This paper shows improved reliability in Industrial Internet of Things (IIoT) distributed automation systems by following along at runtime against expected real-world fault models. This novel idea represents models using extended Petri nets, called control interpreted Petri nets (CIPN), which are basically executable state models. The extensions are synchronization hooks to the various factory IoT devices, events, and controls. An Industrial 4.0 enterprise shows how the traditional top-down automation pyramid is replaced with control systems where functionality is now distributed across IIoT device controllers. Those previous pyramid layers, where control was deterministic and reliable, scale but give up local distributed control (a key IIoT characteristic).
This paper extends the distributed system concepts with new stochastic reward nets (SRNs)--a variant of Petri nets supporting stochastic timing features. SRNs “bridge the gap between the expressiveness of the widely adopted control models and the need for verification of automatically synthesized distributed control systems.” Edge-based reliability and performance runtime monitoring are also required. Collectively, this is called network aware modeling for distributed sequential discrete event control.
These new offline analysis steps, coupled with model property verification, help manage the complexities inherent in highly distributed IIoT systems. Real-world industrial case studies are introduced to show how the fault models are generated. This includes a typical automation controller with controllers, gripper, and movement (like a robot), which is then shown with CIPN (extended Petri nets) where the parts of the robot are managed. This also shows the relatively low level of expressiveness needed to control devices (basically a descriptive control language).