Understanding the Potential of Smart Grid Data Analytics

nosejasonΗλεκτρονική - Συσκευές

21 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

162 εμφανίσεις

Understanding the Potential
of Smart Grid Data Analytics
A GTM Research Whitepaper
JANUARY 2012
Understanding the Potential of Smart Grid Data Analytics
2
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
TABLE OF CONTENTS
1 OVERVIEW: THE AMI DATA AND ANALYTICS OPPORTUNITY 3
2 PLANNING FOR DATA QUALITY 4
3 AMI DATA TYPES 5
3.1 Measurement Data Versus Events and Alerts 5
3.2 Power Consumption Data 6
3.3 Additional Data Types and Functions 6
3.4 Analytical Methods 6
3.4.1 Aggregations 7
3.4.2 Correlations 7
3.4.3 Trending 8
3.4.4 Exception Analysis 8
3.4.5 Forecasts 8
4 DATA AND ANALYTICS APPLICATIONS 9
4.1 Revenue Management 9
4.1.1 Load Forecasting 9
4.1.2 Theft Detection 9
4.1.3 Prepay 9
4.1.4 Rate Plan Modeling 10
4.1.5 Demand Management 11
4.2 Consumer Engagement 11
4.2.1 Conservation Tips and Suggestions 11
4.2.2 Rate Plan Selection 12
4.2.3 Efficiency Program Measurement and Planning 12
4.3 Distribution Optimization 12
4.3.1 Outage Management 13
4.3.2 Distribution Network Planning 13
4.4 AMI Network Management 14
4.4.1 Service Level Management 14
4.4.2 Network and Device Configuration and Troubleshooting 14
4.5 The Future of Smart Grid Data Analytics 15
Understanding the Potential of Smart Grid Data Analytics
3
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
1
OVERVIEW: THE AMI DATA AND ANALYTICS OPPORTUNITY
Smart meters present unprecedented opportunities to push the boundaries of grid visibility
beyond substations and transformers and into the home. The potential benefits of advanced
metering infrastructure (AMI) go well beyond automating the meter-to-cash business
process. It is now possible to view and analyze consumption data in new ways for a plethora
of business applications, including capacity planning, demand management, rate design
and reducing peak power consumption. Further, meters can also capture new metrics and
receive and execute remote commands. Examples include periodic voltage readings to
support voltage optimization; remote connect/disconnect for service provisioning; outage
alerts and power restoration notifications for automated outage management; on-demand
register reads by customer service representatives helping customers; and more.
Organizations seeking to attain the maximum benefits from AMI need a data analytics
strategy. The starting point is to understand the range of business applications AMI data
can enable or enhance. Once these opportunities are identified, planners can determine
which ones make the most business sense to pursue, the supporting data requirements,
and the analytical processes needed to turn raw data into actionable information for
improved decision-making.
This whitepaper explores the AMI data analytics opportunity by highlighting the immediate
opportunities smart meter data creates, as well as some of the data requirements and
analytical methods needed to capture them. It is intended as a useful planning guide for
executives, senior managers, data architects, statisticians and other data management
professionals tasked with planning and executing strategic smart grid initiatives.
Establishing a strategic data and analytics vision is paramount; however, the history of data
and analytics is littered with grand plans that have been bogged down under the weight
of their own complexity. Wherever possible, we provide tips for getting smart grid data
analytics up and running quickly and with a limited number of dependencies.
Understanding the Potential of Smart Grid Data Analytics
4
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
2
PLANNING FOR DATA QUALITY
Generally speaking, high-quality data is accurate, timely and relevant. Specific requirements
for collection frequency, latency (the lag time between data measurement and data use),
and accuracy will vary depending on the data analytics application.
For instance, accurate billing determinants need to be available within billing cycle
timeframes, needs that can be satisfied by periodic data collection. Further, full-featured
meter data management systems are able to deal with incomplete or inaccurate data via a
process known as validation, estimating and editing (VEE). While the primary purpose of
VEE is to ensure data accuracy to generate billing determinants, VEE can also help ensure
that consumption data used for analytical purposes is clean and accurate.
Voltage readings used for technical distribution optimization, on the other hand, present
different data quality requirements. Here, timeliness takes higher priority, and instead of
periodic polling, on-demand voltage readings can quickly be gathered and loaded into
an optimization engine.
The key point is that data quality is relative and determined by the way the data will be
used, as opposed to an overarching abstract principle; both of these considerations are
discussed further in this whitepaper.
Understanding the Potential of Smart Grid Data Analytics
5
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
3
AMI DATA TYPES
Conventional electricity meters are simple devices that measure power consumption using
a running register. Smart meters vastly expand the available range of data and functionality.
This is exciting, but at the same time, it can become confusing to wade through all the
possible permutations of data types, potential end uses and supporting data quality
requirements. Two useful criteria to keep in mind are: 1) differences between measurement
data and events/alerts and 2) power consumption data vs. other types of data.
3.1
Measurement Data Versus Events and Alerts
Both time-interval consumption data and register readings are types of measurement data.
In terms of format, timing and structure, measurement data is predictable. Data collection
can be scheduled and fulfilled via periodic polling at preset intervals, with batches of data
forwarded to a meter data repository for VEE.
Events and alerts, on the other hand, are typically unscheduled messages that happen
randomly when an unusual situation is detected, such as a meter break-in as part of a theft
attempt or an interruption in power delivery (i.e., an outage). Important alerts should be
routed directly to the person and/or applications that need to know about them. A best
practice is to use an enterprise service bus (ESB) messaging middleware with publish and
subscribe capabilities for multipoint broadcasts of critical messages, and to use complex
event processor (CEP) technology to quickly inspect messages, apply business rules and
determine message routing.
Although data management practices are, in general, quite different for measurement
data versus events and alerts, a common misconception is that analytics are limited to
measurement data. To the contrary, event data can be very useful for analytics. A good
example of this is the measurement of outages and reporting to regulators on quality of
service using metrics such as SAIFI (System Average Interruption Frequency Index, or
how often the system-wide average customer experienced a power interruption in the
reporting year) or MAIFI (Momentary Average Interruption Frequency, that is, the number
of momentary outages per customer per year).
Measurement data can also be associated with an event and logged for analysis at a later
point. For instance, momentary voltage sags (i.e., power flickers) associated with meters
along a particular feeder can be logged and compared with transformer data to target
costly vegetation management. Monetary savings from targeted maintenance dispatch can
add up quickly, yielding immediate payback from data analytics investments.
Understanding the Potential of Smart Grid Data Analytics
6
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
3.2
Power Consumption Data
There are two basic types of power consumption data (as measured in kilowatt-hours or
kWh): time-interval data and register reads. Register reads provide absolute values useful
for billing purposes, while interval data provides more granular data (typically at 15-minute
or hourly intervals) for trending and analysis. Interval data is particularly useful for data
analytics, since it is granular and neatly arrayed from a temporal standpoint.
3.3
Additional Data Types and Functions
Examples of new data available from smart meters include power quality data (e.g.,
voltage, reactive power), outage alerts, and tamper alerts. Examples of new functionality
include the ability to deliver price signals and messages to devices inside the home and
remote connect/disconnect for service provisioning.
Conventions for collecting and managing power consumption data are well established.
Determining whether, when and how to collect power quality data is a new smart grid
frontier. Planning is complicated by existing systems and processes. For instance, a turnkey
conservation voltage reduction (CVR) application may not be designed to accommodate
voltage readings from meters.
Leveraging meter data across the organization is a give and take exercise of discovering
opportunities and collaborating to develop creative solutions. AMI pioneers cite a best
practice of creating cross-functional task forces that bring experts from different business
units (such as metering and operations) together to establish business requirements and to
solve technical problems like how to integrate new smart meter data with existing systems.
3.4
Analytical Methods
Collecting accurate, timely and relevant data is the bedrock of any data analytics program,
but the data needs to be put into an appropriate context to become useful information.
Five fundamental analytical data transformations have immediate relevance to smart grid:
aggregations, correlations, trending, exception analysis, and forecasting. Many high-value
analytical processes combine several of these techniques as part of an overall analytical process.
Each technique is discussed in more detail in the following sections, along with relevant
smart grid examples.
Understanding the Potential of Smart Grid Data Analytics
7
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
3.4.1 
Aggregations
Simply put, an aggregation is a summary of data using set criteria. Because smart meter
data is atomistic (i.e., it is associated with a metering endpoint), it can be aggregated in
different ways to serve planning purposes. For instance, the meters connected to individual
transformers can be aggregated together to identify transformer loading patterns.
Combining homes or businesses into demand response pools to deliver sizable demand
reductions (or ‘negawatts’) is another aggregation supported by smart meters. ‘Virtual
meters’ are arbitrary user-defined aggregations that combine data from multiple meters
that share a common characteristic. A typical virtual meter aggregation combines meters
with common linear relationships to support distribution planning and analysis (e.g.,
common substations, feeders or transformers).
3.4.2 
Correlations
Correlations identify statistical relationships between related data that are useful for
building predictions. A basic smart meter correlation is the relationship between outdoor
air temperature and power consumption. The fact that heat waves drive spikes in power
consumption is well known. Statistical correlation using time-interval consumption data
makes it possible to build algorithms that predict the size of demand spikes using forecast
temperature. Correlations can also be built using multiple variables (i.e., multivariate
correlations). For instance, cloud cover, humidity and time of day can be added to the
equation to further refine peak predictions. The ability to align data temporally, spatially,
or across other attributes is important for building correlations. For instance, instead of
relying on measures such as daily minimum, maximum and average temperatures for an
entire metropolitan area, the collection of nearby 15-minute time-interval temperature data
is far more powerful for building weather/power consumption correlations. In fact, data
analytics pioneers are using this type of data to build analytical models that measure the
energy efficiency of individual commercial properties.
Understanding the Potential of Smart Grid Data Analytics
8
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
3.4.3 
Trending
Trending is one of the most basic forms of analytics, and it can be an obvious win for
improving customer relations and service quality using smart meter data. A web page
that shows customers a simple consumption data trend line can help them relate power
consumption to household activity. The ability to overlay multiple trend lines together is
also valuable for purposes such as comparing consumption across similar seasons and
times of day. Trending is a useful analytical process for any time series data.
3.4.4 
Exception Analysis
Exceptions are unexpected or abnormal conditions. A missing meter read, for instance, is
an exception event. The ability to analyze exceptions over time is valuable for identifying
problems in communications and measurement infrastructure, as well as in the distribution
grid. Equipment failure is useful for homing in on a subset of data for other forms of
analysis. In the case of a blown transformer, it may be useful to build a historical trend of
transformer loading prior to the failure. Once pre-failure patterns are identified, they can
be used to build predictive algorithms useful for preventing future failures. Trending of
exception events can also help identify component degradation or operational breakdowns.
3.4.5 
Forecasts
Forecasts are predictions of future events or values using historical data. For instance,
a forecast of power consumption for a new residential subdivision can be created using
historical data from similar homes. Forecasts can also be built using correlation data.
For instance, a forecast could be as simple as predicting one incremental megawatt of
power consumption for each one-degree rise in summer temperature above 78 degrees
Fahrenheit, or it could be a more granular and sophisticated set of algorithms that forecast
maintenance expenses based on the age of equipment, utilization trends and past service
trends for similar equipment.
Understanding the Potential of Smart Grid Data Analytics
9
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
4
DATA AND ANALYTICS APPLICATIONS
Smart grid data and analytics will revolutionize the way power is managed, delivered, and
sold. This section of the whitepaper examines four categories of analytical applications:
revenue management, customer engagement, distribution optimization and AMI network
management. Specific examples of AMI data analytics appear in each of the relevant sections.
4.1
Revenue Management
Key revenue management applications include load forecasting, theft detection, prepay,
rate plan modeling, and demand management.
4.1.1 
Load Forecasting
The ability to accurately predict loads supports multiple utility business processes,
including power generation, power trading, capacity planning, and demand management.
AMI data revolutionizes load forecasting by providing granular point-of-consumption data.
This granular data is useful for building forecasts in a variety of contexts: to determine
power flow loads on specific parts of the distribution infrastructure, to aggregate
consumption up to locational marginal pricing nodes (LMPs) on the transmission grid in
support of power trading, and to plan load shed events (preferably demand response and/
or dynamic pricing, not rolling blackouts). Suffice it to say that improved load forecasting is
a killer analytic app for the smart grid, and time-interval data is the fuel that feeds it.
4.1.2 
Theft Detection
AMI supports theft detection in a number of ways. The first is the elimination
of electromechanical meters that can be tampered with to slow or even reverse
register values. Switching over to new, accurate digital meters quickly weeds out
electromechanical meter bandits.
Potential theft or technical losses can also be identified by comparing smart meter data
with measurements from upstream sensors attached to transformers or feeders. Simple
check-sum comparisons identify power loss. If total consumption at the meter level is
less than at the feeder or transformer level, then field technicians can be dispatched to
investigate the cause of the power loss, including inspection for bypass connections, meter
tampering, or a technical condition (such as runaway current in underground lines).
4.1.3 
Prepay
Together, theft detection and prepay can be considered the killer smart grid apps for
emerging markets. Thanks to pay-as-you-go mobile phone plans, prepay is well established
in emerging markets. Prepay has strong appeal wherever electrification is high but
consumer borrowing (and credit) is low. Offering prepay power purchase plans has multiple
benefits in these “cash societies.” First, prepay is comfortable and familiar to consumers.
Second, prepay helps utilities limit exposure to credit risk. Finally, if the utility offering a
prepay plan provides consumers with appropriate analytical tools, consumers can more
Understanding the Potential of Smart Grid Data Analytics
10
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
effectively manage power purchases and, most importantly, avoid service interruption. It
is important to note that, even in established economies, some utilities have reported that
prepay plans are quite popular as a way for consumers to manage power budgets.
Data analytics, proactive communication and ease of purchase are all key success factors
for prepay plans. The ability to notify consumers that they need to replenish their account
balances is critical. This can be done by comparing remaining balances with the rate of
consumption. More rigorous analytics improve subscriber account management. For
instance, historical consumption patterns can be analyzed to estimate power consumption
“burn rates” for comparison against remaining balances in order to estimate the number of
days before account balances are depleted. Proactive communication via the channel(s) of
subscriber’s choosing (phone, text, web portal, email, etc.) make it easy for consumers to
track their account’s standing.
4.1.4 
Rate Plan Modeling
Many utilities are actively piloting and rolling out new variable price structures. Examples
include time-of-use or dynamic pricing as a way to reduce peak consumption. Smart
meter data can be used to analyze and plan different rate structures, while adhering
to requirements like revenue neutrality or non-discrimination against low-income
demographics or the elderly.
Rate planning makes intensive use of data analytics. A typical sequence is:

Pricing pilot design. Determining the overall objectives of the pilot, what is to be tested and the
desired findings.

Planning rate structures. Using historical consumption data, including timeframes where peak
demand occurred, as well as data from the design and results of other pilots (especially peak
reduction results of different rate structures and their impact on utility bills).

Selecting pilot participants. Sample size and participant diversity needs to be sufficient for
segmentation analysis. Examples of data useful for segmentation include the heating and cooling
systems for various homes (e.g., central air conditioning versus window air conditioners versus no
system), high income versus low income, home office workers versus commuters, etc.

Gathering and analyzing pilot data. This includes the effectiveness of various pricing plans in
reducing peak consumption and their impact on bills.

Submitting proposed rate structures for regulatory approval. This includes supporting analytics that
prove out the benefits of the new rate plan for ratepayers and the utility.

Marketing new rate structures. This also entails providing consumers with tools and analysis to
pick the rate structure that is most beneficial for them (see Rate Plan Selection in the Consumer
Engagement section below).
Advanced statistical analysis, such as cluster analysis, is needed to develop market segmentation.
Understanding the Potential of Smart Grid Data Analytics
11
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
4.1.5 
Demand Management
Demand management program managers can use data analytics to forecast peaks and
to plan demand management events, including when and where to call peak demand
events and who to include (for instance, an entire service territory may be called on to
reduce generating requirements, while a smaller metropolitan area can ease distribution
constraints). Power consumption before, during and after a peak event also needs to be
tracked and analyzed, especially for peak time rebate programs that reward customers
for not consuming power. In this case, an analytical process for identifying the amount of
non-consumption to pay is needed in order to prevent people from ‘gaming’ the system by
ratcheting up demand just prior to an event. Analytics of consumption patterns are also
useful for validating demand reductions claimed by third-party demand response service
providers and for general market operations in open demand response markets.
4.2
Consumer Engagement
Sharing smart meter consumption data with customers opens new opportunities for
actively engaging consumers in energy management. A secure web portal where
customers can log in and view consumption trends is an obvious starting point, but careful
consideration needs to be given to the overall user experience. How easy is it to log in?
Can a forgotten user name or password be retrieved? Can the user select the timeframe
of interest and perform comparisons between time periods? These basics are merely a
starting point for delivering more advanced analytics like benchmark comparisons with
other similar households and proactive recommendations for how to save energy.
Consumer engagement is emerging as a lynchpin requirement for utilities implementing smart
grid and facing questions about what consumer benefit is accruing from smart grid initiatives.
Consumer web portals with analytics capabilities are now a must-have requirement. More
advanced consumer engagement tools enabled by data analytics are discussed below.
4.2.1 
Conservation Tips and Suggestions
Behind-the-scenes profiles and automated algorithms are the ‘secret sauce’ for targeted
power conservation tips and recommendations. Profile data includes home square footage
and data about major heating and cooling systems and other electrical appliances. This
data can be gathered using forms that customers voluntarily complete or external data
sources like tax assessor databases. Examples of advanced analytics for generating targeted
recommendations include correlations between consumption and weather data that
may lead to a recommendation for weather sealing or more efficient heating and cooling
systems, and trending to identify excessive baseline power consumption from “vampire
appliances” like PCs, set-top boxes and chargers that consume power even when not in use.
Understanding the Potential of Smart Grid Data Analytics
12
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
4.2.2 
Rate Plan Selection
Instead of forcing customers to adopt a new pricing model, many utilities (and the public
utility commissions that regulate them) are taking a more cautious opt-in approach where
customers elect to enroll in a new rate plan, such as dynamic pricing or time-of-use pricing.
This is an important economic decision for customers who will expect bill savings in return
for adopting a new plan. Customer decision-support tools can aid the opt-in decision
process. An obvious analytics requirement is the ability for customers to model a new rate
plan using historical consumption to determine whether the plan would have saved them
money. However, since the point of many of these pricing plans is to spur peak-time use
reductions that are a departure from past consumption behavior, additional analytical tools
are needed to help customers understand the impacts of different actions, such as setting
thermostats higher during a critical peak pricing event. A ‘best-plan selector tool’ can also
recommend pricing plans using a wizard interface supported by back-end algorithms.
The end goal for rate plan modeling and selection is a win-win: to coach consumers to
adopt the best plans for their pocketbook – and for utility cost to serve.
4.2.3 
Efficiency Program Measurement and Planning
Because of its multivariate nature, measuring the effectiveness of energy efficiency
programs is a complex undertaking. For instance, if a customer participated in a free
CFL program and also signed up for a home energy audit, how much savings should be
attributed to each program? Limited consumption data compounds the difficulty. For
instance, one monthly kWh consumption number makes it difficult or impossible to create
a detailed timeline that connects consumption reductions with actions taken at a specific
point in time or to adjust for externalities, such as a heat wave or cold snap. For instance,
one would expect that weatherization improvements would yield significant power
conservation during extreme weather events, something that can only be measured with
the benefit of time interval consumption data.
The energy efficiency audit community is just beginning to recognize the opportunities
for improvement made possible by smart meter data. We expect auditors and efficiency
program managers to become important customers for smart meter data.
4.3
Distribution Optimization
Leading utilities are beginning to identify opportunities to optimize power distribution
management using smart meter data. Two immediate opportunities are in the areas
of outage management and distribution network planning. An additional longer-term
opportunity to wield smart meter voltage data for conservation voltage reduction is
discussed further in the Futures section, which concludes this whitepaper.
Understanding the Potential of Smart Grid Data Analytics
13
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
4.3.1 
Outage Management
Enhanced outage management ranks as the first true real-time application for smart
meters. But using ‘last-gasp’ meter outage alerts to drive real-time outage notifications into
an outage management system is not a plug-and-play endeavor. When it comes to outage
notifications, smart meters can be very ‘noisy’ over reporters. Examples include momentary
interruptions in power caused by line flicker (vegetation brushing against power lines, for
instance), or outage ‘message storms’ caused by thousands of meters all reporting power
interruptions as part of a larger outage. Real-time analysis (best performed by a complex
event processor, or CEP) is needed to handle both contingencies.
An example of a CEP in action would be rolling thousands of outage alerts up to a common
upstream node on the distribution grid (such as a common feeder or a substation) to
create one master outage instance and to support targeted crew dispatch. Another
example is a business rule that suppresses the creation of a new outage incident if an
outage restoration alert quickly follows an outage alert – a business rule that eliminates
spurious reporting of line flickers as outages.
4.3.2 
Distribution Network Planning
Smart meter data can be used to improve distribution network planning. Historically,
distribution sizing is a very conservative exercise where planners err on the side of
overcapacity, absent any detailed data on utilization trends, especially for transformers.
Smart meter data can be aggregated to reflect the transformers they are connected to,
and then utilization can be compared to the capacity of the transformer to build detailed
capacity utilization trend analysis.
Examples of questions that this type of analysis can answer include: What percentage
of the time is a transformer operating within 10 percent of its peak rating? Are there
certain times of day or times of year when transformers are nearing overload? What is the
minimum size transformer that could be used to replace an aging transformer?
Utilization patterns can also be compared against pre-failure data for similar transformers
to begin building proactive asset maintenance and failure prevention.
In summary, distribution network planning and analysis using smart meter data is the first
wave of distribution optimization enabled by smart meter data analytics.
Understanding the Potential of Smart Grid Data Analytics
14
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
4.4
AMI Network Management
Savings garnered through remote meter reading can quickly disappear if extensive
manual troubleshooting is needed to configure and manage the AMI network. Any data
communications network needs to be proactively managed and administered, and AMI
networks are no exception.
Going forward, personnel responsible for ensuring the reliability of the end-to-end meter data
collection process are going to need access to data about network performance and reliability.
4.4.1 
Service Level Management
A strong end-to-end AMI solution includes data about meter and AMI network
performance, including response times, data packet losses, message retries, communication
outages and device failures. All of this data needs to be rolled up to create an overview of
AMI service levels. Key metrics include network reliability and on-time message delivery
performance. Service level metrics can be used to support planning decisions, such as
determining where to invest in more capacity and which applications or types of data
need to be throttled back to ensure bandwidth is available for critical data (like collecting
consumption data to support the meter-to-cash process).
4.4.2 
Network and Device Configuration and Troubleshooting
The scope and scale of AMI networks in terms of the number of end-node devices (i.e.,
smart meters) is unprecedented. The device management challenge that smart meters
create should not be underestimated. Each meter needs to be meticulously tracked,
including make and model, warranty information, firmware release and configuration
settings. The same holds true for head end systems (also known as data collectors) that
interface with the meter. Further, the accuracy of each meter needs to be validated to
ensure ratepayers are treated fairly. All of this data needs to be tracked and managed.
Key reports and analytics about smart meters include the number of meters on a specific
version of firmware, the current configurations for all smart meters, problem meters and
network segments that are unreliable and causing missing or inaccurate readings, and so
forth. Analytics data supports asset management tasks like planning upgrades (e.g., adding
repeaters to fix intermittently dark network segments), swapping out bad equipment, and
future purchase decisions.
Understanding the Potential of Smart Grid Data Analytics
15
Copyright © 2012 Greentech Media Inc., eMeter. All Rights Reserved.
4.5
The Future of Smart Grid Data Analytics
The analytics opportunities smart meters present are just beginning to be identified. As this
whitepaper illustrates, time-interval consumption data from smart meters alone presents
a plethora of opportunities ranging from planning new dynamic pricing rate structures to
distribution equipment sizing, load forecasting, theft detection and more. But time-interval
data is merely the tip of the smart grid data analytics iceberg. Smart meters are capable of
producing additional types of data, each opening a range of new opportunities and uses.
Here is a brief summary of additional types of smart meter data and some of the analytics
opportunities we expect to emerge in the near future:

Voltage data.
With smart meters, it is now possible to collect voltage readings from the
edge of the distribution network. This data can be collected and matched with other
voltage readings further upstream in the distribution network, then analyzed to optimize
voltage regulation. Voltage conservation can be used for technical demand response and/
or to improve overall power delivery efficiency.

Power quality data
. Reactive power readings from smart meters can be captured and
analyzed to measure power quality and to determine adjustments in the distribution
network to reduce power harmonics, increase delivery efficiency, and provide a high-
quality product to customers.

Peak demand readings.
Time-interval consumption data follows a time-based sampling
methodology. Within any given timeframe, there will be a maximum draw of power – in other
words, a peak demand reading. Peak demand data can be analyzed to learn more about
consumption patterns, including the identification of ‘heavy-hitter’ appliances like pool
pumps, central air conditioners, electric hot water heaters, and in the future, electric vehicles.

Home area networks (HAN)
. Many of the dynamic pricing trials underway include a HAN
component that connects energy management systems inside the home with utility systems.
While the ultimate vision is some form of elegant machine-to-machine (M2M) interaction
to achieve peak reduction, some level of analytics by utility personnel will be necessary
to orchestrate power consumption. It will be necessary to be able to analyze the portfolio
of available customer load assets at any given time and to interrogate their current status
(potentially including thermostat settings and current indoor temperatures, for instance),
including modeling the amount of power made available by different actions. Although HANs
will be outside of utility direct control, it will still be desirable to create data records for them
and to make them ready to receive communications (provisioning and commissioning).

Electric vehicles
. Electric vehicles are a new type of consumption asset – a mobile device
that can draw power at multiple points across the grid. Data analytics will enable utilities to
visualize and analyze EV charging trends, create new charging plans, and to identify changes in
distribution sizing and planning necessary to accommodate these new power-hungry devices.

Utilities that embrace smart meter data analytics as a core business function will reap
substantial rewards in the form of improved business operations, more effective revenue
management, and technical improvements in power delivery. Data analytics unlocks smart
grid potential and turns opportunity into business reality.
For more info please visit www.greentechmedia.com/sponsored/resource-center/
A joint white paper produced by:
A Greentech Media Company