Inspiration

In the process of working with various Customers (tobacco factory, pharmaceutical factory, packaging materials production), we saw that they spend a lot of resources on improving the efficiency of their equipment (OEE) and understand why critical stops occur that directly affect the release Then we developed a solution based on the classic client-server architecture. Some time passed, cloud services became available for us and for customers. They have many advantages, for example, the use of machine learning. Laura. We want to show our decision on the example of a tobacco factory. We still have a lot of work, but we already have a beginning!

What it does

Employees using interactive dashboards can view various detailed information about the main production KPIs, sources of losses, time and reasons for equipment downtime. They see all the problem points providing an opportunity to understand and eliminate the true causes of production losses. It allows in real time to show the efficiency of the factory, shop or its sites, as well as compare the main production KPIs, such as:

  • OEE - the overall efficiency of equipment operation;
  • MTBF - average time between failures;
  • Stops and their causes All these indicators are very important for an effective industrial enterprise.

How I built it

The Microstop OEE downtime system receives data from cigarette machines directly from PLCs, HMI-panel and SCADA in real time. The data is collected on the IoT gateway and transferred to the Azure IoT Hub via the secure MQTT protocol. Since we do not have the opportunity to use a real installation, we wrote the Cigarette Machine (Maker) emulator, which was laid out on Github. We sent an invitation to testing@devpost.com to get acquainted with our project on Azure.com (should come from artur_******@bat.com)

Challenges I ran into

  1. A wide variety of automation systems in industrial plants. You need to know them well in order to read the required data.
  2. We quickly spent our loan in Azure IoT. This is too little to fully test our solution. Other cloud services, for example, Thingworx gives unlimited use 120 days.

Accomplishments that I'm proud of

We clearly understood that we can transform our classical solution into a cloud infrastructure. This will allow us to quickly introduce the platform to new customers.

What I learned

We spent only 2 weeks to make the first prototype. This is a good result. We realized that we can quickly learn the components of Azure and we can already do cool things.

What's next for Microstop OEE

We will develop our solution. Moreover, there is already an interest in potential customers. We also plan to implement the calculation of indicators:

  • Stops and their causes
  • TEEP - a coefficient of effective use of equipment;
  • MTTR - mean time to recover;
  • Current equipment performance;
  • Percentage of performance of the production task.

When we have enough information on breakdowns, then through the machine learning function we plan to create a model that will be able to predict breakdowns.

Built With

Share this project:
×

Updates