COMPLIANCE INFO_2020 - Laserfiche WebLink

Home Browse Search

SFPUC June 10, 2020 Page 7 of only 8-16 minutes at an average concentration of about 0.3 ug/L. Thus, sampling at 15-minute increments could largely miss an instantaneous release at the Tesla Facility. When a UV lamp breaks, water is spilled into the UV Building from the UV reactor until an automatic shutoff of flow occurs. The current program for operating a UV reactor has the downstream valve fully opening before closing. Once the reactor is isolated, water in the reactor above the suspect broken lamp casing's elevation drains onto the floor. The remaining water in the reactor is drained directly to the trench. The water in the closed-off pipe section is then also discharged into the trench drain. The total volume of water entering the drain is approximately 6,000 gallons through the broken lamp orifices (based on the hydraulic conditions, which depend on the throughflow rate at the time of breakage and the time it takes to close the flow valve to the pipeline) plus another 5,000 gallons drained from the reactor pipeline section after its valves are fully closed. The measured spill volume contained in two rented baker tanks in October 2019 confirmed the estimate of approximately 11,000 gallons per event. All discharged water is contained within the UV Building and eventually flows by gravity into the central, grated trench drain. The visible portion of the concrete trench extends the length of the UV Building. The floor drain is approximately 180-ft long, 4-feet wide and greater than 3- feet deep. Near its downstream end, the trench drain has a shallow 4-inch overflow weir. The downstream end of the trench drain (another—40 feet) has a sump and connects to a 42-inch emergency overflow culvert. If filled with discharged water(11,000 gallons) the length of the trench would be 2-feet deep with discharged water. Figure 3 portrays a schematic of the on-site flow of water and mercury associated with a UV lamp break. The following points describe and quantify (to the extent feasible, based on the available monitoring data) the fate of mercury potentially discharged from a UV reactor because of a UV lamp break. • Based on sampling at Hetch Hetchy Reservoir(the source water) in 2009 and in a raw water tap in the Tesla UV Building and analytical laboratory testing results acquired by this project team on October 11, 2019, the raw water is assumed to contain< 0.0008 ug/L total Hg. Assuming the raw water has no mercury(as it was not detected at 0.0008 ug/L total Hg), a UV lamp's mass dissolved into the discharged water would result in a concentration of 55 ug/L (2.3 grams of mercury into 11,000 gallons). However, water samples collected from the trench during a break/discharge event averaged about 2.6 ug/L. Thus, about 5% of the lamp mercury is accounted for in the discharged water samples. • The methods used to sample and analyze mercury likely biased the results high, meaning that using plastic bottles, sampling without using clean sampling methods, and analyzing using high detection-limit methods likely portray higher concentrations than actually existed. Thus, the previously quantified mercury concentration that is accounted for in the discharge water probably over-estimates the actual concentration. • If the vaporized mercury remains in vapor form and is released uniformly into the UV Building's ambient air, that air would have a mercury concentration of 240 ug/m3. However, a Jerome J405 mercury vapor analyzer employed at various times in summer 2019 measured 8 samples as non-detected(<0.5 ug/m3), one at 0.62 ug/m3 and one at 7.71 ug/m3. Even air in the office space and outside measures 1-8 ug/m3. Thus, <1% of a UV lamp mercury is accounted for in the UV Building's air monitoring data.