POWER May 2021 - 34

COAL POWER
How (and Why) Using
DVR Modeling Will Improve
Power Plant Performance
Traditional power plant performance indicators do not always provide reconcilable
data. This means plant personnel may be making operational changes based on
inaccurate information. An innovative data validation and reconciliation modeling
method could greatly reduce uncertainty and allow proper decisions to be made.
Greg Kanuckel and Frank Todd
he power
T
industry is changing,
and the methods used for tracking
plant performance need to change
with it. The industry has been reliant on
traditional monitoring methodologies
for years, but there's a new method
that promises better results: Data Validation
and Reconciliation (DVR) modeling.
DVR methods have been developed
to help address poor data quality.
They provide a way to accurately determine
key performance parameters for
a fossil cycle.
Historically, important performance
indicators have been determined using
measured plant data, which is obtained
during plant testing or pulled from a data
historian. The same measurement data
is often used as input for plant operations,
equipment monitoring, unit heat
rate curve development, billing and accounting,
emissions monitoring, maintenance
scheduling, and other important
business and operational purposes.
These measurements are performed by
equipment and instrumentation in suboptimal
condition, and therefore have
the potential for significant bias error in
the values. Traditional plant monitoring
strategies cannot account for this, so
biased measurement data has inevitably
been used as input for business and operational
decisions.
DVR methods minimize the potential
effect of significant bias error in process
data. By utilizing measurement
redundancy and uncertainty for the
entire process, and applying mass and
energy conservation laws, plant operators
are able to identify measurements
that are biased, and allow for their correction.
Additionally, DVR can monitor
the process for the data set that represents
the true state of the system,
thereby allowing ongoing system moni34
1.
The pipe configuration shown here illustrates a situation where one measured flow splits
into two separate parallel flows. However, the actual instrument-measured flowrates (shown
on the left) do not satisfy mass conservation laws. The reconciled results (shown on the right)
reduce uncertainty and represent the most probable true state of the system. Courtesy: GSE
Solutions
toring and calculation of accurate performance
indicators.
Understanding the Basics of DVR
All measurements taken by instruments
are subject to measurement error. The
type of instrument used, where and
how it is installed, and degraded conditions
will influence the amount of error.
A 100% perfect measurement system
just doesn't exist, so it's better to understand
the error and to account for it.
DVR methodology identifies the amount
of error that is most probable for a measured
value-how far the measurement
is from the " true " value.
The basic idea of the DVR method
is to establish redundancies between
process measurements, and apply the
measurement uncertainties, and mass
and energy balances, to the system to
determine a solution that best satisfies
the physical constraints. Figure 1
helps illustrate the DVR methodology.
In it, one measured flow splits into two
separate parallel flows, and each flow is
measured. For such a configuration in reality,
the conservation of mass dictates
that the parallel flowrates should sum to
equal exactly the inlet flowrate.
However, these are not perfect flowmeters.
Each has an uncertainty of
+/-5% of the measured flowrate. This
is reflected in the summing of the meawww.powermag.com
sured
flowrates shown on the left of Figure
1. The total measured outlet flow of
680 lb/hr + 360 lb/hr = 1,040 lb/hr, which
does not match the measured inlet flow
of 1,000 lb/hr, and therefore mass conservation
is not satisfied.
This example shows functional redundance.
The redundancy present in
the system allows for the measurements
to be reconciled, or corrected,
to satisfy the mass balance. The mathematics
for this reconciliation process
are described in the German standard
VDI-2048.
The
reconciliation
process accomplishes
two things. First, the measured
values are converted to values that satisfy
the mass and energy balances. Second,
the uncertainties associated with
each measurement are also reconciled
to achieve a more accurate description
of the process.
The reconciled results and uncertainties
are shown on the right side of Figure
1. You can see in this illustration that
the inlet flowrate is equal to the sum
of the outlet flowrates. Additionally, the
uncertainties for each flowrate measurement
have decreased. This set of
reconciled results represent the most
probable " true " state of the system,
according to the measured values and
input uncertainties.
Real-world power plant applications
POWER | May 2021

http://www.powermag.com

POWER May 2021

Table of Contents for the Digital Edition of POWER May 2021