See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/265220212

Physical Processes Resulting in Geysers in Rapidly Filling Storm-Water Tunnels

Article  in  Journal of Irrigation and Drainage Engineering · March 2011

DOI: 10.1061/(ASCE)IR.1943-4774.0000176

CITATIONS READS
29 1,074

6 authors, including:

Steven J. Wright James W. Lewis

University of Michigan U.S. Army Corps of Engineers
102 PUBLICATIONS   1,399 CITATIONS    13 PUBLICATIONS   163 CITATIONS   

SEE PROFILE SEE PROFILE

Jose G. Vasconcelos
Auburn University
115 PUBLICATIONS   966 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Description of unsteady, two-phase flow conditions in stormwater systems. View project

Sediment-water flows in stormwater applications View project

All content following this page was uploaded by Steven J. Wright on 06 June 2016.

The user has requested enhancement of the downloaded file.

World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5380

Physical Processes Resulting in Geyser Formation

In Rapidly Filling Stormwater Tunnels

Steven J. Wright1, James W. Lewis2 and Jose G. Vasconcelos3

1
Member ASCE, Professor, Civil and Environmental Engineering, 113 EWRE,
University of Michigan, Ann Arbor, MI 48109-2125; PH (734) 764-7148; FAX
(734) 763-2275, email: sjwright@umich.edu
2
PhD student, Civil and Environmental Engineering, The University of Michigan,
email: jimlewis@umich.edu
3
Professor, University of Brasilia, Brazil, email: jvasconcelos@unb.br

ABSTRACT

A significant concern in the design of large, underground conduits for storm

or combined sewer systems is the avoidance of geyser formation, a phenomenon
which involves the explosive release of water through vertical shafts. Current
practice for transient analyses of these systems involves application of numerical
models to simulate the rapid filling processes. Since these models do not consider
interactions between the inflows and air being displaced during the filling process,
it is felt that these models cannot predict geyser formation.
Data from a large diameter stormwater tunnel in Minneapolis, Minnesota has
been analyzed to determine the events that led to a series of observed geyser events.
Data include pressure and velocity measurements within the tunnel that can be
correlated with a videotape of the geyser events themselves. The pressure records
are inconsistent with hydrostatic pressure lifting water to the ground surface.
Features of the pressure records can be interpreted to indicate the release of large air
pockets through the dropshaft. A qualitative comparison of these results with
smaller scale laboratory experiments indicates similarities in the pressure responses.
These results suggest that the entrapment of large air pockets is an important
component to the geysering process and that tunnel design procedures need to
properly account for these effects.

INTRODUCTION

The phenomenon commonly referred to as “geysering” that occurs in below grade

sewer systems is manifest as an explosive release of water through vertical
ventilation shafts or manholes. Figure 1 presents images of two such events, one in
Novosibirsk, Siberia and the other in Minneapolis, Minnesota. These clips are parts
of larger images and based on other features such as nearby buildings, the geyser
rises at least 20 m above the land surface. Such occurrences following the
construction of the Chicago Tunnel and Reservoir Project (TARP) in the early
1980s initiated interest in development of analysis methods so that geyser formation
World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5381

a.) Novosibirsk, Siberia b.) Minneapolis, Minnesota

Figure 1. Images of geysers formed in underground sewer systems.

could be predicted during design and modifications implemented to avoid such

problems. Numerical models of transient mixed flow where both free surface and
pressurized flow can occur simultaneously in nearly horizontal conduits were
developed for this purpose with the model by Cardle and Song (1988) being a
notable early contribution to the literature on the topic. Additional investigations in
more recent years include Capart, et al (1997), Politano, et al (2007), and
Vasconcelos, et al (2006). These represent various approaches to the problem of
predicting the simultaneous occurrence of both free surface and pressurized flow
with sufficient numerical accuracy to predict the hydraulic bores that may form in a
rapidly filling conduit.
In spite of the numerical sophistication associated with the implementation of
these models, it is not clear that any of them are relevant to the issue of geyser
formation. To date, there is very little discussion in the literature on the physical
processes involved in geyser formation. The geyser images in Figure 1 could be
interpreted as a mixture of air and water, raising the possibility of air-water
interactions. Small scale laboratory experiments reported in Vasconcelos (2005)
indicated that the most likely explanation for geyser formation was the escape of
discrete air pockets through vertical shafts that were in a pressurized state at the
time of air arrival such that water was forced upwards ahead of the rising air pocket.
Vasconcelos and Wright (2006) describe a number of different ways in which
trapped air pockets can form in a rapidly filling pipeline, even with very simple
geometries. Wright, et al (2008) suggest that a common situation for trapping large
air pockets involves the reflection of a free surface bore at a conduit transition that
results in the filling of the pipe at the transition with air behind the bore becoming
pressurized. However, the small scale of the experiments from which these
observations were made does not permit generalization to the much larger prototype
World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5382

scale of large natural systems. Other studies that discuss the phenomenon of
geysering are curiously quiet on this issue. For example, Guo (1989) and Guo and
Song (1990) discuss geyser formation but then present simple hydraulic analyses
that don’t consider air interactions. In Guo and Song (1991), the following
discussion is presented “…If the water level rises above ground surface, the geyser
occurs. It has been ascertained that if the dropshaft is ventilated, as most are, the
cover could not be blown off by air pressure alone. That is, most blowoffs are
caused by the impact force of the rising water. Therefore, it is sufficient to study
the hydrodynamics alone.” Although it is true that adequate ventilation will prevent
buildup of air pressures sufficient to displace manhole covers, a simple calculation
will show that air pressures as small as 3 kPa can be sufficient to lift a typical
manhole cover on the order of 0.8 m diameter. The first writer was involved in an
application where providing additional ventilation at an appropriate location
eliminated a recurring problem with manhole cover displacements in a 1.8 m
diameter combined sewer system. However, the above statement does not address
the fundamental question of whether geyser formation involves an air-water
interaction. This question is of critical importance to design issues since a
numerical model that requires the hydraulic grade line to rise to the ground surface
in order for a geyser to occur would entirely miss an occurrence that is associated
with an air-water interaction. Fortunately, data has been collected on one large
scale system that allows this issue to be addressed.

SYSTEM DESCRIPTION

The image of the Minneapolis geyser depicted in Figure 1b was extracted

from a video recorded as part of a data collection effort due to a recurrence of
geyser formation at two dropshafts in a stormwater collection system that flowed
beneath the median of Interstate 35W in Minneapolis, Minnesota. The video
recorder was installed at a dropshaft near the 35th Street overpass and was
accompanied by the installation of velocity and pressure transducers. Records
spanned the time from at least 1997 through 2005 with 13 events captured between
1999 and 2005 (and additional ones prior to 1999). Individual events often
involved more than one discrete geyser event. Data is presented below for a rainfall
event that occurred on the morning of July 11, 2004 and involved nine individual
geyser events that formed and subsided at intervals of approximately two minutes.
A video record of two geyser events is provided on the Minneapolis Star Tribune
website at www.startribune.com/video/15509842.html/. It is noted that the writers
possess other video records from this same installation that depict even stronger
geysers.
The general details of the system are as follows:

• Stormwater tunnel diameter of 3.66 m;

• Dropshaft diameter of 2.44 m;
• Tunnel invert located 28.6 m below grade;
• Pressure transducers installed at 0.47 and 2.88 m above the local tunnel
invert. The data acquisition system was programmed to record pressures only when
World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5383

water depths in the tunnel exceeded 1.46 m. Data was collected every five seconds
for water depths (as determined by the velocity meter) between 1.46 and 2.93 m and
every second for indicated depths greater than 2.93 m;
• Velocities were recorded at five minute intervals with an American Sigma
area-velocity meter.

RESULTS

The particular data presented below were collected on July 11, 2004 at
approximately 5:30 AM. Nine independent geysers were observed in the video
record of the event; there is some variability in strength as some of the geysers were
not explosive, each geyser lasted for about 10-25 seconds with about 75-90 seconds
between separating the onset of each one. The velocity record indicates that the
velocity was relatively constant at about one meter per second between about 5:30
and 8:50 AM with no indication of fluctuations that might be indicative of inertial
oscillations in the pipeline. Figures 2 and 3 present the pressure and velocity
records, respectively, for a time period spanning the occurrence of the geyser
events. The pressure record is from the lower pressure transducer and converted to
a pressure head relative to the tunnel invert. Superimposed on the figures are the
visual observations of the geyser occurrences indicated in the video record. This is
indicated by the vertical rise and fall in the line; the height of the line has no
relationship to the pressures indicated on the vertical axis.

5
Pressure Head (m.)

0
5:28:19 AM 5:38:19 AM 5:48:19 AM
Time
Pressure Geyser Observation

Figure 2. Pressure head relative to tunnel invert in

tunnel recorded during geyser events.
World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5384

Velocity (m/s)

0
5:28:19 AM 5:38:19 AM 5:48:19 AM
Time
Velocity Geyser Observation

Figure 3 Velocity recorded during geyser events.

A number of observations can be made from the results presented in Figures 2

and 3. First of all, the pressure head never went above about 6 m during the entire
event which is only nominally greater than the tunnel diameter of 3.66 m. Since the
tunnel invert is 28.6 m below grade, the discussion in Guo and Song (1991)
discussed in the Introduction would exclude the possibility of geyser formation as
the hydraulic grade line never approaches the land surface. This clearly indicates
that some other mechanism must be invoked to explain the observed geysers.
Another interesting observation is related to the observed geyser height. If it is
assumed that the geyser height is approximately 20 m from the video record, the
exit velocity of a water jet can be estimated by the simple conversion of velocity to
elevation head as V2/(2g) = h with h the rise height of the water jet. Using h = 20 m
would result in an exit velocity of 19.8 m/s, suggesting an upward discharge in the
dropshaft of 93 m3/s. This compares to a tunnel discharge of approximately 9 m3/s
during the geyser events according to the velocities presented in Figure 3. Thus,
there is an insufficient source of water to produce a water jet rising that high in the
air. Finally neither the pressure nor the velocity records in Figures 2 or 3 suggest
any indication of inertial oscillations that would be necessary to produce the
discrete geyser events according to the explanation suggested by Guo and Song
(1990). Although there does appear to be a velocity oscillation in Figure 3 during
the time that the geyser events occur, a careful inspection of the timing indicates
that the velocity is continuously declining during the geysering interval and then
increases again once the events cease. This is more indicative of additional flow
resistance during that time period such as partial flow blockage due to air pocket
World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5385

migration upstream against the water flow. Therefore, there is a total lack of
evidence to support the concept that geysers are created by water column rise in a
vertical shaft to above grade. A different explanation must be invoked to explain
the geyser formation in these particular events. While the data is only applicable to
the particular physical setting and the particular event, it is felt that the observations
can be more generally applied to understand geyser formation.
A more careful inspection of the pressure records provides additional insight.
Assuming that the tunnel becomes pressurized when the pressure height exceeds the
diameter, this occurs about six minutes prior to the commencement of the first
geyser event. The last geyser event finished about two minutes before the system
transferred back to a free surface state. Figure 4 is an expansion of the pressure
record to examine the pressure variation during the sequence of geyser events. It is
seen that each geyser lasted for about 10-25 seconds with about 75-90 seconds
separating the onset of the next one. The pressure tends to rise gradually between
geyser events and then drop abruptly at the onset of the geyser. This behavior is
inconsistent with a “hydraulic” or water flow only explanation of geyser formation
but would be consistent with the arrival of discrete air pockets at the dropshaft with
a pressure drop as the air releases into the vertical shaft. Although this is not
conclusive proof, the data are far more consistent with an air-water interaction,
specifically with the expulsion of discrete air pockets through vertical shafts than
other explanations considered. It is also noted that the apparent regularity of the
geyser events displayed in Figure 4 is not necessarily consistent with observations
during other events and the time intervals between individual occurrences is also
inconsistent between events. The writers believe that the geometry of the particular
system results in the propagation of discrete air pockets from the downstream
direction, but are awaiting further investigation to confirm this hypothesis.

LABORATORY OBSERVATIONS
An adaptation of the experimental setup described by Vasconcelos (2005) was
used to measure pressures during laboratory events that were interpreted as geysers.
Figure 5 indicates the experimental setup. A 0.038 m diameter, 2.5 m tall
ventilation shaft was installed on the top of the horizontal pipeline at about its
midpoint. Experiments were performed with the pipe initially partially full of
stagnant water. The two-way valve indicated in Figure 5 was opened in two steps.
The first step produced a free surface hydraulic bore that propagated to the
downstream end and reflected off the closed end of the pipe, with a higher free
surface bore propagating back towards the inlet. An additional partial valve
opening increased the upstream flow rate and produced a pressurized condition with
a significant amount of air trapped between the two fronts. The timing of the
second inflow increase was selected to result in the two bores interacting near the
ventilation shaft so that water rose a few cm into the shaft prior to the arrival of the
discrete entrapped air pockets. The flow was switched off once the water level
began to rise in the inflow shaft so that the pressure head within the pipeline never
exceeded about 1.2 m. Regardless of this, water was expelled out the top of the 2.5
m high ventilation shaft at a height far above the hydrostatic pressure level. Each
arriving air pocket lifted the water successively higher in the shaft. Although there
World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5386

was no attempt to reproduce the unknown inflow conditions associated with the
field event described above, the height of the laboratory ventilation shaft, when
scaled by the pipe diameter, is greater than the 28.6 m for the Minneapolis
stormwater tunnel and it appears reasonable to conclude that air can indeed lift
water to the heights observed in the video.

7
Pressure Head (m.)

4
5:35:00 AM 5:36:00 AM 5:37:00 AM 5:38:00 AM 5:39:00 AM
Time
Lower Probe Geyser Observation

Figure 4. Output from lower pressure probe during geyser events.

Figure 5. Schematic of laboratory apparatus.

World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5387

Two piezo-resistive pressure transducers (Endevco model 8510B-1) were

connected at the bottom of the pipe at distances of 6.45 m and 11.70 m downstream
of the center of the inflow pipe. Outputs from these transducers were sent to a data
acquisition system (National Instruments DAQPad MIO-16XE-50) at a frequency
of 10 Hz. A typical pressure record from the transducer located upstream of the
ventilation shaft is presented in Figure 6. The effect of the air pockets rising in the
shaft is seen as sudden drops in the overall trend of the pressure variation. This
feature is similar to the behavior indicated in Figure 4.

1.4

1.2

1
Pipe Crown
Unvented
Pressure (m)

0.8

0.6

0.4

0.2

0
0 10 20 30 40 50 60
Time (seconds)

Figure 6. Pressure in pipeline upstream from ventilation shaft during

expulsion of entrapped air pockets.

DISCUSSION AND CONCLUSIONS

The field measurements collected during a rainfall event in Minneapolis,
Minnesota in 2004 included visual observations of the dropshaft at the ground level
along with velocity and pressure measurements within the stormwater tunnel
immediately upstream from the dropshaft. The most plausible explanation for the
observed geyser formation is the interaction of trapped air with water initially
standing in the drop shaft due to the existence of surcharge conditions. The
measured pressures within the pipeline were incapable of lifting that water even
close to the ground surface, let alone eject it 20 m into the air. Converting the
estimated rise height of the geyser into a vertical velocity in the dropshaft requires a
discharge an order of magnitude than actually measured to be flowing within the
tunnel at the time of geyser formation. Furthermore, the measured pressures and
World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5388

velocities do not support the presence of large scale inertial oscillations within the
tunnel such as might be required to lift the water to the ground surface in a series of
short duration events. The patterns of water pressure drop during air expulsion
through the ventilation shaft are similar in the field and laboratory measurements,
suggesting that in spite of possible limitations due to the small laboratory scale,
there are similarities between the mechanisms for water rise in the vertical shaft in
both situations.
There are several implications associated with these findings. The process of
geyser formation is apparently not directly connected with surging in the tunnel
system as suggested by the previous literature on the subject. Numerical models
that are currently applied to simulate transients in rapidly filling tunnel systems do
not account for the air phase. The predicted results of these models should be
interpreted with caution with respect to the issue of geyser formation. Predictions
that transient hydraulic grade lines remain below grade should not be interpreted to
suggest that geyser formation will not occur. Model capability to predict the
location of air entrapment within a system is useful, even if the subsequent motion
of the air cannot be predicted with a single phase flow model. This information can
be used judiciously to make design decisions about the location and capacity of air
ventilation required in a system.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the cooperation of Dr. Christopher

Ellis (originally with the St. Anthony Falls Hydraulics Laboratory) and Bruce Irish
of the Minnesota Department of Transportation in providing the field data used in
this manuscript.

REFERENCES

Capart, H., Sillen, X., and Zech, Y. 1997. Numerical and Experimental Water
Transients in Sewer Pipes. J. Hydr. Res. 35 (5), 659-670.
Cardle, J.A. and Song, C.S.S. 1988. Mathematical Modeling of Unsteady
Flow in Storm Sewers. Int. J. Eng. Fluid Mechanics, 1(4), 495-518.
Guo, Q. 1989. Geysering in Urban Storm Drainage Systems. Proc. XXIII
IAHR Cong. Ottawa, Canada S75-S81.
Guo, Q., and Song, C.S.S. 1990. Surging in Urban Storm Drainage Systems. J.
Hydraulic Engineering, V. 116(12), 1523-1537.
Guo, Q., and Song, C.S.S. 1991. Dropshaft Hydrodynamics Under Transient
Conditions. J. Hydraulic Engineering, V. 117(8), 1042-1055.
Politano, M., Odgaard, A.J., and Klecan, W. 2007. Case Study: Numerical
Evaluation of Hydraulic Transients in a Combined Sewer Overflow Tunnel System.
J. Hyd. Eng. 133 (10) 1103-1110.
Vasconcelos, J.G. 2005. Dynamic Approach to the Description of Flow
Regime Transition in Stormwater Systems. PhD dissertation, Environmental
Engineering, The University of Michigan, Ann Arbor, Michigan.
World Environmental and Water Resources Congress 2009: Great Rivers © 2009 ASCE 5389

Vasconcelos, J.G., Wright, S.J. and Roe, P.L. 2006. Improved Simulation of
Flow Regime Transition in Sewers: Two-component pressure approach. J.
Hydraulic Engineering, 132(6), 553-562.
Vasconcelos J.G. and Wright, S.J. 2006. Mechanisms for Air Pocket
Entrapment in Stormwater Storage Yunnels. Proceedings of World Water and
Environmental Resources Congress, Omaha, Nebraska, Paper 40856-14275.
S.J. Wright, C.T. Creech, J.M. Lewis and J. G. Vasconcelos, “Mechanisms of
Flow Regime Transition in Rapidly Filling Stormwater Storage Tunnels,” Journal
of Environmental Fluid Mechanics, 2008, 8, 605-616.

View publication stats