Analytics Product for Predictive Maintenance & Data Visualization in the Semi-Conductor Industry

The client set up an IoT infrastructure that collects all critical data points along the process chain. The recipes and conditions of the process chambers are recorded and stored in real-time by sensors installed in the processing module. The client commissioned Indium to develop the analytics layer of its IoT infrastructure, which would: Measure the efficiency of wafer production, Monitor wafer production lines, Identify outlier process modules, Predict defects in process modules, and Enable predictive maintenance. Although the semiconductor industry is highly complex and domain-intensive, Indium Software was able to deliver all the requirements and a few additional value-added features by leveraging its Big Data, Predictive Analytics, and Visualization capabilities.

The client is one of the world’s most advanced semiconductor manufacturing enablers; today, nearly every advanced chip is built using its technology. Its manufacturing centers across the Americas, Europe, and Asia develop and supply wafer fabrication equipment and services to every semiconductor manufacturer worldwide. The client is a California-based value-added services provider for semiconductor manufacturers. They develop and supply wafer fabrication equipment and services to build innovative devices. Creating chips involves a chain of individual steps, with each process module producing multiple wafers in defined recipes of temperature, pressure, and other conditions.

The application is a data bank of the all- critical data points gathered in the process chain. The recipes or the conditions of the process chambers are recorded and stored in real-time by sensors installed in the process module.

The data produced was continuous, time series in nature and significantly huge in size. This led to an unusual IoT problem. However, the Data was an asset to business which when put to insightful use, facilitates real-time monitoring and defect prediction in the process module.

The client wanted a solution to make use of the available data to achieve the following goals:

• Measure efficiency of the wafer productions
• Monitor production line of wafers
• Identify outliers process modules
• Prediction of defects in process modules
• Predictive maintenance