(202d) Living on the Edge with Distributed Embedded Model Predictive Control

The Industrial Internet of Things (IIoT) offers new opportunities for operators to improve efficiencies, the development of open, secure, and interoperable systems architectures is key to delivering that potential. Cooperative distributed model predictive control is designed to manage dispersed systems, and can stabilize and optimize an industrial process having many manipulated and controlled variables. Rather than using a single, centralized solution, multiple control agents are used in a distributed design. These control agents operate in a cooperative manner to achieve a shared objective function which approaches the global optimum even when significant interactions are present. The optimization also encompasses plant capacitance, recycle streams and mixed fast and slow process dynamics. This technology has the potential to harness the power of IIoT connectivity to deliver robust plant wide control with significant operational benefits. Additionally, it has the advantage of a simple, modular design and utilizes existing control standards and infrastructure.