Mitigating Geysers in Deep Sewer Tunnels

Sewage drop shafts function as a powerful air moving devices by creating water-in-air waterfalls that move air like fans or air-in water fluid columns that transport and compress quantities of air when water velocities are greater than the bubble rise velocity. This creates situations where air in unvented tunnels can transfer pressure at the sewer obvert to the surface at shafts that, even when vented can lead to situations where air and water forcefully exits manholes at the surface and damages surrounding infrastructure (Figure 1). This damage can be severe in large wet weather events. One such event was in St. Louis Missouri on August 30, 2021, which caused a precast drainage structure top to separate from the main structure (referred to herein as the Ikea shaft, the result being visible in Figure 2 below). Another depressurization at Arkansas and McKean caused damage to the pavement near a drop shaft and is generally a nuisance to residents. For the Ikea shaft, SWMM models for the area were analyzed to estimate hydraulic flow through the sewers during cloudburst events. This accompanied creation of an inventory of all drop shafts associated with the tunnel of interest using record drawings. Model results were superimposed onto the inventory to determine how flow enters the deep tunnels, the quantity of flow at each inlet and outlet, and identify manholes where no flow enters the sewer from above. This review identified that the Ikea shaft was the first (in the direction of flow) that did not have flow from above that could otherwise prevent air pressure at the sewer obvert transferring to the surface, and that over half the drop capacity of the deep sewer system is upstream of the Ikea shaft.A previous evaluation of the structure had been performed. This investigation concluded that the drop shaft experienced a rise in the hydraulic grade line during wet weather, which caused the overflow events. As a result, a grate vent (visible in Figure 3) was installed on the structure. This helped mitigate issues, but a rising hydraulic grade line did not explain ongoing cover slab displacements. It was determined that the investigation had not directly considered the deep sewers upstream of the Ikea shaft, vent locations, and steady-state phenomena like air flow and entrainment effects. In response to these observations, one-dimensional model of the deep sewer system connected to the Ikea shaft was to estimate maximum entrained air flow rates based the geometry of the deep sewers and connected shafts (long section shown in Figure 4). To estimate maximum entrained air flow rates a pressurization model was developed by doing the following: - Derived simplified pressure loss/gain relationships for the sewers and drop shafts - Assuming bubbly or misty flow with air and sewage velocities equal. - Allowing for the compression/expansion of air along/down sewers/drops and the resulting change in mixture velocity along/down sewer/drops and the resulting impact on pressure loss/gain. - Based on conventional Colebrook-White friction factors and average fluid properties. - Forwards-calculated sewer flow rates based on input sewage and air flow rates at each drop. - Back-calculated sewer obvert pressure from the outfall to the sewer obvert for each drop based on the forwards-calculated flow rates, with iteration to allow adjustment and resolution of the resulting mixture fluid properties. - Used the derived pressure loss/gain relationship for the drops to calculate an air-sewage mixture column depth/height that would give the back-calculated sewer obvert pressure. - Used Newton's method to adjust the input air flow rates to give an air-sewage mixture column surface at drops below the incoming sewer invert by the distance necessary for the sewage to free fall to the mixture velocity at the column surface (as indicated in Figure 5). - With minimal sewage flows in drops downstream of the Ikea shaft, varied upstream sewage flow rates to indicate entrained air flow rates that could supplied to the Ikea shaft and the pressure at which they could be supplied (plotted in Figure 6 for three scenarios due to uncertainty of the exact location of the Ikea shaft in the network). A simplified model was selected for the evaluation because higher air entrainment rates are not considered plausible as there would be no driving force for air entrainment. The driving force for air entrainment is taken to be the deceleration of free-falling sewage entering the air-sewage mixture column. Maximum possible air entrainment is conservatively taken to be the air entrainment rate that gives an air-sewage mixture surface level at which there is no longer any free-falling sewage deceleration energy to drive air entrainment. Physically achievable air entrainment rates will be lower due to some energy being required to drive entrainment and overcome bubble rise velocities. This allowed for a flexible model that did not have high computational costs.↵ The model calculated that the maximum entrained air flow rates increase with obvert pressure before peaking and then declining with increasing obvert pressure, which is shown in Figure 6. These pressures were calculated to be well above the lifting pressure of cover slabs. Adequate venting of entrained air flow would be required to prevent high pressures developing under cover slabs. The analysis concluded that the historical depressurization incidents were a result of high entrained air flow rates and obvert pressures forcibly venting at the Ikea shaft. This location forcibly vents because the site lacks sufficient sewage inflow to prevents the obvert air pressure transferring to the surface. Being the first such shaft (in the direction of flow), it likely protected other downstream shafts without inflow that have a pressure retaining design. Other downstream drop shafts could also have a large vent grate or are much further downstream where obvert pressures are much lower and would therefore be less likely to violently depressurize. The maximum entrained air flow rates for the Ikea shaft indicated an order of magnitude increase in venting area was required. To mitigate this issue, GHD designed the new shaft cap in Figure 7 for the Ikea shaft with more than an order-of-magnitude increase in venting area, increased weight and some pressure balancing that reduces the net lifting force. This design enables a two orders-of-magnitude increase in the air flow rate that can be vented when compared with the previous cover slab (Figure 3). This was a relatively inexpensive modification that did not require complex models or deep energy dissipation structures to design and install. The methodology could be applied to similar sewers to assist with venting analysis and design without the need of more complex and computationally intensive modeling techniques. Modifications are currently in design and are expected to be installed prior to the conference.