The WSSC is the largest water and wastewater utility in the Washington, D.C. metropolitan area, providing services to 1.6 million residents of Montgomery and Prince George’s counties. It has a service area of nearly 1000 square miles (2590km2) including four reservoirs, two water filtration plants, six wastewater treatment plants, more than 5,300 miles (8530km) of potable water pipelines and over 5,300 miles of sewers.
The sewer system model was created to enable the WSSC to meet the requirements of a Sanitary Sewer Overflow Consent Decree and estimate the current and long-term capacity of the system and its offline storage facility.
The decree was issued to ensure the Commission met its Clean Water Act commitments for discharges into receiving waters, and specified that the organization should use a computer model of the collection system to identify portions that might not have sufficient capacity to accommodate present or anticipated future flows, plan sewer improvements, and make decisions about future development of the system.
Rock Creek Basin is one of 23 sanitary sewer basins modeled in InfoWorks CS by CDM and sub-consultants McKissack & McKissack and Environ-Civil Engineering as part of the modeling study required by the Consent Decree.
A 1985 agreement limits peak water transmission flows from the Rock Creek basin into the District of Columbia Water and Sewer Authority (DC-WASA) sewers to 56.5 million gallons per day (214 million litres per day) per day at the Washington, D.C. boundary. Because of this, the Rock Creek basin includes an offline storage facility so that the WSSC can keep flows to the required levels.
The basin itself is extremely complex – it has a drainage area of around 41,725 acres (16,700ha) and is divided into 199 mini-basins each averaging 210 acres (85ha) in area. Around 709 miles (1141km) of sewer exist within its boundaries, ranging in size from 4 inches (100mm) to 54 inches (1.37m) in diameter. Almost half of the sewers in the basin were constructed prior to 1950, with the oldest sewers concentrated in the southern part of the basin.
The Rock Creek storage facility was completed in 1991 and contains five storage basins with a total available storage volume of six million gallons (22,700m3).
During storms, a remotely-operated diversion structure channels the flows from a 54-inch diameter trunk sewer into two wet wells, where the flow is pumped by six constant speed pumps into the five storage basins. The existing pumping capacity allowed for a maximum flow diversion of 24 MGD (91,000m3/d).
The stored flows are gradually fed back into the trunk sewer, when flows recede sufficiently following a storm. Control of the facility is manual, from the WSSC central control center. This control is complicated by the fact that the storage facility is over seven miles from the point where the peak flows need to be controlled, and a significant portion of the Rock Creek basin flows enter the system between the storage facility and the point where they are discharged back into the DC-WASA system – the main reason why an InfoWorks model is needed. The current approach is to divert flows to the storage facility when flows at the boundary line reach 42 MGD (159MLD).
The Consent Decree required all pipes of 10 in (250mm) diameter and larger to be modeled, as well as selected 6 in (150mm) and 8 in (200mm) pipes. This resulted in 113 miles (182km) of modeled sewer in 2,422 pipe segments as well as models of the stormwater facility, two pumping facilities and 3.4 miles (5.47km) of force mains.
Manhole and pipe data for the model was obtained from the WSSC GIS and mapped to the Infoworks CS model data fields. The InfoWorks CS solution automatically identified missing and inaccurate data. Construction record drawings were used where available to reflect the sewer system as accurate as possible. In a few places, data had to be inferred to complete the model.
The two wastewater pumping stations, the Reddy Branch and North Branch stations, were included to represent the motive force for flows from the upstream 10in and larger diameter sewers to the downstream gravity sewers.
A simplified procedure was used to model these relatively small elements of the system, based on the estimated firm capacity, maximum capacity, number and type of pumps, (whether constant speed, two-speed or variable-speed), wet well dimensions and the actual wet well operating levels.
The pumped flow rates, which are controlled automatically in the model based on simulated wet well water level depths and real pump on-off configurations were compared to the observed flow data and estimated station capacity.
The Rock Creek storage facility system – including the diversion structure, influent gravity sewer, wet wells, pumps and storage basins – were input to the InfoWorks CS model using data from the original construction drawings. Control algorithms were built into the model to accurately represent the facility’s operating procedures.
Although this procedure provides an approximate representation of the general operating conditions, as discussed with WSSC staff, the operators also have some flexibility to operate depending on the system conditions and estimates of incoming rainfall from rainfall radar data.
Determining dry weather flows and I&I
Analysis determined the existing dry weather flows at the system’s 23 permanent wastewater flow meters, and at two Rockville-WSSC interconnection billing flow meters. The data for the parallel meters was analyzed as one by summing the flows to provide the total flow amount from all upstream areas.
Additional flow data was analyzed to calibrate and validate the model of the Rock Creek Storage Facility’s operating procedures during wet weather events. Flow monitoring data used in the model was provided in 15-minute increments over a four-year time span. Further review of rainfall and groundwater conditions found that this was a representative period in terms of rainfall and dry weather patterns.
Meter errors and inconsistencies in flow data were identified, and results were found to be less accurate partly where meters were installed in series. Large values were being subtracted from large values to gauge relatively small incremental values for the additional area between two meters. Overall, the average dry weather flow data for the system during the monitoring period ranged between 586 and 2,300 GPD (2,218 and 8,706 litres/day).
The flow monitoring data was also used to determine the rainfall-dependent inflow and infiltration at each of the permanent flow meters. CDM’s SHAPE program was used to identify 20 to 30 storm events and separate out I&I flows from the total flows observed for each event.
The fraction of the rainfall that enters the sanitary system upstream from the meters (the R-value) was calculated for each event by dividing these I&I flows by the rainfall volumes. Higher R-values suggest areas with a greater number of sewer defects, and sewer system inspection and repairs can be prioritized based on this ranking.
The InfoWorks CS model allows flows generated within an area to be applied to a single manhole. To enable flows to be more accurately applied to multiple manholes and ensure that all gravity sewer sections being modeled have flows, the Rock Creek mini-basins were further divided into 406 micro-basins with an average area around 103 acres (41.7ha).
The mini and micro-basins generally contain all of the areas that could, potentially, be served by WSSC’s sewer system. The boundaries may contain large areas such as golf courses and parking lots that make no contribution to the system. Significant areas of Rock Creek have no sewers, particularly the northern area where land use planning limits development.
To identify and exclude large unsewered areas within the micro-basins the GIS system was used to undertake an initial screening. Property parcels in proximity to pipes of 8in diameter or less were selected as sewered. Larger sewers were not used in this initial screening process so that areas of undeveloped land where main sewers pass through would not be counted. The parcel selection was refined using digital orthography and other mapping together with the WSSC sewer system and property parcel GIS layers. The process defined 32% of the Rock Creek Basin area as sewered.
The estimation allowed wastewater flows to be assigned more accurately to individual micro-basins, so flows were distributed more accurately within the modeled network.
Calibrating the model
The InfoWorks model was calibrated using data from the 25 flow meters and the two billing meters, but not the meters at the Washington, D.C. boundary line, as these did not have a complete 15-minute data record.
The model was calibrated using criteria defined in the Wastewater Planning Users Group (WaPUG) Code of Practice for hydraulic modeling of sewer systems.
Dry weather flow criteria were that the simulated flow volume and peak flow rate should be within 10% of those observed, and the shape and magnitude of the simulated and observed hydrographs should be similar.
The wet-weather calibration criteria were:
- The volumes of the simulated flow and simulated peak flow rate should be within 10% of those observed for two of the three events and should be no more than 20% different for any event.
- Greater emphasis should be placed on matching peak flows rather than volumes as the model is to be used to evaluate sewer system capacity.
- The depth of simulated flow in surcharged areas should be between +18 inches and -4 inches of those observed.
- The unsurcharged simulated depth of flow at key points (such as near flow splits) should be within +/- 4 inches.
- The timing of the peaks and troughs of the hydrographs should be similar over the event.
The wet weather flow calibration storm events were chosen carefully to identify those where their R-value was reasonably close to the average weighted R-values for each meter. The selection emphasized storms with larger rainfall volumes and peak intensities, as these more closely matched the two-year design storm event used to evaluate the system capacity and storage facility, and were therefore better suited to this calibration.
Once the model was calibrated to two storm events, a third rainfall event was simulated and compared against the observed data as an additional verification step. For all analyzed meters and storms, the average difference between simulated and observed peak flows was 3.3% and the average volume difference was 4.6%.
The calibrated and verified InfoWorks CS model was used to simulate flows in the sewer system and storage facility of the Rock Creek Basin under conditions specified in the Consent Decree. Modeled scenarios included baseline dry weather flows and the two-year 24-hour Soil Conservation Service Type II storm for 2005 population conditions.
The model showed that most portions of the sewer system that had been mapped in the Infoworks CS solution model had enough capacity to carry dry weather flows and had a reasonable allowance for rainfall-induced peak flows without using the storage facility. The lower reaches of the trunk sewers from the storage facility to the D.C. boundary line had enough dry weather flow capacity but were less able to convey additional wet-weather flows.
The model scenario with the two-year design storm applied the current storage facility operating procedures (the facility is utilized when flows reach 42 MGD at the Washington, D.C. boundary line, and fills as long as downstream flows are over 35 MGD (132.5MLD) or total storage volume is not reached).
For this scenario, the model results predicted that for the chosen design storm the storage facility would operate for almost seven hours and would fill to 90% of its capacity. Using the storage facility reduces the maximum flow at the Washington, D.C. line to 53.4 MGD (202MLD), 3 MGD (11.4MLD) below the peak discharge limit. The facility also shaves peak flows in the downstream parallel trunk sewers and controls surcharging within acceptable limits.
The model also identified segments of the system within four categories of state of surcharge:
- where simulated maximum depth of flow is less than 80% of pipe diameter, meaning there is more than enough capacity for the simulated peak wet-weather flows;
- where simulated maximum depth is over 80% but less than 100% of pipe diameter, meaning the pipe is approaching capacity but is not surcharging at the simulated peak flow;
- where the pipe is full or surcharged during simulated peak flows due to peak flows exceeding the full-flow capacity in downstream sewers, causing backup although the segment has sufficient capacity to carry the peak flow;
- where the peak flow exceeds the full-flow pipe capacity, resulting in surcharging, which is acceptable as long as the hydraulic grade line elevation does not rise so much that it results in an overflow or user backup.
The results proved the storage facility was essential to ensuring the network met the boundary limits during the design storm. It also showed that if another large storm occurred before the storage facility had fully drained from a storm, there could be a potential violation of the limit. To limit the chances of this happening, storage facility operations were evaluated in the model to find an appropriate operating range to ensure violations do not occur and to minimize the volume stored.
From a number of model runs, refined operating procedures were developed, under which the storage facility is activated when flows reach a new operational limit of 46 MGD (174MLD) rather than 42 MGD at the Washington, D.C. line. In addition, flow storage stops as soon as the flow rate begins to fall off in both the trunk sewers immediately upstream of the facility and where the Washington, D.C. line starts.
In this scenario, for the two-year design storm, the facility fills for 3.5 hours to just 56% capacity, and does not exceed the flow limits. Closing the diversion gate as soon as flows begin to slacken would allow the storage facility to keep an extra two million gallons in available storage.
The Rock Creek sanitary sewer system model’s main use is as a planning tool for future improvements and development of the sewer system as required by the Consent Decree. The results of the model’s use for storage facility operation improvements show that the proposed operating procedures are effective in reducing peak flows at the basin boundary to below the agreed limits while retaining two million gallons of additional storage volume compared to the existing operation procedures.
The consultants intend to verify the proposed operating procedures for observed storm events and additional design storm events with different distribution types and return periods, and refine the storage facility drainage procedures during back-to-back wet weather events before implementing the proposed operating procedures.