Limitations of Gravity-Based Separators
Professionals in the storm water industry are familiar with the use of gravity separation systems for the removal of total suspended solids (TSS). There are several different configurations available to specifying engineers, including an expanding number of proprietary configurations, all of which rely on the same basic principle.
Reducing the velocity of the influent flow allows sediment being transported by the water column to drop out. To some extent, the effectiveness of these systems is determined by the dimension and shape of the internal components.
The primary factors determining effectiveness, however, are particle size distribution and percent by mass of incoming sediment. In other words, a higher portion of very fine sediment will mean lower removal efficiencies for specified storm water treatment systems.
This may seem obvious, but when reading information provided by different manufacturers of storm water treatment systems, it is important to analyze the data side by side. Unfortunately, until recently, many regulations regarding storm water treatment did not specify what gradation of sediment manufacturers should remove. Regulations were notoriously vague; for example, they might state that storm water treatment systems should remove 80% of TSS. Manufacturers were then able to take advantage of regulations that were vague regarding the variability of particle size distribution in storm water and its effect on the removal efficiency of various treatment systems.
A classic example would be the approval of two different storm water treatment systems, one of which is capable of removing 80% of TSS with a d50 of 50 microns, and one which is capable of removing 80% of TSS with a d50 of 125 microns. Without seeing the data side by side, it is easy to overlook the fact that one system is capable of removing significantly smaller particles than the other while still obtaining the same 80% reduction in TSS. Expressed differently, if a sediment gradation with a d50 of 50 microns was fed into both systems, only one of them would obtain an 80% TSS removal efficiency while the other would achieve something much lower (the actual removal efficiency of which could only be determined by analyzing the percent by mass of different particles within the sediment gradation).
Particle Size & Settling Velocity
Before analyzing the efficiency of two different systems, it is important to understand how variations in particle size affect settling velocity. After all, the removal efficiency of a storm water treatment system is dependent on particles settling to the bottom of the system before water leaves the structure. To better understand this process, imagine a rectangular vault as displayed in Figure 1.
As water enters the structure, it has to travel a horizontal distance of 8 ft to the outlet pipe. If there is no turbulence to take into consideration, determining whether a particle is removed would be a simple matter of calculating how far the particle falls within that 8-ft horizontal distance. Sediment that does not fall below the invert of the outlet pipe will be carried downstream in the effluent; sediment that drops below the invert of the outlet pipe will accumulate on the vault floor. If the inlet and outlet pipes are both 12 in. in diameter, then sediment being transported near the crown of the inlet pipe has to drop a minimum of 12 in. to prevent being carried downstream.
Let’s perform a calculation to determine whether or not we capture a 125-micron particle. First, we need to know the residence time of water within the vault. Assuming that water is displaced uniformly as it travels through the structure, we can divide the volume of the vault by the inlet flow rate to determine the residence time in the structure. Based on the dimensions of the vault, we know that the volume of the sump is 128 cu ft. If the inlet flow rate is 1 cu ft per second, then the residence time of water and any associated sediment would be 128 seconds. In order to calculate the settling velocity of a particle, Stokes’ law for laminar flows may be used: 2
In Equation 1, Vs equals the settling velocity of the particle, g is the gravitational acceleration, ps is the density of the particle, p is the density of the fluid, d is the diameter of the particle and Mu is the viscosity of the fluid. The calculation’s results show that a 125-micron particle has a settling velocity of approximately 0.55 in. per second. With a residence time of 128 seconds, our particle will fall approximately 70 in., significantly more than the required 12 in.
Now let’s examine the 50-micron particles discussed previously. Using the first equation, a 50-micron particle has a settling velocity of approximately 0.09 in. per second. This translates to approximately 11.5 in. of drop in 128 seconds. At first glance, one might say this is close enough to 12 in. and that capture of a 50-micron particle seems plausible, however, a number of variables have been eliminated from the calculations, most importantly the effects of turbulence. Turbulence will have a significant impact on settling velocity and will tend to keep particles suspended for a longer period of time.
TSS Removal Efficiency
During the summer and fall of 2004, the University of Minnesota’s St. Anthony Falls Laboratory (SAFL) conducted testing on a BaySaver separator system to determine its TSS removal efficiency. The result of this testing was an equation that utilizes a dimension-less coefficient called Peclet, which represents the relationship between settling and turbulence. The Peclet number for each system is unique and based on the dimensions of the separator unit and the primary and storage structures. The Peclet number is defined as: 1
In Equation 2, Vs is the settling velocity of a particle, L1 is length in feet and D is a turbulence diffusion coefficient (sq ft/sec). With this equation, as the length or width of the structure increases, the Peclet number increases as well. As was explained earlier, larger structures in general will have a higher TSS removal efficiency than smaller ones; therefore, it should be intuitive that as the Peclet number for a particular system increases, so does its efficiency. The diffusion coefficient also plays a significant role in that as this number increases, the Peclet number and therefore removal efficiency of the system will decrease. This can be calculated using the following equation: 1
In Equation 3, coeff is a dimension-less coefficient, U is the flow velocity in the structure, L' is the depth of the structure, L2 is the width of the structure and Q is the flow through the structure. Referring back to the original example, the hypothetical vault with influent particles 50 microns in diameter has a Peclet number of 0.132. The same vault structure with 125-micron particles has a Peclet number of 0.828. Although not accurate when applied to the hypothetical situation, the following equation will at least give an idea of how turbulence within a structure will affect the removal efficiency of a specific size particle: 1
If we enter the Peclet numbers for the 50- and 125-micron particles into the fourth equation, we come up with a removal efficiency of 0% and 28.4%, respectively. This example illustrates the significant impact that turbulence has on the settling velocity of a particle.
There are many different configurations available to engineers and regulators, all of which have unique characteristics designed to minimize the amount of turbulence within the structure and maximize the removal efficiency of the system. Despite this fact, it is evident from our original example (in which the effects of turbulence were completely eliminated) that there are still limitations to the removal efficiency of a gravity-based system. More importantly, there are very realistic constraints imposed by the effects of turbulence within a structure.
This is not to say that these systems do not serve a purpose. They have proven effective as pretreatment devices designed to remove coarse sediment and floatable contaminants from storm water runoff. They can also significantly reduce the frequency and cost of maintenance on downstream structures, and until now, they have allowed storm water treatment to exist in highly urbanized environments where natural storm water treatment structures may not fit. Tightening regulations, however, have required a shift toward the use of filtration technologies such as the BayFilter. These systems are capable of removing much finer sediment, and recent competition in the market has brought the cost of these structures down significantly.
This black box example has served to illustrate the realistic constraints of gravity-based storm water treatment devices. As regulations across the country become more stringent, the nature of their use will hopefully shift from stand-alone treatment systems to pretreatment structures for more specialized filtration technologies.