Discrete-Event Based Monitoring and Diagnosis of Manufacturing Processes

We describe a decentralized control method that augments Boolean logic control to provide greatly enhanced diagnostics and monitoring of systems with discrete-valued sensors and actuators. Unlike conventional Boolean logic where logic is associated with the actuators, the new method defines a behavior and timing model for both sensors and actuators (called Control Elements). The behavior model is defined by a sequence of events called the event signature, in a Control Element´s (CE) neighborhood. The event signature encodes a subset of the process state into each CE. We describe a method for the continuous evaluation of this signature to compute a measure of the likelihood of a CE to change state. This likelihood measure called an Expectation Function, is used to check and enforce the correct behavior of a CE. Also, each CE learns a statistical temporal model in real time. The temporal model predicts delays in the states of the CE as a function of the previous delays and the current delays in the evaluation of its event signature. The distributed behavior model and the on-line estimated timing model are applied to a simulated bottle filling station to illustrate the detection and-diagnosis of incorrect behavior and failures of Control Elements.