Note how closely HazenFilter predicts actual headloss
all things water®
SPRING
(R represents how well the predictions fit the data on a scale of 0-to-1, with 1 being a
100% fit. )
After calibrating HazenFilter with data from two BAF units at NNWW, the utility then tested the model against data from a third BAF unit with
1.0-mm media:
VALIDATION 1
HazenFilter accurately predicts BAF headloss
(Headloss is a drop in the pressure needed to pull water through a filter. When it gets too high, the filter needs to
be cleaned.)
160 days
200 days
240 days
Headloss (m)
HazenFilter prediction
INTRODUCTION
CALIBRATION
VALIDATION 1
Headloss actual
Loading rate
VALIDATION 2
Media size: 1.0 mm |
R
2
=0.84
SUMMER
5
4
3
2
1
0
2
0
)
40 days
Media size: 1.4 mm |
80 days
120 days
Time (days)
=0.85
HazenFilter achieved an 85% fit for this BAF unit.
i
2
WINTER
As the lines show, HazenFilter’s predictions came very close to the actual results. NNWW also tested the calibrated model against eight months of data from a fourth BAF unit with 1.4-mm media:
VALIDATION 2
The rate at which wastewater flows through the filters, expressed in volume per unit area per unit time.
These strong results prove that utilities can use HazenFilter—rather than expensive physical tests—to predict filter performance under a wide range of conditions over long periods of time.
HAZENFILTER
Newport News Waterworks (NNWW) used eight months of historical data to test how well HazenFilter predicted headloss in two of the utility's biofiltration (BAF) systems.
CALIBRATION
Headloss is a drop in the pressure needed to pull water through a filter, measured in meters of liquid or “head." When it gets too high, the filter needs to be cleaned.
HazenFilter
6
5
4
3
2
1
0
Loading rate (gpm/ft