Decision Error Probability in a Two-stage Communication Network for Smart Grids with Imperfect Data Links

TL;DR

This paper evaluates the probability of decision errors in a two-stage smart grid communication network with imperfect links, analyzing how different decision rules and consumption patterns affect accuracy.

Contribution

It introduces an analytical framework for assessing decision error probabilities in a two-hop, 1-bit communication scheme for smart grids with binary symmetric channels.

Findings

Error probability depends on consumption frequency and decision rule used.

AND, OR, and MAJORITY rules have different performance under various conditions.

Numerical results demonstrate the impact of consumption patterns on decision accuracy.

Abstract

This paper analyzes a scenario where the distribution system operator needs to estimate whether the average power demand in a given period is above a predetermined threshold using an 1-bit memoryless scheme. Specifically, individual smart-meters periodically monitor the average power demand of their respective households to inform the system operator if it is above a predetermined level using only a 1-bit signal. The communication link between the meters and the operator occurs in two hops and is modeled as binary symmetric channels. The first hop connects individual smart meters to their corresponding aggregator, while the second connects different aggregators to the system operator. A decision about the power demand also happens in two stages based on the received information bit. We consider here three decision rules: AND, OR and MAJORITY. Our analytical results indicate the circumstances (i.e. how frequent the meters experience the consumption above the defined threshold) and the design setting (i.e. decision rules) that a low error probability can be attained. We illustrate our approach with numerical results from actual daily consumption from 12 households and 3 aggregators.