A gas-treatment system consisting of three modules generates data for more than 300 different process variables, alarms and events every day. This raw data is written to a large number of text files.

Objective

The client required a software tool that allows to monitor the performance of the system. For each module, the following data must be visualized for a chosen date range:

  1. Process variables in individual graphs.
  2. Events marked by a timestamp.
  3. Alarms with a given start time and end time.

The text of the entire application must be displayed in German.

Challenges

  • The raw data contains a large number of variables spread out across many text files.
  • All the options of the graphs (color, name, alarms, events, …) must be configurable by the client with no coding experience.
  • The solution must be installed on a server with no database access and no internet connection.

Solution

The delivered solution consists of three components.

1. Automated R-scripts to process the raw data

First, a program was written to clean and organize the raw data from the text files. Since no database nor internet are available, the data is written to structured CSV-files labeled by date. This program is then scheduled to automatically run every morning.

2. Interactive dashboard to visualize the processed data

Next, an interactive dashboard was developed to visualize the processed data. It offers the following features:

  • Selection of one of the three gas-treatment modules.
  • Selection of a date range over which to plot the data.
  • Selection of the plotted curves by name.
  • A legend showing numeric values on mouse-over.
  • Interactive graphs that allow to zoom in on an arbitrarily small interval.
  • Display of alarms and events using colored shading and vertical lines with text.

To make searching for particular alarms and events easy, the app contains a second tab with a large event table. This table allows to:

  • Search alarms and events using wildcard or pattern matching.
  • Filter alarms and events by a specified date range.
  • Order the rows by any column.
  • Copy the contents current table to the clipboard or export it as a CSV or Excel file.

3. An Excel file to configure the dashboard

This Excel file allows a user with no programming experience fully customize the dashboard:

  • Which variables are plotted on which graph and which axis.
  • The color of the curves and alarm shading.
  • The variable display name and axis labels.
  • Which alarms and events are added to a particular graph.

Technology

  • The data processing program is written in the R programming language.
  • The program was scheduled to run daily using a Windows batch file and vbscript.
  • The dashboard is developed with R and Shiny.

Want to know more?

Contact me to discuss a custom solution for your specific problem.