Flow field and sediment removal in a stormwater sump utilizing internal structures

Urban stormwater runoff is one of the major sources of urban water pollution, comprising particulate matters of varying sizes, heavy metals, and organic matters (Aryal & Lee 2009). These sedimentary particles are the primary carriers of pollutants and pathogens and their direct discharge to receiving water can lead to varying degrees of contamination. Catchbasin sumps are an essential component of stormwater collection networks, which have proven to be effective in retaining floatables, capturing large debris, and reducing the buildup of deposition in sewers (Lager et al. 1977; Hvitved-Jacobsen et al. 2010). Underground stormwater treatment devices were also commonly used to reduce suspended solids (Roseen et al. 2005; Fassman 2006). In comparison to end-of-pipe treatment devices, stormwater sumps occupy a smaller land footprint, which is highly desirable in densely populated urban areas (Field et al. 2003; Erickson et al. 2013). In addition, sumps can be used in series to achieve multi-stage solids control. Therefore, maximizing the capability of existing sumps to remove sediment is an intriguing area of interest, such as incorporating simple inserts and the like (Aronson et al. 1983; Brar et al. 2016). An optimal modified stormwater sump must not only enhance its particulate retention efficiency but also minimize its impact on drainage capacity and its maintenance cost (Pitt & Field 2004).

Numerous studies on the standard sump, by laboratory testing, field monitoring, and/or computational fluid dynamics modeling, reveal that the configuration and size, inflow rate, and particle settling velocity are all important factors in determining the sediment removal efficiency of sumps (Wilson et al. 2007; Pathapati & Sansalone 2009; Howard et al. 2012; Kaushal et al. 2012; McIntire et al. 2012). Nondimensional parameters were found to be useful to analyze the overall functionality and efficiency of sumps, including the Hazen number, Froude number, and Péclet number (Wilson et al. 2009; Guo 2017; Yang et al. 2022). Howard et al. (2012) evaluated the removal efficiency of standard stormwater sumps of varying sizes and found that the sumps can effectively remove suspended sediment from stormwater runoff, although not as effectively as dedicated hydrodynamic separators. The relationships between the removal efficiency and system parameters were identified by an integral approach in earlier studies, and there is still a lack of quantified information and in-depth analysis on hydraulic characteristics of the flow field in the sump that are crucial for optimizing and novel designs.

A number of sump inserts and/or internal structures were studied to improve the management of wet-weather flow solids (Pitt & Field 2004; Andoh et al. 2007; Brar et al. 2016). Various sump inserts were found to be useful for gross solids removal, although considerable variability has been reported in the results (Lau et al. 2001; Walch et al. 2004; Alam et al. 2018; Basham et al. 2019). Filter fabrics were used with special care for capturing fine particulates, as they get clogged so quickly that the maintenance cost becomes an issue (Pitt & Field 2004). When dealing with a large flow rate, catchbasin inserts are inclined to cause flow choking and scouring contamination (Walch et al. 2004; Brar et al. 2016). A hooded outlet design was recommended as it can withstand extreme flows with little scouring losses in Pitt & Field (2004). A few internal structure add-ons were tested and found to provide a good balance between solids control and hydraulic performance (Howard et al. 2011; Ma & Zhu 2014). However, most of such designs lack a well-defined physical foundation and are challenging to continuously optimize using established metrics in order to attain better performance. This difficulty is due, in part, to the absence of a determined functional relationship between the flow features and solids movement, especially for sumps.

The present study aimed to investigate the flow characteristics of a sump with and without modifications and to provide insight into the solids retention efficiency as a function of various physical characteristics, including particle size, settling velocity, flow rate, and sump geometry. In laboratory testing, a particle image velocimetry (PIV) device was employed to quantify and analyze the flow field. Additionally, the study explored scaling relationships for predicting the performance of stormwater sumps of different sizes.

It was discovered that over 50% of particles had a diameter greater than 125 μm in German & Svensson (2002). Additionally, Sansalone et al. (2010) conducted statistical surveys and found that roughly 10% of particles were less than 100 μm in diameter, while particles with diameters between 100 and 400 μm constituted approximately 25–60% of the distribution. Therefore, for the experiments in this study, non-cohesive quartz sand particles with a density of 2,650 kg/m³ were selected within a size range of 100–150 μm, with a mean value of 125 μm. The discharge range associated with free surface conditions is Q = 0.1–0.3 L/s, while the discharge range related to full pipe flow conditions is Q = 0.4–0.5 L/s. To conduct removal efficiency testing, sediment was continuously fed into the entrance pipe in the form of slurry, with a sand feeding rate in the range 36–286 mg/s (Table 1). The sediment concentration of the incoming flow was in the range of 72–2,860 mg/L. The quantity of sediment particles fed into the pipeline was pre-determined by weight. The captured solids in the sump were collected, dried by a drying chamber (DHG-9240A Heating and Drying Oven), and then weighed by a high precision electronic scale (QUINTIX35-1CN). Therefore, the removal efficiency, conventionally defined as the ratio of the sediment particles retained in the sump to the total introduced, was assessed using a mass-based method.

A camera (SONY FDR-AX60) with a high resolution of 3,840 × 2,160 was used to record the flow field at 50 frames per second and the two-dimensional PIV technique was applied to obtain the velocity field in the sump. The particles with a diameter of 10 μm and density of 1,010 kg/m³ were chosen as PIV tracers. The particle size and density ensured minimal interference with the flow while allowing for reliable tracking and measurement of flow velocities. Following the approach of Beg et al. (2018), to minimize the error caused by refraction through the curved manhole wall, the sump was put within a square acrylic tank filled with water, in front of which the camera lens was positioned. The laser was directed from the top of the sump as a laser sheet with a thickness of approximately 5 mm. To ensure the relative stability of the water flow, we conducted continuous operation for approximately 20 min. During this period, the stability of the system was assessed by measuring the stability of the liquid level of the sump. To ensure the reproducibility of the experiment, each operating condition was replicated at least three times.

Eddy diffusion is a crucial parameter for assessing particle retention efficiency, which can be directly related to the fluctuating component of the velocity (Siegel et al. 1990). It follows that the eddy diffusion can be represented by the turbulent intensity or the total kinematic energy (TKE). In Figure 4(c), we have provided the values of calculated turbulent kinetic energy (TKE) corresponding to the different test cases. The data in this figure clearly demonstrate a noticeable trend: as the discharge rate increases, there is a corresponding rise in the TKE values. This observation implies that the intensity of eddy diffusion also increases as the discharge rate is raised.

However, to satisfy the request for further discussion and explanation, we need to delve deeper into the implications of this relationship. The connection between TKE and eddy diffusion is pivotal in understanding the mechanisms at play here. Eddy diffusion refers to the dispersion caused by turbulent eddies in a fluid flow. When the TKE increases, it signifies a higher level of turbulence within the flow field. This heightened turbulence leads to greater mixing and diffusion of particles or substances present in the fluid.

The relationship between TKE and eddy diffusion can be attributed to the fact that higher TKE values indicate a more energetic flow, with stronger and more frequent eddies. These turbulent eddies contribute significantly to particle transport and dispersion, influencing factors such as particle retention and mixing efficiency. Therefore, the observed correlation between increasing discharge, rising TKE, and enhanced eddy diffusion highlights the sensitivity of particle behavior to fluid dynamics.

In conclusion, while the presented data clearly illustrate the connection between rising TKE and increasing discharge, a more comprehensive analysis reveals that this relationship is intricately tied to the concept of eddy diffusion. The heightened TKE values signify a more turbulent flow, resulting in amplified eddy diffusion, which in turn has notable implications for particle retention efficiency and mixing patterns.

Scale effects refer to the disparities between experimental outcomes and real-world scenarios attributed to differences in scale. When investigating sediment removal efficiency in a stormwater sump, acknowledging and addressing these scale effects is crucial. Therefore, it is imperative to scrutinize these effects thoroughly to gain a deeper understanding of their influence on sediment removal efficiency.

It can be observed in Figure 12(a) and 12(b) that when the model is scaled up by 1.5 times, the predicted values agree well with the experimental results in Howard et al. (2012). When the model is scaled up by six times (to the standard size), there are notable differences at low flow rates, and the discrepancies reduce as the efficiencies both approach zero at high flow rates, as expected. For instance, at a flow rate of 50 L/s, the predicted retention rate is about 10% higher than the actual value, and this gap will further decrease as the flow rate increases.

Figure 12(c) and 12(d) illustrate the changes in sediment retention rates obtained by upscaling the improved model. Following sump enhancements, the sediment retention rate for a sump with a diameter of 120 cm increased from 20 to 50% under a flow rate of 50 L/s, marking an efficiency improvement of approximately 30%. It is evident that the particle removal efficiency of stormwater sumps with proper internal structures increases significantly.

To enhance the retention efficiency of stormwater sumps, two general approaches have been explored in previous studies. Howard et al. (2011) achieved promising results by reducing the inlet flow rate through the incorporation of Saint Anthony Falls Laboratory (SAFL) baffles, which proved particularly effective during low flow rates. On the other hand, Ma & Zhu (2014) adopted a different strategy by partitioning the rectangular sump into multiple channels, thereby increasing the particle retention rate by extending the residence time. Their comparative analysis demonstrated an efficiency improvement of approximately 10%. In our experiments, an innovative device was introduced that applies the principle of inclined plate sedimentation to further enhance retention efficiency. By reducing the settling distance, this approach yielded a substantial increase in the retention rate, measuring over 30% for a large flow rate range. These findings underscore the significance of exploring various retention-enhancing techniques, each offering unique advantages in stormwater sump design. As future research endeavors continue, it is essential to evaluate the feasibility and potential synergies of integrating multiple techniques to achieve even higher retention rates and foster sustainable urban water management practices.

The data from previous studies, such as Saddoris et al. (2010), Howard et al. (2011), and McIntire et al. (2012), suggest an empirical relationship that exists between particulate removal efficiency and Pe/Fr² in conventional sumps. The current study further confirms this relationship holds even in optimized sumps, with only necessary corrections for the dominant dimensionless parameters. The addition of internal structures substantially alters certain parameters related to particulate removal efficiency. Once the parameters are uniformly described for both conventional and optimized sumps, the empirical formulas become universal and can guide sump optimization.

The small scale of the physical model used here necessitated careful comparison to previous larger-scale experimental data to quantify scale effects. This analysis delineates the application scope of conclusions from the current study. Accounting for scale dependencies enables first-order estimation of particulate removal efficiency in full-scale systems based on empirical relationships.

The findings of this study, especially the relationship equation between Pe_m/Fr² and sediment removal rate after modification, offer data-based guidance for the design and improvement of stormwater sumps. It should be noted that, in practical applications, as the bottom space fills up, there is a risk of captured solids causing clogging or being washed out. Further investigations are warranted to optimize the volume allocated for sediment storage.

This paper presents hydraulic characteristics of conventional and improved storm sumps and examines the corresponding solids removal efficiency. The following conclusions can be drawn:

• (1)

  The adoption of the enhanced model, which incorporates methodologies such as lengthening the water flow path, attenuating flow velocity, and facilitating particle adhesion, yields a substantial increase in the particle removal efficiency of stormwater wells. The observed efficiency improvement achieved with the upgraded model is approximately in the range of 20–30%.

• (2)

  The results demonstrate that using the modified Péclet number (Pe_m) provides a better estimation of particle retention rates. The addition of internal structures can be considered as effectively modifying Pe_m. Overall, the modified Peclet number enables unified empirical modeling of particle retention rates across sump configurations by encapsulating the effects of internal structures on the effective settling parameters.

• (3)

  When the scale ratio (prototype size : model size) is small, particle removal efficiency can be modeled with reasonably good accuracy. However, for scale ratios greater than 6, scale effects become significant and should be considered in the modeling to properly represent particle removal efficiency.

This investigation was supported by the Ningbo Young Technology Innovation Leading Talent Program (2023QL028), Natural Science Foundation of China (52009087), the Natural Science Foundation of Zhejiang Province (LGF19E090005, LY21E090003) and the key Research and Development Program of Zhejiang Province (2020C03082).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.