The Application of Artificial Intelligence Technology in Human Centered Manufacturing in Industry 5.0

Main Article Content

Jiawei Zhang

Abstract

In order to clarify the cognitive process of human beings and the influencing factors of human errors in the process of manufacturing capability evaluation, quantitatively analyze the reliability of human beings in the manufacturing process, and more accurately evaluate the manufacturing capability of production lines, the author proposes the application of artificial intelligence technology in human centered manufacturing in Industry 5.0. In response to the dynamic nature of data in the operation process of manufacturing production units and the varying importance of indicators to evaluation objects at different times, the author proposes an objective weighting method that combines indicator sensitivity with entropy weight method to solve the problem of existing weighting methods only considering the fluctuation of indicator data and ignoring the importance of evaluation indicators to all evaluated objects. The combination of subjective weights established by the Analytic Hierarchy Process (AHP) is used to obtain the final combination weight of indicators. At the same time, evaluate and analyze the factors that contribute to human error behavior to obtain the human reliability of the unit, and introduce it into the comprehensive evaluation of unit manufacturing capability. Based on the time series data in the evaluation, a time dimension factor combined with grey correlation analysis is introduced to conduct a dynamic comprehensive evaluation of unit manufacturing capacity in time series, and the production unit manufacturing capacity index is obtained. The example results show that 10 indicator data from the past 10 time periods were selected for evaluation, and the closer the time period, the more important the data is. The time factor for each time period is (0.0048, 0.0068, 0.0126, 0.0266, 0.0582, 0.0704, 0.1232, 0.1685, 0.2212, 0.3070). The unit capability value obtained through dynamic horizontal and vertical comprehensive evaluation is most consistent with the capability value obtained by the author’s method. Under the four methods, although there are differences in the capacity values of each unit, the fluctuation is within a reasonable range, indicating that the author’s evaluation method is reasonable and feasible. The feasibility and effectiveness of the evaluation method have been validated.

Article Details

Section
Special Issue - High-performance Computing Algorithms for Material Sciences