Power plants live or die by their water pre-treatment. The winning playbook pairs coagulation–flocculation with clarification and media filtration, tuned by jar tests to the “knee” of the dose curve.
Raw water feeding heat recovery steam generators (HRSGs, heat recovery steam generators) arrives loaded with colloidal clays, natural organic matter (NOM), and turbidity. The standard defense is conventional pretreatment — coagulation–flocculation, clarification, and media filtration — usually aiming for effluent turbidity below roughly 1–5 NTU (nephelometric turbidity units) to protect downstream membranes, demineralizers, or boilers (www.watertechnologies.com).
In coagulation, metal salts or polymers neutralize particle surface charge and form hydroxide precipitates that enmesh colloids; in flocculation, gentle mixing encourages larger aggregates (“flocs”) that settle (www.watertechnologies.com). Typical coagulants include alum (aluminum sulfate, Al₂(SO₄)₃), ferric chloride (FeCl₃), ferric sulfate, pre‑polymerized alum (PAC, polyaluminum chloride; ACH, aluminum chlorohydrate), and cationic polymeric coagulants/flocculants such as polyamines and PolyDADMAC (polydiallyldimethylammonium chloride). For example, at optimal conditions (pH about 6–7), 1 mg/L alum salts produce Al(OH)₃ precipitate that captures turbidity (water.mecc.edu; www.watertechnologies.com).
In tropical surface waters, NOM often complicates this chemistry. Sururi et al. (2023) report that with moderate river turbidities, lab‑made alum (“tawas”) at 20 mg/L and acidic pH achieved ≈85% turbidity removal, while similar studies sometimes saw ≤70% — a reminder that water chemistry (organics, alkalinity, temperature) can swing outcomes (ejurnal.itenas.ac.id; ejurnal.itenas.ac.id). Key drivers are coagulant type, dose, initial pH/alkalinity, and mixing intensity (www.researchgate.net; ejurnal.itenas.ac.id).
Coagulant chemistry and pH windows
Metal‑salt coagulants tend to depress pH, often necessitating lime or caustic supplementation; alum works best around pH 5.5–6.5 (water.mecc.edu), while ferric salts function across a broader range (approximately pH 5–11) (www.watertechnologies.com).
Pre‑hydrolyzed aluminum products and cationic polymers can coagulate at higher pH or without added alkalinity, lowering chemical demand (www.watertechnologies.com). Polymers do not add solids or acidify; they “bridge” particles and can reduce sludge volume by roughly 50–90% compared with alum (www.watertechnologies.com). In practice, waters with 10–60 NTU are often clarified most effectively by an inorganic coagulant plus a cationic polymer, whereas turbidities above 60 NTU can frequently be managed with a polymeric coagulant alone; very low turbidities (<10 NTU) sometimes require an inorganic “seed” to kick‑start floc formation (www.watertechnologies.com). Plants typically feed these reagents via accurate metering, where an in‑line dosing pump helps control coagulant and flocculant addition.
Because any given raw water suspension is non‑uniform, jar testing is essential to identify the best coagulant and flocculant for that source (www.watertechnologies.com; www.sugarprocesstech.com). Plants sourcing these chemistries often standardize on site‑qualified coagulants and flocculants, including polymeric options such as PAC (polyaluminum chloride) and ACH (aluminum chlorohydrate), which are screened against the incoming water matrix.
Jar testing workflow and parameters
Jar tests — bench‑scale experiments mimicking full‑scale coagulation/flocculation — follow a simple sequence: (1) characterize the raw water (turbidity, pH, alkalinity, temperature); (2) fill several identical beakers (about 1–2 L each) with the sample; (3) dose each jar with a different coagulant type and/or dosage — for example, alum, ferric chloride, and PAC at 5, 10, 20, 30 mg/L (as Al or Fe); (4) rapid mix at high shear (≈100–300 rpm) for ~30–60 seconds; (5) adjust pH/alkalinity if needed (e.g., ~6.5 for alum); (6) slow flocculation at ≈30–40 rpm for 15–30 minutes; (7) settle quietly for ~30–60 minutes; (8) evaluate supernatant turbidity and inspect floc size/character (water.mecc.edu; brieflands.com; studylib.net).
Plot turbidity versus coagulant dose to locate the optimum — the lowest dosage meeting the clarity target. More coagulant is not always better: beyond a point, removal plateaus or worsens, wasting chemicals (water.mecc.edu). One study found alum at 40 mg/L could achieve 99.9% turbidity removal on a 10 NTU feed, versus lower removal at smaller doses (brieflands.com). Jar tests often yield an S‑curve; in economic terms, the choice is the “knee” where extra chemical gives diminishing returns (water.mecc.edu). Good practice also tracks final pH, alkalinity, and residual metal; for example, a target residual Al/Fe below about 0.2 mg/L is standard for drinking water (www.researchgate.net).
Screening polymer flocculants
After selecting a primary coagulant, a second jar series screens polymer flocculants at, say, 0.1–2 mg/L. For example, a cationic PolyDADMAC can be tested at 0, 0.5, 1.0, 1.5 mg/L. Add the polymer toward the end of the rapid‑mix stage (briefly at high speed ~30–60 seconds) before slow flocculation (water.mecc.edu). In Veolia’s engineering practice, adding a high‑molecular‑weight polymer after the inorganic coagulant often “develop[s] a larger floc at low dosage” and can salvage cases where coagulant alone failed to clarify (www.watertechnologies.com). If turbidity is already very low, polymer use may be omitted to save cost. Plants that standardize polymer supply typically keep qualified PAC and ACH on hand within their water‑treatment ancillaries for quick switching when raw water shifts.
Field and lab outcomes across waters
On a 10 NTU kaolin suspension (kaolin is a clay mineral commonly used in jar tests), Kord Mostafapour et al. (2008) tested alum, ferric, and PAC at pH 5.5–7.5 and found turbidity removal >99% at high dose: alum up to 40 mg/L yielded 99.9% removal; ferric reached >99% at 20 mg/L; PAC reached ~99% at 40 mg/L. All three performed similarly at optimal doses in that low‑turbidity test (brieflands.com; brieflands.com).
In a more challenging case, Husaini et al. (2018) compared lab‑made coagulants to a commercial alum on mine wastewater (~2000 NTU feed). The new alum (“tawas”) removed turbidity from ~2000 to 151 NTU (92.45% removal), compared with the commercial alum’s 548 NTU (72.6% removal). A pilot‑scale PAC reduced mine tailing turbidity from 130.7 to 2.92 NTU (97.77%) versus the plant’s PAC to 4.67 NTU (96.43%); total suspended solids fell from 196.3 to 38.7 mg/L versus 196.3 to 30.7 mg/L, respectively (jurnal.tekmira.esdm.go.id; jurnal.tekmira.esdm.go.id). The takeaway is straightforward: coagulant choice and dose can deliver order‑of‑magnitude clarity differences.
For tropical surface waters, Sururi & Hardika (2023) underscore that organic‑rich feeds challenge coagulation. They report alum dose ~20 mg/L (acidic pH) at ≈85.3% turbidity removal on moderately polluted raw water, with other similar studies at ≤70% — evidence that organics, alkalinity, and temperature matter. They also note that contamination by trace organics (e.g., tryptophan) can “reduce efficiency, necessitating higher coagulant dose and risking coagulation failure” (ejurnal.itenas.ac.id; ejurnal.itenas.ac.id).
From jar to plant trains
Once the optimal chemicals and dose are identified, plants deploy them in rapid‑mix tanks, flocculating channels, and settling basins ahead of media filters. Many power facilities opt for a conventional clarifier to remove suspended solids at practical detention times (typically 0.5–4 hours) before dual‑media filtration; lamella designs can reduce footprint, where a lamella settler (inclined plates) increases surface area and throughput. Downstream, dual‑media filters often use sand/silica media and, for longevity and stratification performance, anthracite; color or taste/odor issues may be handled with activated carbon. In high‑pressure contexts, operators select robust housings (for example, steel filter housings).
Pretreatment clarity also protects membranes. Plants commonly route clarified water to ultrafiltration or reverse osmosis; UF is a common pretreatment step (ultrafiltration) ahead of RO, and HRSG make‑up is frequently produced on brackish‑water RO trains. Where seawater is the source, large industrial users run SWRO systems. For polishing to low conductivity, ion exchange units and demineralizers are typical endpoints (demineralizers).
Operators re‑run jar tests whenever raw water quality shifts significantly. As Water.WPT McMinnville’s guide notes, “the best [coagulant] dose may change from day to day” if feed turbidity fluctuates, so jar tests should be repeated with each major change to keep dosing optimized for performance and cost (water.mecc.edu).
Troubleshooting clarification problems
Even with sound jar‑test design, clarifiers can misbehave. Common symptoms, causes, and fixes include:
- Poor floc formation or persistently high effluent turbidity. Often due to under‑dosing, an ill‑matched coagulant, or off‑spec pH. Jar‑test manuals note that “coagulant underfeed… causes the sample to appear cloudy with little or no floc and almost no settling” (studylib.net). Remedies include increasing coagulant dose or adjusting pH toward the optimum (alum ≈5.5–6.5), verified by alkalinity/pH checks (water.mecc.edu). If pH is too high, add acid or more alum; if too low, add alkali. Ensure rapid mixing is vigorous enough, and consider alternative coagulants (e.g., PAC where water is very alkaline or has low alkalinity) (www.watertechnologies.com).
- Floc stays too fine or floats. Excessive shear or overdosing high‑MW polymers can break floc; oils and surfactants can buoy it. In jar tests, “overfeeding coagulant will form dense floc that will not settle” (studylib.net). Try lowering coagulant/polymer dose and flocculation speed; check for organics and pre‑treat oily/high‑TOC waters (including pre‑oxidation where appropriate). In some cases, adding a small amount of aluminum sulfate can collapse buoyant polymer flocs.
- Rising sludge blanket or carry‑over. Points to overloading (high flow/solids) or overdosing. If flow exceeds clarifier design, solids won’t settle (short retention time) (www.lautanairindonesia.com). Dosage matters, too: “if dosage is excessive, it can actually increase the volume of sludge due to the presence of excess chemicals in the settled sludge” (www.lautanairindonesia.com). Keep flows within design, trim chemical dose, monitor return sludge rates in activated sludge clarifiers (too much recycle overloads), and maintain rakes/scrapers and sludge pumps (www.lautanairindonesia.com).
- Excess metal residuals in effluent. If filtered water shows iron/aluminum carryover or color, dose/pH may be off. Many references cite residual Al below ~0.2 mg/L for drinking water (www.researchgate.net). Measure Fe/Al, adjust pH or lower dose, and ensure adequate flocculation time. If carry‑over persists, consider switching coagulants (e.g., ferric for color‑bound NOM, or polymeric coagulant with reduced metal dose); polymer coagulants yield no metal residuals (www.watertechnologies.com).
- Poor performance on organic‑rich waters. Organic matter such as tannins, humics, or tryptophan can inhibit coagulation, “necessitating higher coagulant dose and risking coagulation failure” (ejurnal.itenas.ac.id). Screen coagulant‑aid polymers (polyDADMAC, polyamine), trial pre‑oxidation, and compare PAC/composite coagulants versus alum on colored water. If color/turbidity persists, a clarifier cocktail (alum plus carbon) or recycle adjustments may be needed (www.watertechnologies.com).
- Channeling or short‑circuiting. Uneven flow from inlet baffles or sludge buildup can dump solids. Inspect hydraulics (baffles, launder weirs) and clear floc mats. Jar tests cannot predict mechanical issues; routine inspections remain necessary.
In all cases, re‑running a jar test with the current raw water is the most reliable diagnostic. Compare jar‑test clarity to plant performance: if lab jars don’t clear at the present dose, dose is likely too low; if they do, look to mixing or pH. Multiple replicates, turbidity instrumentation, and incremental dose steps help pinpoint issues and guide corrections (water.mecc.edu; brieflands.com).
Source anchors and further reading
Industry handbooks outline coagulation/flocculation fundamentals and coagulant comparisons, notably the Veolia Water Handbook’s clarification chapter, which details charge neutralization by metal salts or polymers, flocculation, sedimentation, typical formulas, optimal pH ranges (alum best at pH 6–7; ferric effective ~5–11), the substantial sludge reduction from polymeric coagulants (≈50–90% vs. salts), and the need for “specific testing” to find broadly effective coagulants and flocculants (www.watertechnologies.com; www.watertechnologies.com; www.watertechnologies.com; www.watertechnologies.com; www.watertechnologies.com).
Jar‑test methodology is covered in practical primers and handbooks. Water.Mecc’s guide specifies rapid‑mix (∼100–200 rpm), slow‑floc (∼30 rpm), settling, and selecting the “best turbidity removal ‘for the money’,” not merely the maximum drop (water.mecc.edu; water.mecc.edu). Open‑access protocols by Pivokonský et al. (2022) provide optimization guidelines (www.researchgate.net; www.researchgate.net). Reported jar‑test outcomes include near‑complete turbidity removal (~99–99.9%) on 10 NTU kaolin feeds at alum 40 mg/L, ferric 20 mg/L, or PAC 40 mg/L (brieflands.com). Indonesian studies (Sururi & Hardika, 2023; Husaini et al., 2018) emphasize variabilities and gains from formulation changes, with mine effluent tests showing 92–98% turbidity removal in optimized cases (ejurnal.itenas.ac.id; ejurnal.itenas.ac.id; jurnal.tekmira.esdm.go.id; jurnal.tekmira.esdm.go.id).
Operational troubleshooting tips draw on lab manuals and case studies, including warnings that “if water has minimal turbidity, flocs can be hard to see… coagulant underfeed… [and] overfeed… will form dense floc that will not settle” (studylib.net) and clarifier overload notes on excessive flow and chemical dosage from Lautan Air Indonesia (www.lautanairindonesia.com; www.lautanairindonesia.com). These practical insights, combined with data‑driven jar tests, enable operators to diagnose and correct most coag‑floc‑clarification issues.
