Cloud4NFICA-Nearness Factor-Based Incremental Clustering Algorithm Using Microsoft Azure for the Analysis of Intelligent Meter Data

Cloud4NFICA-Nearness Factor-Based Incremental Clustering Algorithm Using Microsoft Azure for the Analysis of Intelligent Meter Data

Archana Yashodip Chaudhari, Preeti Mulay
Copyright: © 2020 |Pages: 19
DOI: 10.4018/IJIRR.2020040102
OnDemand:
(Individual Articles)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Intelligent electricity meters (IEMs) form a key infrastructure necessary for the growth of smart grids. IEMs generate a considerable amount of electricity data incrementally. However, on an influx of new data, traditional clustering task re-cluster all of the data from scratch. The incremental clustering method is an essential way to solve the problem of clustering with dynamic data. Given the volume of IEM data and the number of data types involved, an incremental clustering method is highly complex. Microsoft Azure provide the processing power necessary to handle incremental clustering analytics. The proposed Cloud4NFICA is a scalable platform of a nearness factor-based incremental clustering algorithm. This research uses the real dataset of Irish households collected by IEMs and related socioeconomic data. Cloud4NFICA is incremental in nature, hence accommodates the influx of new data. Cloud4NFICA was designed as an infrastructure as a service. It is visible from the study that the developed system performs well on the scalability aspect.
Article Preview
Top

1. Introduction

Artificial Intelligence (AI) related inventions are booming, shifting from theory to commercial application. AI is an amalgamation of Machine Learning (ML), Deep Learning (DL), etc. Some areas of AI is developing more quickly as compared to others. ML is the leading AI technique disclosed in power system. ML-related techniques acquire more than one-third of all identified inventions. ML techniques categorized into Supervised, Unsupervised, and Reinforcement learning. Clustering is the subcategory of unsupervised learning — however, traditional clustering lack in the concept of new learning. Incremental learning via incremental clustering algorithm is the novel concept in ML.

According to the World Resources Institute (WRI), (Chakrabarty, August 08, 2018), energy-related emissions make up more than two-thirds of India's overall emissions and represent more than three times the next largest source (the industry sector). Of the country's energy sector emissions, 77 percent are from electricity generation, making this a key target for reductions to meet India’s climate, or Nationally Determined Contribution (NDC), and commitments(India, 2018). Although India’s emissions are still comparatively low on a per-capita basis, the very size of the nation's population and the scope for them to increase the country's emissions of global concern; India's emissions from the energy sector are the fourth largest in the world (Navroz K. Dubash, August 22, 2018) (behind China, the United States, and the European Union), and rising.

One of the challenges faced by India’s electricity sector is the capacity of the grid to accommodate renewable energy as the country invests in renewable energy to meet national targets. It will have a potential surplus, rather than a shortfall, in electricity, and yet it faces technical constraints in using that electricity. According to analyst firm Brookings, (Tongia, 2018) to avoid curtailing renewable energy generation, the country needs “a stronger grid, cheap storage, and the ability to shift load to match supply conditions.”

In parallel with these challenges, in India today, utility companies still rely on manual electricity meters reading to track consumption at residential and industrial locations. The utility companies send employees to read these meters each month and then generate monthly electricity bills. This system does not provide the utilities with insight into electricity consumption patterns, for example, based on time-of-day usage, nor does it enable consumers to understand their electricity consumption as part of the larger picture of electricity availability.

Complete Article List

Search this Journal:
Reset
Volume 14: 1 Issue (2024)
Volume 13: 1 Issue (2023)
Volume 12: 4 Issues (2022): 3 Released, 1 Forthcoming
Volume 11: 4 Issues (2021)
Volume 10: 4 Issues (2020)
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing