Contributions to Smart Metering Protocol Design and Data Analytics

Abstract

The next generation Advanced Metering Infrastructure (AMI), with the aid of two-way Smart Metering Network (SMN), is expected to support many advanced functions, such as remote reading and control, demand response, etc. In order to satisfy the communication requirements of these applications, the information transportation protocol design and data analytics are of fundamental importance to the design of SMN and associated research topics that need to be discussed. A suitable and well-designed Multiple Access Control (MAC) protocol is critical to the transportation of information in SMN. First, SMN traffic may be classified according to their communication requirements; thus each type may require a different MAC protocol which is specially designed for it. Furthermore, as discussed in Chapter 2, because the number of Smart Meters (communication nodes) involved in such a single SMN is much larger than those in today's local area networks, the traditional MAC protocols are unlikely to perform well in SMN. In order to solve this scalability issue for event-driven traffic, in Chapter 3, we propose two grouping based MAC protocols: TDMA-DCF and Group Leader DCF-TDMA. Both of them reduce the competitive random channel contentions in SMN effectively by dividing all the Smart Meters into several groups. The group division and management schemes are also clearly presented. We conduct comprehensive throughput and delay analysis on these two schemes for unsaturated traffic conditions and hidden node events. From the numerical results, we observe that the performance of these two grouping based MAC protocols are significantly better than those of traditional random access protocols. In order to solve the scalability issue for periodic data transportation, in Chapter 4, we propose a modified PCF scheme with the aid of cognitive radio technology, in which the Smart Meters are allowed to use the white space to report the periodic data to the data aggregator as secondary users. The comprehensive throughput analysis is also presented. The numerical and simulation results through NS-3 show that the modified PCF with cognitive radio significantly outperforms the traditional one in SMN. In Chapter 5, we focus on a specific type of periodic reporting data, power factor measurement, which is an important information to the power grid security and infrequently changing. Therefore, we exploit the time-invariant nature of this measurement and propose a new MAC protocol by using compressive sensing, a famous signal processing and reconstruction technique. The simulation results show that this proposed scheme can solve the scalability issue effectively and outperform the traditional MAC protocols. An effective analysis of Smart Metering data is critical to demand response, one of the most important applications in SMN, whereby customers use aggregated power consumption data (available locally via Smart Meters) and real-time pricing information (send pre-emptively from the utility) to schedule their future energy use to enhance conservation goals. This requires enabling customers to determine disaggregated consumption at individual appliance level from aggregated power consumption data. Therefore, in Chapter 6, we propose a real-time disaggregation algorithm based on a new Markov Chain model for the power consumption of individual electrical appliances. The model incorporates both temporal and inter-device correlations which are used to estimate the appliance-level disaggregated power consumption through the application of the Viterbi Algorithm. The performance is vetted through significant testing with real minute-resolution power consumption data and proves the accurate estimation.