Sioshansi F.P. Smart Grid: Integrating Renewable, Distributed & Efficient Energy

Подождите немного. Документ загружается.

storing the data in a database; and sending it to the billing system

according to each customer's billing schedule.

When evaluating solutions as part of a smart grid investment,

utilities generally have two overriding and fundamental

objectives:

• Maximize Business Benefits Realization—most utilities

that have deployed advanced metering or smart meter

systems have yet to realize benefits beyond reduced meter

reading costs. A good application platform enables added

benefits through the addition of specific data applications,

the ability to apply business rules to raw data such as

outage alarms, and a robust integration platform to allow

the implementation of new applications and new data

collection technologies.

• Manage and Minimize Risk—prominent smart grid

implementation risks include technology obsolescence,

integration risk—as a rule, the more interfaces required, the

greater the risk—and changes required to core utility

business systems such as billing and the customer

information system (CIS), and scalability. The complexities

of the IT challenge should not be underestimated. Robust

applications and formal IT integration approaches must be

applied to the implementation and operation.

Achieving these objectives requires understanding the

challenges. As with any large IT project, there are many

obstacles that must be overcome to achieve the desired

results. The goal is typically a reliable system that generates

accurate results—for example, customer bills—and one that

enables the use of efficient business processes—for example,

processes that are primarily automated and require minimal

human intervention.

541

Standing in the way of achieving this goal are five key

challenges to smart grid IT systems:

• Obtaining data reliably from smart meters, sensors,

controllers, smart appliances, and any other devices

included in the overall smart grid project.

• Ensuring that the data accurately record or represent

what they are supposed to record or represent.

• Defining data systems of record—which database is the

master, when, as with the smart grid, multiple IT

applications are using the same data.

• Managing a large and diverse communications

network involving hundreds of thousands or millions of

communications devices. Management of such networks is

new for utilities—indeed, for any company, because the

smart grid has a whole new class and scale of devices to

manage.

• Integrating disparate software applications. For

example, data from smart meter network head-end software

developed a year ago may need to be sent to a billing

system that is three decades old.

The following sections describe these challenges in greater

detail and specificity. The concept is that solving the problem

begins with understanding the problem.

Smart Grid Data Reliability

The smart grid establishes a new level of required data

reliability for utilities. “Reliable” data are data that can be

relied upon for their intended uses; the goal is for customers

and other data users, such as utility personnel, to have a high

542

level of confidence in the data. Higher reliability is needed for

smart grid data because the data will be used for more

complex and/or numerous functions. These range from

demand response to outage management. One indicator of the

significance of this issue is that, as a rule of thumb, a typical,

well-run, large-scale smart meter system misses up to 4

percent of the interval usage data it is supposed to record and

retrieve each month. For a million-meter system, this totals

over 28 million missing data intervals per month. Retrieval of

total consumption values from meters—the meter's register

read—is more reliable, because only one such read is needed

per day rather than 24 (hourly interval data) or 96

(quarter-hourly interval data). Even so, a typical, well-run,

large-scale smart meter system can miss more than 1 percent

of such data.

In fact, the Ontario MDM/R requires that valid

data be provided from at least 98% of smart meters.

In the

USA, each utility has set its own requirement for the amount

of missing data that is permissible.

Pacific Gas and Electric Company, Advanced Metering Infrastructure,

January 2010 Semi-Annual Assessment Report and SmartMeter™

Program Quarterly Report (Updated), January 31, 2010 at 35.

IESO MDM/R Operational Best Practices, April 2011.

Demand Response

One example of complexity is the use of interval data to

create billing determinants for dynamic pricing and other

demand response programs, further described in the chapter

by Chapter 3. For time-of-use rates, the interval data must be

processed into peak and off-peak usage. The 3,000 or so

intervals recorded each month per meter must be processed

into one total for peak times and a second for off-peak times.

543

In the past, utilities have needed and handled only a single

number each month. To deal with this new data volume,

utilities implementing smart meters are also implementing

MDM systems.

Another complex demand response function is measuring

peak demand reduction. This involves calculating a

“baseline” load, which is the load the customer would have

consumed absent the peak demand event. The reduction is

determined by subtracting the actual load from the baseline

load. The baseline load is often determined by taking the

average load during the same time period (e.g., noon to 6 PM)

on the three highest of the ten previous similar days.

The

baseline must be accurate, because it must comply with

regulator-approved tariffs and is used to calculate cash

payments. In addition, the methodology must be sound—and

auditable—to deal with the issue of customers who may

attempt to “game” the system and create artificially high

baselines. Finally, because the baseline load changes every

day as the calendar advances, this calculation must be done

every day—or at least every event day.

PJM, PJM Manual 11: Energy & Ancillary Services Market Operations,

Revision 45, June 23, 2010. Available at http://www.pjm.com/~/media/

documents/manuals/m11.ashx.

Consumer Engagement—Bill-to-Date Calculation

Another instance of complexity is bill-to-date calculation.

Customers have said in surveys that they would like to see

this information so they can manage their usage and

electricity bills better.

This function has been piloted and is

already in use by some utilities and Texas electricity retailers.

Bill-to-date requires daily bill calculations, as opposed to

544

the traditional monthly, with the daily estimated amounts

shown on a utility's or retailer's website.

Chapter 15 further describes the attractive features of “bill-to-date”

functionality.

eMeter Strategic Consulting, PowerCentsDC Final Report, September

2010 (available at http://www.powercentsdc.org).

Ibid.; Jeff Read, Reliant's e-Sense Looks Smart, Smart Meter Marketing,

January 7, 2011, available at http://www.smartmetermarketing.com/

reliants-e-sense-looks-smart/.

Bill-to-date also requires reconciliation. This process occurs

after the actual bill is issued. The purpose is to make sure that

the total of the daily estimates calculated prior to the bill

issuance equals the actual bill that was sent to the customer.

This is done when each bill is issued by taking the actual bill

amount and going back and adjusting the daily estimates to

match the total bill actually issued. In practice, this means a

slight decrease or increase in many of the daily estimates to

ensure that the total of the daily estimates matches the actual

bill issued. For example, the daily estimates for a given

customer may add up to $99.00. The actual bill may then be

$100.00. Though the difference is small, the customer's

confidence in data provided by the utility depends on the data

being accurate. No customer would stand for a utility saying,

“Your bill is approximately $100.00; please pay this amount.”

Of course, customers understand estimates during a billing

period. But unless told it is an estimate, customers assume the

bill is 100% reliable.

545

The purpose of the daily bill estimate reconciliation process is

to avoid customer questions and confusion caused by

differences between the total of daily estimates shown on the

utility's website, or emailed to customers, and the actual bill

amount. The frequency of situations where the final bill does

not match the total of the daily estimates depends on the

specific utility regulatory situation. Common causes of

differences between estimated and actual bills are rate

changes and seasonal changes. Often, regulators make rate

changes retroactive; since the estimates are done before the

retroactive rate change, the estimates will not match the actual

bill.

In the case of seasonal changes, winter rates are applied to

part of the bill and summer rates to the other part. This occurs

for utilities with differences in seasonal rates, which make up

perhaps half of U.S. utilities. There are different ways of

applying the seasonal changes. Some utilities prorate the

usage when a bill is issued after a seasonal change, allocating

usage to summer and winter based on the number of days in

the billing period in each of the seasons. For example, there

may be 10 summer days and 20 winter days, in which case

the utility would assume 10/30 of the consumption occurred

during the summer and 20/30 during the winter period. This

proration cannot be done until the final meter read is taken at

the end of the billing period. Because this calculation is done

based on the proportion of days in each season, the total of

the daily estimates will not match the actual bill unless the

end-user's consumption is perfectly balanced between the

seasons. In the case of proration, a perfect balance occurs

only when the average daily consumption in the summer days

equals the average daily consumption in the winter days.

Accordingly, the vast majority of bills issued after a seasonal

546

change will have a difference between the total of the daily

estimates and the actual bill amount. By reconciling the daily

estimates—adjusting them to match the actual bill

total—utilities can maintain confidence of customers in the

bill calculations.

Outage Alerts

Other important issues relating to data reliability involve

outage alerts from meters and restoration verification

messages. Most smart meters emit a “last gasp” when the

power goes out. The last gasp is a radio transmission made

possible by a capacitor in the meter that stores energy and can

use it after the power goes out to send the outage alert

message. Utility outage management systems can use these

last gasp messages to identify the location of outages and,

with some additional intelligence, sometimes identify the

substation or feeder fault location that caused the outage. The

additional intelligence requires mapping the individual meters

against distribution system elements, such as feeders and

substations. For example, if all the alerts are from a single

feeder, the outage management system of the utility can

conclude that the specific feeder is the source of the problem.

There are a few problems with last gasp messages that must

be managed by utility IT systems in order to provide

confidence to data users that the outage alerts can be relied

upon. The first is that all last gasp messages may not be

received by radio receivers in the smart meter

communications network. The business rule that solves this

problem is straightforward: if the message is not received, the

utility cannot assume power is on. The second problem is that

meter-generated outage alerts may overwhelm the outage

management system. This occurs especially in the case of a

547

large-scale outage usually caused by a major storm; tens or

hundreds of thousands of meters can lose power. The ideal

solution in these cases—as implemented by utilities such as

JEA (formerly Jacksonville Electric Authority)

—is to

“filter” the outage messages and forward only high-priority

alerts. Most utilities define these as alerts coming from

hospitals, fire stations, police departments, and other public

health and safety authorities. The meters on such facilities are

designated “bellwether” meters to signify their high value and

priority.

JEA, Meter Data Management System, Presentation at Autovation

Conference, Nashville, Tennessee, October 2006.

The third problem is false alarms. This problem is perhaps the

most insidious, because it causes distribution system

operators to have low confidence in the outage alerts from

smart meters generally. False alarms are created for many

reasons. Meters may report outages when reclosers have

actually already restored power. This situation can be solved

by programming the meters to report outages only if the

outage exceeds the time threshold programmed into the

reclosers, but this requires that the system managing the

meters maintain the synchronization of values between the

reclosers and meters. In other cases, meters report outages

caused by crews conducting scheduled work. The utility

system can be set up so that such scheduled outages are

excluded from reporting, provided the smart meter system,

the work management system—the utility CIS often performs

this function—and the outage management system are

coordinated.

548

The other source of false outage alarms is meter detection

issues. Some utilities have turned off smart meter outage alert

systems, because the systems generate too many false alarms.

For various and sometimes unknown reasons, smart meters

may send outage alerts that do not reflect actual outages. This

situation occurs when the meter's outage detection mechanism

is triggered by an event—perhaps a voltage sag—that the

meter interprets to be an outage. In fact, the event may not

have caused an outage. Overall, the IT system must have the

flexibility to implement business rules that, over time, can

learn that these outage alerts are false and not to forward them

to the outage management system. Interpreting meter alerts

improves with experience and requires adjustments over time

as the utility distribution operations team becomes more

familiar with the sensing capabilities of the meters and the

like.

The IT system can help if it has the capability to segregate

different types of outage alerts or perform other analysis, such

as a correlation analysis. If multiple alerts are received at the

same time from different meters attached to the same circuit,

the likelihood is high that an actual outage has occurred.

Another example is based on history; some meters have a

history of reporting spurious outage alert data. The right IT

application for handling outage data can track such “suspect”

meters and exclude alerts from them as likely false positives.

Data Accuracy

Data accuracy is about whether data delivered by smart grid

systems are correct. Accurate data are reported data that

correctly represent the situation that actually occurred.

Examples of accuracy are that a certain number of kWh was

549

consumed during an interval period or that a certain voltage

occurred as reported by the smart grid system. Four common

data accuracy issues are as follows, and each is described in

greater detail later:

• Sum check: do the intervals add up to the correct total?

• Device mis-programming: were the parameters set

properly?

• False zero values: is a zero in a dataset an outage, zero

usage, or missing data?

• Meter reset: did the meter improperly reset its values to

zero?

Accuracy differs from reliability. Accuracy is simpler; it is

about whether a value is correct. Reliability encompasses the

broader context described above: will the data arrive on time,

will it be complete, will it represent what it is supposed to

represent? An hourly interval smart meter provides an

example: from a reliability standpoint, the goal is to receive

all of the interval usage data recorded each month from the

meter, which would be approximately 720 intervals (30 days

times 24 hours). From an accuracy standpoint, the goal is that

the interval data accurately reflect the kWh consumed.

These four specific data accuracy issues are further described

below.

Sum Check

One test for data accuracy for a smart meter is a “sum check.”

If the data are accurate, the sum of the intervals will equal the

difference between the starting and ending point. An example

illustrates this. Let's assume the starting value—the “register”

read of the meter—on June 1 is 10,000 kWh and the ending

550

Sioshansi F.P. Smart Grid: Integrating Renewable, Distributed &amp; Efficient Energy

Подождите немного. Документ загружается.

Sioshansi F.P. Smart Grid: Integrating Renewable, Distributed & Efficient Energy