Minor Stoppages Analysis on a Shiny App

Marcelo Tardelli
February 23rd, 2016

General Concepts

The Minor Stoppages Analysis app was built up as the final project of Coursera's Developing Data Products course.

  • It's main goal is to be used as a tool to analyse Minor Stoppages of automated packaging equipment, using the data that can be obtained directly from their PLC (or “machine computer”).

  • It adjusts real data to the theoretical statistical distributions and their related parameters as input for further process simulations on productivity and efficiency improvement of those production lines through specific software.

  • It can be found at: https://mjtardelli.shinyapps.io/msanalysis/

Definitions

A Minor Stoppage is an operational stoppage of a machine / line that :

  • is shorter than 10 minutes (600 seconds) - this value may change from company to company;
  • to restore the machine's proper operation, there is no need for a maintenance technician intervention and no spare parts are exchanged;
  • have direct impact on machine's OEE (Overall Equipment Effectiveness) and reduces significantly the line throughput.

Understanding the behavior of stoppages drive more effective actions to reduce or eliminate them, allowing simulations that demand good knowledge on the statistical distribution parameters of such events.

Setting Up the App for the Analysis

Instructions

  • In the main panel of the app you just have to define the dates interval, packaging line and shift to be studied;

  • Graphs and parameters are automatically updated.

Analysis - Graphs and Results

In the “Time Between Minor Stoppages” tab :

  • A time series graph of the selected data and a histogram with the Mean Time Between Stoppages plotted;
  • A graph with the real and the theoretical Weibull plots;
  • The Weibull distribution parameters and regression data tables and plots.

In the “Minor Stoppages Duration” tab :

  • A time series graph of the selected data and a histogram with the Mean Stoppage Time plotted;
  • A graph with the real and the theoretical Lognormal plots;
  • The Lognormal distribution parameters and regression data tables and plots.