Is grid modernisation success hidden in data analytics?
Big data analytics continues to be a top deployment priority for businesses as its use will help them remain competitive while accelerating innovation. Globally, the big data analytics market is estimated to grow 4.5 times, garnering revenue of $68.09 billion by 2025.
Having acknowledged that big data analytics forms an integral part of smooth business operations, it is complex and yet an evolving exercise. This intricacy was highlighted in part one of an article on big data analytics, published in ESI Africa Issue 3 2020. In this second instalment, we now delve deeper into other aspects such as data architecture, technologies, and applications.
Big data system architecture for the power grid comprises reliable, scalable, and automated data pipelines across grid systems. This system relies on communications technologies that collect raw data and convert those data into information that provides insight and value. The technologies involved are: 1) data acquisition, storage, and querying applications used by electric utilities; and 2) data-analysis models and methodologies.
A logical architecture for big data and analytics has three components:
1. Information management: high-volume data acquisition, multi-structured data organisation and discovery, and low-latency data processing.
2. Real-time analytics: speed of thought analysis, interactive dashboards, advanced analytics, and event processing.
3. Intelligent processing: application embedded analysis, optimised rules and recommendations, guided user navigation, and performance & strategy management.
Utilities have successfully used the above architecture for acquisition, storage, and analysis of smart meter and advanced metering infrastructure (AMI) data on customer energy usage. Meter data management (MDM) can collect, store, and process customer data acquired from smart meters as well as from their predecessors, interval meters.
In a general energy MDM system architecture, at one end are meters that sense and collect energy-related time-series data. The AMI network transmits these data using communication technologies such as wired and wireless networks, and the utility head-end system collects and aggregates the data. Once the data are collected securely and their accuracy are validated, the utility enterprise system’s interfaces link the data to different utility applications. The utility MDM system can leverage the big data architecture outlined above to use customer smart-meter data for different purposes, such as managing loads through demand response (DR), providing customer service and billing, managing grid outages, and enabling customer participation in electricity markets through rate tariffs.
Using the above data acquisition and management techniques and the architecture to support advanced applications of big-data technologies creates challenges. Disparate utility head-end systems (e.g. MDM and DR management systems) need to integrate and exchange data among themselves. They also need to interface with diverse distributed energy resources and enable markets that participate in the safe, reliable operation of the electricity grid.
Data acquisition technology
The two most basic forms of data acquisition technologies on the smart grid are: 1) automatic meter reading (AMR) or AMI, and 2) supervisory control and data acquisition (SCADA) or distribution automation (DA). AMI is an integrated system of smart meters, communications networks, and data management systems that enables two-way communication between utilities and customers through a smart meter. The AMI provides a number of functions that remotely measure electricity use, connect and disconnect service, detect tampering, identify and isolate outages, and monitor voltage.
Customers are provided access to usage data for informational purposes. SCADA/DA systems support efficient and reliable power systems within the utility’s network. When integrated with MDM systems, SCADA/DA systems can monitor electricity transmission and distribution system equipment over large areas, allowing utilities to quantify power-quality issues related to voltage/ current and control assets within their networks. These systems employ automated decision making, effective fault detection, and power restoration to support reliable power supply to customers.
Data communication technology
Utilities transmit data using communication technologies with multiple protocols, frequency bands, and transfer rates depending on the purpose, location, cost, security and privacy requirements of the data or technology. Smart-grid communication technologies fall into two primary categories: wired and wireless. A recent study compared wired and wireless communication technologies and evaluated their applications on the smart grid.
Wired communication technologies
Most utility service providers prefer wired communication, transmitting energy data over power lines. The most important advantages of wired communication are reliability and insensitivity to interference.
Types of wired communications include:
1. Power-line communications, which send data over existing power cables. There are two classes: broadband and low- and high-data-rate narrowband. Broadband operates in the 1.8-250 megahertz (MHz) range and has a physical layer rate ranging from several megabits per second (Mbps) to several hundred Mbps. Low- and high-data-rate narrowband operate in the 3 kilohertz (kHz) to ~500kHz range and have physical layer rates of 1-10 kilobits per second (Kbps) for low data rate and 10-500Kbps for high data rate.
2. Fiber-optic communication is a fundamental communication technology for a wide-area network because it has a relatively high data rate and is immune to noise. High data rates range from 155Mbps to 40 gigabits per second (Gbps).
3. Digital subscriber line (DSL) is used to transmit digital data over telephone lines. There are three DSL systems: asymmetric (ADSL), highspeed (HDSL), and very-high-data rate (VDSL). ADSL has data rates of up to 8Mbps downstream and 800Kbps upstream; HDSL has data rates of up to 2Mbps, and VDSL has data rates of up to 100Mbps.
4. Coaxial cable communications on the cable infrastructure, which can provide data rates of up to 170Mbps.
Wireless communication technologies
1. ZigBee is a wireless personal area network protocol that provides data rates ranging from 20Kbps to 250Kbps.
2. A wireless local area network (WLAN) is based on Institute of Electrical and Electronics Engineers (IEEE) standard 802.11. A WLAN provides data ranging from 2Mbps to 600Mbps.
3. A wireless mesh network has many nodes of mesh clients and routers.
4. Z-Wave is a proprietary wireless technology that is suitable for short-range communications and supports data rates of up to 40kbps.
5. WiMAX is a 4G wireless technology based on the IEEE 802.16 set of standards. It provides data rates of up to 75Mbps.
6. The cellular network is a communication network in which the last link is wireless. The data rates depend on which generation of the network is used: 2G, 2.5G, 3G, and 4G provide data rates of 14.4Kbps, 144Kbps, 2Mbps, and 14Mbps, respectively.
7. Satellite communication transfers signals between two nodes and has data rates of up to 1Mbps.
Big data analytics applications
Data’s value lies in the information or insight that analytics can derive from the data. Descriptive, and predictive analytics are commonly used for this purpose.
Descriptive data analytics models – aggregate and mine data to provide insight into the past. Cluster analysis is a commonly used, unsupervised learning technique that can help identify different types of energy consumption behaviour. It has been applied to individual industrial, commercial, and residential customers and is usually employed in descriptive models. A good clustering method produces high-quality clusters with low inter-cluster similarity and high intra-cluster similarity; in other words, members of a cluster are more like each other than they are like members of a different cluster.
Predictive data analytics models – A variety of statistical, modelling, data mining, and machine-learning techniques are utilised to study recent and historical data to make predictions about the future. Predictive analysis employs three models: load-shape regression model, change-point regression model, seasonality & trend decomposition. Due to the transition from centralised systems to local distributed energy resources, there is a need to deploy advanced sensors and measurement equipment and use such data analytical methods for a safe, reliable, and resilient planning and operation of the grid.
Final word: electricity grid stakeholders – including utility operators, electricity customers, and product vendors – can leverage the findings of this study to identify opportunities and technologies for big data and analytics-related applications for demand-side management and in support of modernising the electric grid. ESI
This article is based on an adaption of a 2019 paper titled Big-Data Analytics for Electric Grid and Demand-Side Management, written by Rongxin Yin, Girish Ghatikar, and Mary Ann Piette – of the Electric Power Research Institute. View a full list of references and diagrams. www.researchgate.net/publication/336029983