Lake Erie is the fourth largest of North America’s Great Lakes and home to the magnificent Niagara Falls, which drain Lake Erie into Lake Ontario. The area and its ecosystem have long been threatened by pollution, however. The Environmental Protection Agency is concerned with the effects of combined sewer overflows (CSOs). When sewer capacity is overwhelmed during heavy rainfall, CSOs allow excess rainwater mixed with raw sewage to be discharged into nearby water bodies – in this case, into Lake Erie itself. The agency now requires 98% of discharge to be captured and kept from entering the lake.
To meet this target, regions bounding the lake are investing heavily in systems designed to address the capacity limitations of existing sewers – hence the US$3 billion programme being implemented by Northeast Ohio Regional Sewer District.
Euclid Creek tunnel – 5.5km long, 7.3m in diameter and at an average depth of 61m – will have a storage capacity of 17,200cu m. It runs adjacent to and partly beneath Lake Erie. Water and sewage collected near the surface by sewers is transported to the storage tunnels, built deep underground, through structures called drop shafts. While gravity does the hard work, the structures must control the speed and energy of the liquid as it descends and remove trapped air that could impede flow in the tunnel.
Two types of drop structure had been used locally for decades and were considered suitable for the Euclid Creek tunnel: vortex and baffle drops. Our team liked the idea of using baffle drops as they are simple, easy to build and cost-effective. However, the Euclid system needed to handle flows four times that of existing structures – and there was no data available to predict if this was possible.
Vortex structures have been modelled to understand how they work. In a ‘tangential inlet’ vortex drop, water is carried along an approach structure, a chute that heads downhill and narrows steadily until the flow is discharged through a slot at the end. The accelerating liquid rushes through this and is directed at an angle onto the inside wall of the drop structure to induce a whirlpool effect, causing the flow to spiral around the wall, all the way to the bottom.
But there are significant disadvantages to this type of drop. Multiple inlets at varying elevations would disrupt the spiral effect and so can’t be incorporated. Air becomes trapped in the flow at the base of the drop. If not removed, this air would be carried into the storage tunnel itself, reducing capacity and potentially causing odour problems. To tackle this, the water has to pass through a deaeration chamber before entering the tunnel.
For Euclid Creek, the approach structure would have been several storeys below ground at its entry point to the drop shaft, with a slot 6m high. The deaeration chamber would have been around 25m long and 5m high – and 60m underground. Building one of each for every shaft would have been difficult and expensive.
Space saving solution
Baffle drop structures require neither approach structures nor deaeration chambers. They work by routing the liquid through a series of alternating stepped platforms. The flow pools on each platform, absorbing the energy of the liquid falling from above. As the flow doesn’t need to be routed around the circumference of the shaft, the structure can be divided vertically to provide a dry side for worker access – another addition on a vortex construction programme.
Baffle drops had been used for more than 90 years locally, but only for flows of around 3.3cu m/s. For Euclid the drops needed to handle 14.9cu m/s. The most accurate way to simulate complex flows such as these was to build a physical model, but this had yet to be done. There was no scientific industry information to turn to.
World-first hydraulic research
The solution, which was approved and funded by our client, was to partner with the Iowa Institute of Hydraulics Research to build a scale model of the baffle structure we had designed, and feed the data gleaned back into the design process. Over the years, a standardised configuration of baffles had been developed, which worked in practice but which had no explanatory methodology. Our goal was to understand exactly how water behaves as it moves through such a structure. Once we understood precisely how it acts, we could extract the data to accurately forecast how the Euclid Creek structures would perform – and if they would convey the flows.
The studies yielded proof that baffle structures could incorporate multiple inlets without affecting flow, and demonstrated that our design would be able to cope with the expected flows. The modelling threw up some surprising results too. It was well known that baffle structures required no deaeration chambers – but not why. When we piped coloured smoke through our model, we could see that air was being stripped from the water and passing through apertures in the dividing wall. The flow leaving the model was mostly devoid of air bubbles, unlike the maelstrom of water leaving the base of a vortex drop.
Removing the need for deaeration chambers and approach structures saved significant cost. And because baffle drops can accommodate multiple flow inlets, fewer shafts could be specified compared to using vortex drops. All told, using baffle drops has saved our client 38% cost per structure.
Tenfold capacity increase
Our research is an industry-first, and is being applied on major projects in Canada and New Zealand, as well as elsewhere in the US on structures with more than double the capacity of the Euclid Creek drops. We have also found that, by offsetting the dividing wall to decrease the size of the worker access, we can increase flow capacity without increasing shaft diameter. Our work has demonstrated that these structures can handle flow up to 33cu m/s – ten times what’s been common until now.