Water Journal : Water Journal July 2012
refereed paper UV disinfection water JULY 2012 65 power consumption and sleeve function. Recommended activities that most of the utilities don't perform regularly include verifying individual lamp intensity, checking quartz sleeves for excessive fouling due to iron or hardness, visually inspecting ballast fans, and assessing the operation and performance of the automatic cleaning system. Although all the utilities use some type of supervisory control and data acquisition system to continuously monitor flow, UVT, UV dose and other critical system parameters, monthly reporting to local regulators varies widely. Some utilities are required to submit a 35-page monthly compliance report summarising system performance, UV sensor checks, UVT analyser checks and off-specification events. Other utilities are not required to submit such reports to their local regulator. Other site-specific operational issues experienced by water utilities include excessive air entrainment in the UV reactors, causing reactor shutdown; excessive ballast failures, causing frequent UV reactor shutdown; stuck wiper systems, causing lamp breaks; and electrical wire melting. Most of these problems have since been resolved. For unfiltered systems, excessive lamp fouling from seasonally increased iron requires additional sleeve cleaning for a few weeks each year. An automated cleaning system usually cleans the sleeves effectively. Most LPHO lamp systems do not have automated cleaning systems, but typically experience less fouling than MP lamps due to the lower lamp temperature. All LPHO will require periodic manual cleaning of the lamp sleeves. Figure 2. Broken UV lamp sleeve (top) and solid mercury in lamp sleeve (bottom). Because all lamps currently in use contain mercury, water utilities must address online and offline lamp breaks. Most of the participating utilities experienced less than one online lamp break every five years of operation. A few utilities have observed one or more lamp breaks per year for several years, possibly due to a cleaning system malfunction or old lamp sleeves. When online lamps break, emergency operating procedures should be followed, usually consisting of isolating the UV train, draining contaminated water and treating prior to disposal to the sanitary sewer, disposing of the collected glass and mercury according to mercury-disposal guidelines, and collecting water samples downstream to ensure compliance with the MCL for mercury. Off-line lamp breaks may occur while handling UV lamps. Mercury clean-up kits are used to collect elemental mercury and broken glass. All of the participating utilities have plans to address the issue of both on-line and off-line lamp breaks. Figure 2 shows an example of a broken lamp sleeve and mercury settled in a lamp sleeve. Operating Costs for UV Systems The utilities are frequently unable to sufficiently reduce lamp power to match the required or operating UV dose during normal flow and water quality conditions. This problem was more pronounced for smaller utilities that use only a single reactor during normal operating conditions. Larger systems could minimise power use by operating fewer UV reactors. Power turndown of a UV reactor was driven by lamp turndown capabilities, which ranged from 30% to 60%. The UV reactors couldn't turn off individual lamps or banks of lamps within the reactors. Therefore, the UV reactors commonly dosed at two to three times the required dose for extended periods of time, which consumes additional energy. Since the UV reactors were pre-validated, the turndown capacity is dictated by its validation testing approach and the worst-case design criteria of the specific installation (see Figure 3). LPHO systems typically have greater turndown capability and lower overall power consumption than MP systems, and therefore may be able to provide equal disinfection with lower power costs. Although not specifically addressed in the UVDGM, operating costs can vary widely but are generally much lower than for other process units, such as high service pumping. The primary costs associated with operating a UV system include power consumption, lamp replacement, expendable parts replacement, and UV sensor and UVT analyser calibration. Power consumption costs are driven by water quality, disinfection target (RED) and local energy costs. Annual operating costs for the participating utilities were converted to UV Dose (mJ/cm2) Flow Rate Target UV Dose Maximum Power Level Minimum Power Level "Overdosing" Region = Wasted Energy Worst-Case Operating Condition Figure 3. Schematic of turndown limitations for UV reactor. 0.0 1.0 2.0 3.0 4.0 0% 30% 100% Chlorine Residual (mg/L) UV Reactor Power Setting Chlorine Decay through UV System Before UV Figure 4. Example of chlorine decay across a UV reactor due to high UV doses.
Water Journal August 2012
Water Journal May 2012