Scholarlink

<p>An exploratory statistical analysis was conducted to assess the effectiveness of a scaled manufacturing cell. The cell was equipped with presence sensors, robotic arms with five degrees of freedom, and a Raspberry Pi control system with local storage. Three iterations were performed, involving the transfer of measurements through seven workstations within the manufacturing cell. From the iterations, the control system collected data on the measurement’s residence time and sent it to the Software as a Service called MAQUIA4.0®. Then, the Shapiro–Wilk and Jarque–Bera proved a better option for parametric data analysis, since they permit a robust analysis of normal distributions, while for non-parametric data, the KED is useful. The results showed outliers due sensors issues at station 3, causing a process effectiveness about 87.5% (12.5% of significance). Thus, sensor was replaced returning the manufacturing cell returned at its 100% effectiveness. While the second and third iterations produced typical (expected) data when analyzing the times of each sample. Based on the results, it is concluded that the histogram of station number 5 does not show statistical evidence to reject the hypothesis of normality. Finally, as future work can be the application of functional data analysis to track the process through observations of quality features in the form of one or more profiles (in the electronic industry). The above, by adding good agreements within industrial and normalization will have a significant impact into the 4RI manufacturing systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Exploratory statistical analysis for decision-making in a scaled manufacturing system of medical equipment

  • 期刊论文

  • 发表日期：

    2025年4月21日

查看全文
  • Darío Bringas-Posadas,
  • Antonio Trejo-Morales,
  • Daniel Gómez-Torres,
  • Edgar A. Franco-Urquiza,
  • Celso E. Cruz-González

摘要

An exploratory statistical analysis was conducted to assess the effectiveness of a scaled manufacturing cell. The cell was equipped with presence sensors, robotic arms with five degrees of freedom, and a Raspberry Pi control system with local storage. Three iterations were performed, involving the transfer of measurements through seven workstations within the manufacturing cell. From the iterations, the control system collected data on the measurement’s residence time and sent it to the Software as a Service called MAQUIA4.0®. Then, the Shapiro–Wilk and Jarque–Bera proved a better option for parametric data analysis, since they permit a robust analysis of normal distributions, while for non-parametric data, the KED is useful. The results showed outliers due sensors issues at station 3, causing a process effectiveness about 87.5% (12.5% of significance). Thus, sensor was replaced returning the manufacturing cell returned at its 100% effectiveness. While the second and third iterations produced typical (expected) data when analyzing the times of each sample. Based on the results, it is concluded that the histogram of station number 5 does not show statistical evidence to reject the hypothesis of normality. Finally, as future work can be the application of functional data analysis to track the process through observations of quality features in the form of one or more profiles (in the electronic industry). The above, by adding good agreements within industrial and normalization will have a significant impact into the 4RI manufacturing systems.