Event-Triggered Distributed Consensus Filtering Based on Interval Analysis for Heterogeneous Multienergy Networks With Coupling Output Constraints

Abstract

This paper addresses the distributed consensus filtering problem of heterogeneous multi-energy networks (MENs), where subnetworks exhibit different dynamics and are coupled by output constraints. The global consensus is reached by propagating constraints between subnetworks in the form of interval vectors. Guaranteed intervals satisfying given constraints are returned, which are then used to correct the local states through the probability density function (PDF) truncation. Given the dependency and wrapping effects during constraint propagation, a hybrid Newton-forward backward propagation (FBP) contractor is proposed for avoiding overly pessimistic estimates. It extends the FBP contractor by replacing the natural inclusion function with centered forms. Additionally, an event-triggered mechanism is designed to enhance the efficiency of constraint propagation, which allows the threshold values to be adaptively adjusted according to interval widths. Finally, the proposal is assessed in a steel industrial gas-electricity network with 25 nodes, and the results verify the performance in estimation accuracy, especially for coupling nodes.