Decision error probability in a two-stage communication network for smart grids with imperfect sensing and data links

• Main Author: Ramezanipour, I. (Iran)
• Format: Dissertation
• Language: English
• Published: University of Oulu 2016
• Subjects: Master's Degree Programme in Wireless Communication Engineering
• Online Access: http://urn.fi/URN:NBN:fi:oulu-201602031115; http://nbn-resolving.de/urn:nbn:fi:oulu-201602031115

This thesis 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 a 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. In the first set of analysis, the decision making only happens by the network operator in the second hop and aggregators in the first hop only work as relay nodes which only forward the information it has received from the smart meters. AND and OR decision rules are studied in this scenario. Moreover, in the second set of analysis, the decision about the power demand happens in two stages based on the received information bit. Meaning that the decision making happens both by the aggregators in the first hop and network operator in the second hop. We consider here three decision rules in the second scenario: 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 both theoretical and numerical results from actual daily consumptions from 12 households and 3 aggregators. Also, we derive closed-form equations for the average decision error probability as a function of the system parameters (e.g. number of sensors, communication error, sensing error) and the input signal characterization. The first set of simulations are done in Matlab. Since the second set of data are provided in Excel; thus, the simulations are done using Visual Basic.