Question 4 of 5: Coagulation Mechanisms; Water Distribution Systems

Question 4: Coagulation Mechanisms; Water Distribution Systems (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) — Three destabilisation mechanisms (12 marks)

Colloidal particles in a natural water — clay, silica, humic matter, bacteria — carry a negative surface charge and are held apart by electrostatic repulsion that overwhelms the van der Waals attraction between them. They are stable in the thermodynamic sense: Brownian collisions occur constantly, but the energy barrier prevents attachment, so the suspension never aggregates and the particles never settle. Coagulation is the deliberate destruction of that barrier, and it is achieved by three physically distinct routes, which in a real jar test operate together and in different proportions depending on dose, pH and turbidity.

Ionic (double) layer compression. Around each negatively charged particle the water organises itself into a Stern layer of tightly bound counter-ions and a diffuse layer in which the counter-ion excess decays with distance. The thickness of that diffuse layer is the Debye length, which depends only on the ionic strength of the bulk solution:

Adding an indifferent electrolyte raises I and compresses the diffuse layer, so the repulsive potential decays over a shorter distance and the energy barrier falls. A soft surface water of ionic strength 0.001 M carries a 9.6 nm double layer; raising the ionic strength to 0.01 M shrinks it to 3.0 nm and to 0.1 M shrinks it to below 1 nm, at which point van der Waals attraction dominates at every separation and the suspension coagulates without any charge neutralisation at all. Because I depends on the square of the ionic charge, counter-ion valence is decisive: the Schulze–Hardy rule gives critical coagulation concentrations in the ratio $1 : 1/2^6 : 1/3^6$, that is $1 : 1/64 : 1/729$ for mono-, di- and trivalent counter-ions. This is why aluminium and ferric salts, not sodium salts, are the coagulants of practice — and it is also why sea water is naturally coagulating and why an estuary deposits its sediment load where fresh water meets salt.

Adsorption and charge neutralisation, leading to sweep coagulation. When alum or ferric chloride is added at normal pH, the metal ion hydrolyses through a series of positively charged, polynuclear hydroxo species which adsorb specifically onto the negative particle surface and neutralise its charge; the zeta potential moves toward zero and the particles attach on contact. Overdosing reverses the charge and restabilises the suspension — the reason the jar-test curve has a minimum rather than a plateau. If the dose is pushed further, beyond the solubility limit of the metal hydroxide (roughly 20–60 mg/L of alum at pH 6–7.5), amorphous Al(OH)₃(s) precipitates throughout the volume as a voluminous, low-density floc, and the colloids are physically enmeshed and swept out as that precipitate settles. This is sweep coagulation or sweep floc. It is remarkably robust — it works on low-turbidity waters that have too few particles to collide with one another, it is insensitive to modest overdosing, and it removes natural organic matter by adsorption onto the fresh hydroxide surface — but it consumes far more coagulant than charge neutralisation and produces a much larger volume of sludge.

Inter-particle bridging. A long-chain, high-molecular-weight polymer — a polyacrylamide polyelectrolyte, or a natural polymer such as starch — adsorbs at a few segments onto one particle while the rest of the chain extends into solution as loops and tails, which then adsorb onto a second particle. The two are physically bridged by the polymer, and the resulting floc is far stronger and more shear-resistant than one held by charge neutralisation alone. Bridging is not primarily electrostatic: an anionic polymer can bridge negative particles even though both carry the same sign, provided divalent cations act as anchor points. It has two characteristic failure modes that must be named. An overdose saturates every available surface site so that no chain finds a vacant site on a second particle, and the suspension restabilises — steric stabilisation. And excessive shear in the rapid-mix stage folds the extended chain flat onto the particle it first touched, destroying the bridge before it forms; this is why polymer is dosed as a flocculant aid after the rapid mix, into a gently stirred flocculation basin at a velocity gradient G of 20–70 s⁻¹, and not into the 700–1000 s⁻¹ of the flash mixer.

Part (b) — Requirements of a distribution system; grid-iron versus dead-end (13 marks)

Key requirements. An adequate distribution system must deliver, simultaneously and continuously:

• Adequate quantity at adequate pressure. Sufficient capacity for maximum-hour domestic demand, and separately for maximum-day demand plus fire flow, which usually governs the pipe sizing in residential areas. Normal service pressure is 275–415 kPa (40–60 psi), with a code minimum of 140 kPa (20 psi, about 14 m of head) maintained at every point during the fire-flow event, and a practical upper limit near 700 kPa to protect fixtures.
• Preservation of water quality. A detectable disinfectant residual everywhere, water age short enough to prevent nitrification and DBP growth, no cross-connections, minimum velocities sufficient to avoid stagnation, and mains no larger than needed — oversizing for fire flow is the most common cause of poor water quality in a new subdivision.
• Reliability and redundancy. Any single main can be taken out of service for repair without cutting off supply; storage is distributed so that a pump failure or a power outage does not immediately depressurise the system.
• Operability. Isolation valves close enough together that a break shuts down few services (typically no more than one block, three valves per tee), hydrants spaced 75–150 m in residential areas, blow-offs at low points and air valves at high points.
• Durability and integrity. Materials, bedding and cover appropriate to the ground and the frost depth; corrosion protection; minimal leakage, with real losses held under about 10 per cent.
• Economy. Minimum whole-life cost of pipe, pumping energy and maintenance, and the capacity to be extended as the service area grows.

The hydraulic difference, quantified. One calculation explains almost every entry in the comparison table. Given. A 250 mm main, Hazen–Williams C = 130, serving a demand of 40 L/s at a point 600 m from the source. Find. The head loss when that point is fed from one direction and when it is fed from two.

Fed from one end, the whole 40 L/s travels the whole 600 m and $h_f = 1.72$ m. Closed into a loop, the same demand is fed from both directions, so each leg carries about 20 L/s over about 300 m, giving $h_f = 0.24$ m. The ratio is exactly $0.5\times0.5^{1.852} = 0.139$:

Grid-iron (looped) system. Mains are laid on the street pattern and cross-connected at every intersection so that flow can reach any node from several directions. Advantages: no dead ends, so no stagnation and a uniform disinfectant residual; a break or a shutdown isolates one block instead of a whole branch; head loss and pressure fluctuation are much lower, as the calculation shows, so smaller pipes can carry the same fire flow; fire flow can be drawn from several mains at once; the system extends easily in any direction. Disadvantages: substantially more pipe, more valves and more fittings, so a higher capital cost; the flow distribution is statically indeterminate and must be solved iteratively (Hardy Cross or a network model), so design and calibration take more effort; more valves means more maintenance and a greater chance that a valve is left in the wrong position.

Dead-end (branched or tree) system. A trunk main with sub-mains and branches, each terminating. Advantages: the least pipe for a given coverage and therefore the lowest capital cost; flow direction and magnitude in every pipe are known by inspection, so design is simple and can be done by hand; it suits the ribbon development of rural roads, and valve counts are low. Disadvantages: every branch is a single point of failure — one break cuts off everyone downstream; water stagnates at each terminal, so the residual decays, sediment and biofilm accumulate, and taste, odour and coliform complaints cluster at dead ends, which then require a routine flushing programme that wastes water; the whole demand of a branch travels its full length, so head loss is high and pressures at the extremities are poor, particularly during a fire; and fire flow at any point can be supplied from one direction only.

Practice. Canadian municipal design (MMCD and the AWWA distribution standards) requires looping wherever it is practicable, and permits dead ends only for short temporary extensions or genuinely isolated services, where a flushing hydrant or automatic flushing device is then mandated at the terminus.