Predictive Performance Method for Hydrodynamic Separators Using the Peclet Number
Hydrodynamic separation technology is commonly used for the removal of suspended sediment from stormwater runoff. Laboratory performance tests of hydrodynamic separator (HDS) systems rely on a number of controlled parameters such as flow rates, particle size distribution, sediment concentrations, sediment specific gravity and water temperature. Considerable challenges exist when making direct performance evaluation comparisons between different HDS systems, especially when different test sediment particle sizes were used. This presentation explores a predictive performance method using the Peclet Number that allows for a comparison between laboratory-based HDS performance curves and predicted performance curves for different particle sizes. An independently verified HDS performance test using a sediment median (d50) size of 110 microns is compared to a series of predictive performance curves for sediment d50 particle sizes of 45, 50, 67, 90 and 125 microns for the same device. These performance curves are further compared to both known and predicted performance curves derived from a separate HDS system design and test verification. These performance curve comparisons illustrate a shift in the device’s performance expectations depending on sediment size.
The unitless Peclet Number (Pe) is expressed as Pe = (D*H*VS) / Q, where the test parameters of Flow Rate (Q in cfs), Horizontal Flow Dimension (D in feet), Vertical Flow Dimension (H in feet) and Particle Settling Velocity (VS in feet/sec) are considered. This performance evaluation method applies to both vortex- and vault-type HDS designs. The D value for a vortex- or vault-type device is based on their respective diameter or long axis of the effective treatment area. The H value of both types is based on the depth of water within the effective treatment area. Particle settling velocity is a critical aspect of this performance evaluation method since it accounts for variables including the selected particle size diameter, specific gravity and water temperature. A spreadsheet has been developed to calculate VS by three methods including Stoke’s Law and those proposed by Cheng and Ferguson & Church. This method is based on Stoke’s Law due to its wider recognition, a test sediment specific gravity of 2.65 and water temperature of 20°C.
For the 110 µm HDS performance test, the Q, D and H values are known for the tested device and the VS value is calculated. Pe values are then calculated for each tested flow rate. Each Q value is converted to surface area loading rate expressed in gallons per minute per square foot (gpm/ft2). To predict Q (and loading rate) values for each of the calculated Pe and associated removal efficiency values for a different d50 particle size, the equation is rearranged to Q = (D*H*VS) / Pe, and the VS value is recalculated for the selected particle size. While keeping both the calculated Pe and removal efficiency values constant, a predictive performance curve that plots removal efficiency versus loading rate is then generated for the selected d50 particle size. These performance curves can also be used to predict a peak loading rate for a specific target sediment removal efficiency goal, such as 50% or 80% on a per storm event basis. Sizing charts are also used to scale larger or smaller HDS models from that of the tested model to provide the same level of performance as demonstrated from the performance test. A method is also described to generate HDS sizing charts for the different particle sizes using the peak loading rates derived from the predicted performance curves.