Power BI OEE monitoring system

Iot Case study – OEE monitoring system

This case study displays the implementation of an IoT based OEE system using various Microsoft Azure services. The solution aims to address the challenges of integration, digitalization, and automation in business processes, specifically focusing on manufacturing.

Key services used in this solution include Azure Data Factory for integrating various data sources into a central repository, Azure IoT Hub for connecting and managing IoT devices, Microsoft PowerApps for creating a user-friendly human-machine interface, Stream Analytics for real-time data visualization, Power BI for historical data analysis and reporting, and Power Automate and Azure Logic Apps for creating an alert and notification system.

The solution provides a scalable and cost-effective approach, allowing businesses to pay only for the resources they consume. Additionally, it highlights the flexibility and democratization of the Azure platform, empowering users to maintain and further develop the system with specialized training in specific tools and services.

IoT case study – OEE monitoring system

Assumptions of IoT OEE system implementation

In an ideal scenario, a system that monitors manufacturing efficiency KPIs (including OEE) should have the following characteristics:

  • it should be integrated with an ERP system to retrieve information on production orders and real-time product data;
  • it should be technology-agnostic, independent of manufacturing technologies, machine suppliers, and data exchange protocols;
  • it should provide a user-friendly human-machine interface, allowing operators to input information into the system and retrieve basic data on current and historical events;
  • it should enable real-time streaming of data at the product or downtime level, with the ability to access this information anywhere, anytime;
  • it should gather historical data for further analysis, such as trend analysis and downtime reasons, and present them in a user-friendly format;
  • it should support advanced analytics on accumulated data using machine learning and artificial intelligence scripts;
  • it should incorporate an alert system that sends notifications to operators, planners, and warehouse personnel;
  • it should be flexible enough to adapt to any type of business;
  • it should be part of a larger framework where all factory data flows into a central repository, powering a global reporting system that facilitates internal information and knowledge exchange among different factories;
  • it should not tie the company to a single vendor and should be capable of further development by alternative providers or appropriately trained factory personnel.

The above list may seem idealistic, but in a “composable approach,” each of these points is realized through autonomous, dedicated modules, services, or pieces of IoT class hardware.

Data architecture

The Diagram below presents a “composable approach” solution based on Azure services and IoT terminals with sensors.

  1. Integration of various data into a single source of information – almost every manufacturing company uses an ERP system. Azure Data Factory allows you to extract data from any manufacturing management system, process it and load it into a central repository (such as Azure SQL Database). This database becomes the primary “source of truth” for the system being developed. As a result, the solution can combine multiple types of data from various sources, enabling seamless communication among them through the repository.
  2. Adaptation to different types of machines and technologies is done through Azure IoT Hub, which takes data from independent IoT-class sensors placed directly on the machines (without interfering with their operation). These sensors collect production data, downtime information, and machine status. The data is pre-processed and sent to the central repository mentioned earlier. In addition, the IoT Hub allows remote management of sensors.
  3. The human-machine interface is built using Microsoft PowerApps. The application’s design and functionality can be tailored to the specific needs of the production process, and it can be installed on any mobile device or accessed through a standard web browser. As a result, operators have the ability to communicate with the system and indirectly with the management team. They can also make more autonomous decisions based on a customized set of information, such as equipment failure or the progress of a production order.
  4. Real-time data visualization is crucial for effective production management. This functionality is provided by Stream Analytics. It’s worth noting that data can be sent to analysis services (such as Power BI) and other Azure platform services.
  5. Historical analysis is readily available because all IoT sensor data is stored in a central repository (SQL database), where it is combined with data from other sources (ERP, MES, operator application). These combined datasets are used to create Power BI data models tailored to specific reporting needs, such as downtime and failure reports, unused capacity reports, or production waste volumes and costs.
  6. Advanced analytics using artificial intelligence or machine learning scripts are made possible with services such as Azure ML Studio. In practice, this may involve training a model for early detection of patterns preceding machine failure based on vibration and temperature data from IoT sensors. This aligns with the concept of preventive maintenance.
  7. The alert and notification system is created based on logic and conditions tailored to the characteristics of a specific production environment. It is managed by Power Automate and Azure Logic Apps. As a result, designated personnel are informed through automated email messages and in-app notifications.
  8. The structure created is easily scalable, both within a single factory and the entire organization. Adding new users or increasing the capacity of the cloud database is a matter of a few clicks in Azure Portal. Adding another factory according to a predefined standard is also relatively quick. It requires the purchase and configuration of additional licenses within existing services and standard data exchange connectors. As a result, historical and real-time data can be presented for all factories in the organization, creating a digital twin of the manufacturing part of the supply chain.
  9. The Azure platform, as well as its individual services, are incredibly “democratic” technologies. Each module is well-documented and, to some extent, independent. This allows for the system’s maintenance and development to be taken over by personnel trained in specific tools after the implementation. This greatly reduces risks and ensures cost-effectiveness resulting from high competition in the Microsoft service market.
  10. All of the above-mentioned services receive data generated through Antdata’s IoT-class terminals, which ensure smooth communication between IoT sensors and the Azure cloud via the Edge device (server).

The solution presented in the case study is an example of one among many possible compositional systems. What’s more, such a system can be expanded with additional modules, such as Azure Custom Vision, which can be used to identify quality defects in products using images captured live from the camera.

Get in touch with us

Book a consultation call and we will get back to you.

Scroll to Top