensured.

Description

technical field [0001] The invention relates to the technical field of intelligent manufacturing and data management, in particular to a method for preprocessing multi-modal manufacturing data in an intelligent factory. Background technique [0002] In the era of industrial big data in smart factories, massive multi-modal manufacturing data exists widely. How to dig out the hidden huge value from these manufacturing data is the main concern of industrial big data research at this stage. However, the low-quality modalities, real-time processing, unbalanced modalities, and high-dimensionality of attributes of multimodal manufacturing data pose severe challenges to the design and implementation of data mining. In the data mining process of smart factories, data preprocessing takes about 60% of the entire process. Therefore, how to complete multi-modal manufacturing data preprocessing more efficiently and ensure data correctness, consistency, integrity and reliability is the ke...

AI Technical Summary

Problems solved by technology

On the other hand, in the process of multi-modal manufacturing data preprocessing, the intervention of experts in related fields is mostly required. When abnormal situations occur in the processing process, manual auxiliary decision-making is usually required, which is inefficient and cannot meet the needs of smart factories. Actual demand

Images