Question 4 of 5: Filtration and UV Disinfection in Water Treatment

Question 4: Filtration and UV Disinfection in Water Treatment (25 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Part (a) — Schmutzdecke and backwash (8 marks). The two terms in this part belong to two different filters, and saying so is the substance of the answer rather than a quibble.

The schmutzdecke — German for "dirt cover" — is the biologically active mat of algae, bacteria, protozoa and retained particulate matter that develops on the surface of a slow sand filter over the first days to weeks of a run. It is the defining feature of slow sand filtration, and it does most of the work: it strains particles finer than the sand grains themselves could retain, its protozoa and rotifers actively predate bacteria, and its microbial community biodegrades dissolved organic carbon. A slow sand filter run at 0.1 to 0.4 m/h on fine sand ($d_{10} \approx 0.15$ to 0.35 mm) is therefore not merely a physical strainer but a biological reactor, and it needs a "ripening" period after cleaning before it reaches full performance. Its cleaning method matches its mechanism: the filter is drained and the top 20 to 25 mm are scraped off and washed or replaced, every one to three months. It is never backwashed.

A rapid sand filter works on the opposite principle. It runs at 5 to 15 m/h on coarser media ($d_{10} \approx 0.45$ to 1.2 mm, commonly dual-media anthracite over sand), and it depends on upstream coagulation and flocculation to destabilise the particles so that they can be captured within the depth of the bed rather than on its surface. Because floc penetrates the bed and is removed by attachment throughout it, no surface mat forms — and if one did, it would appear as premature headloss and a short run, not as a benefit. Backwash is the cleaning of such a filter by reversing the flow: treated water, sometimes preceded or accompanied by an air scour, is forced upwards through the underdrain at 40 to 50 m/h for 8 to 12 minutes, expanding the bed by 20 to 50 percent and fluidising it, so that the shear of the rising water and the collisions between grains detach the accumulated floc, which is carried out over the wash-water troughs to waste. A filter is backwashed when any one of three triggers is reached: terminal headloss (typically 2.4 to 3 m), turbidity breakthrough, or a maximum run time of 24 to 72 hours imposed to prevent floc from becoming attached too strongly to release.

Given. A 40 m² filter run at 10 m/h for 48 hours and backwashed at 45 m/h for 10 minutes. Find. The backwash volume as a fraction of production. Backwash uses $45 \times (10/60) \times 40 = 300$ m³, against production of $10 \times 48 \times 40 = 19\,200$ m³, so

$$\boxed{\frac{300}{19\,200} = 0.0156 \approx 1.6\ \text{percent of production}}$$

which sits in the usual 1 to 5 percent band and is the number that justifies backwash-water recovery. Two operating cautions complete the answer: the rate must be matched to water temperature, because cold water is more viscous and a rate set for summer will over-expand and carry media to waste in winter; and the first filtrate after a backwash carries a turbidity spike, which is why filter-to-waste or a slow-start ramp is used.

Part (b) — Design and operating principles of UV disinfection (9 marks). Ultraviolet disinfection inactivates micro-organisms photochemically rather than chemically. Photons near 254 nm are absorbed by adjacent pyrimidine bases in DNA and RNA, forming cyclobutane pyrimidine dimers that block transcription and replication; the organism is not lysed but is rendered unable to reproduce and therefore unable to infect. The governing design quantity is the dose, or UV fluence, the product of intensity and residence time:

$$D = I \times t \quad [\mathrm{mJ/cm^2}]$$

Given. A reactor delivering an average intensity of 12 mW/cm² with a residence time of 3.5 s. Find. The delivered dose. Substituting, $D = 12 \times 3.5 = 42$ mJ/cm², which meets the 40 mJ/cm² validated design dose adopted in North American practice for 4-log virus credit and comfortably exceeds the 10 to 12 mJ/cm² needed for 3-log Cryptosporidium and Giardia. That contrast is UV's decisive advantage: the chlorine-resistant protozoa that dominate Canadian surface-water risk are among the most UV-sensitive organisms known, while adenovirus is the outlier requiring roughly 186 mJ/cm².

Five design and operating principles follow. (1) Water quality governs delivered dose. UV transmittance at 254 nm over 1 cm, UVT, is the controlling water-quality parameter; the absorption coefficient is $a = -\log_{10}(\mathrm{UVT})$, so at UVT 88 percent, $a = 0.0555$ cm⁻¹ and Beer–Lambert gives a residual intensity of $10^{-0.0555 \times 4} = 0.60$ at a 4 cm path, whereas a poor water at UVT 65 percent leaves only 0.18 — more than a threefold penalty for the same lamp. Iron, humic colour, and turbidity all depress UVT, and particles additionally shield organisms embedded in them, which is why UV always follows filtration. (2) Hydraulics govern the dose distribution. Every element of water must receive the dose, so short-circuiting is the enemy; reactors are validated as complete assemblies by bioassay against a challenge organism, and the validated dose is always well below the calculated average. (3) Lamps and sleeves must be monitored and maintained. Low-pressure high-output lamps are near-monochromatic at 254 nm and efficient; medium-pressure lamps are polychromatic, more powerful per unit and better suited to large plants. Output falls with age to roughly 80 percent at end of life (9000 to 15 000 hours), and quartz sleeves foul with calcium, iron and manganese deposits, so both an end-of-lamp-life factor and a fouling factor are applied in design and sleeves are cleaned mechanically or chemically on line. (4) Control is by on-line dose monitoring: calibrated UV intensity sensors, a UVT analyser and the flow signal are combined to compute delivered dose in real time and to pace lamp power or the number of banks in service. (5) The absence of a residual is the defining limitation. UV leaves nothing behind in the distribution system, so a chlorine or chloramine residual must still be added downstream; against this, UV forms no halogenated by-products, needs no chemical storage or handling, and requires very short contact time. Photoreactivation and dark repair, by which some bacteria reverse dimer damage, are handled by designing at doses well above the minimum inactivation requirement.

Part (c) — Mechanisms of contaminant removal in filtration (8 marks). Granular filtration is conventionally analysed as two sequential processes: transport, which brings a particle from the bulk flow to the surface of a grain, and attachment, which decides whether it stays there.

The transport mechanisms are four. Straining retains particles larger than the pore openings between grains; it acts only at the bed surface, dominates in slow sand and cake filtration, and in a rapid filter causes headloss without useful depth removal. Interception captures a particle whose streamline passes within one particle radius of a grain surface, and is the dominant mechanism for particles above roughly 1 µm. Sedimentation captures particles dense enough for gravity to pull them across streamlines onto the upper surfaces of grains, and matters for the larger, denser floc. Diffusion — Brownian motion — dominates below about 1 µm, and grows more effective as particles get smaller and as the water gets warmer. Because interception and sedimentation improve with size while diffusion improves with smallness, single-collector efficiency passes through a minimum near 1 to 2 µm: the most difficult particles to filter are, unhelpfully, the same size as Cryptosporidium oocysts. Hydrodynamic action and inertia contribute at higher velocities but are minor in water filtration.

Transport is necessary but not sufficient, and attachment is where chemistry enters. Natural particles and the media grains both carry negative surface charge, so an intercepted particle is repelled by the electrical double layer and re-enters the flow. Coagulation with alum, ferric salts or a cationic polymer compresses the double layer and neutralises charge, allowing van der Waals attraction, charge neutralisation and polymer bridging to hold the particle to the grain. This is why a rapid sand filter is only as good as its coagulation: without adequate destabilisation, particles are transported to the grains and simply bounce off, and the filter passes turbidity while showing very little headloss. Adsorption of dissolved species — natural organic matter, taste-and-odour compounds — occurs on the accumulated floc and, in biologically active filters, biological oxidation removes assimilable organic carbon, ammonia, iron and manganese as well.

Detachment is the reverse process, and it governs the end of a filter run: as deposits accumulate, pore velocity and shear rise, previously captured floc is torn loose, and turbidity breaks through — usually before terminal headloss is reached in a well-designed dual-media bed. Media design follows directly from these mechanisms: coarse anthracite over fine sand puts the large, easily captured floc in the coarse upper layer and reserves the fine lower layer for polishing, so that the whole depth is used and the run ends on headloss and breakthrough together.