A Framework for Integrating Diverse Data Types for Live Streaming in Industrial Automation

In the modern manufacturing system, the industrial automation plays an important role in improving the efficiency, productivity, and overall performance of the system by increasing throughput, reducing downtime and more. However, challenges arise in implementation of real-time continuous data streaming and the integration of diverse data types, posing significant track events. To overcome the challenges, the present paper proposes a methodological approach that employs linear mixed models for data integration and online stochastic gradient descent with the Adaptive Gradient algorithm for processing streaming data in real-time. The use of linear mixed models proves advantageous, allowing for the effective integration of data from diverse sources. The methodology captures complex relationships within the data, offering a comprehensive understanding of the industrial process. A case study of predictive maintenance systems for industrial machinery is demonstrated for empirical validation for the online stochastic gradient descent with Adaptive Gradient facilitates real-time processing of streaming data. This adaptive algorithm ensures continuous adjustment to evolving data patterns, and providing a holistic view of industrial processes. Through empirical results, the demonstration of the system’s efficiency in increasing throughput and in-built system visualization is achieved. This paper contributes to the advancement of industrial automation by presenting an innovative solution for handling disparate data types in a unified framework. The findings cover the way for improved operational efficiency and reliability in complex industrial environments.