لماذا جرعة المادة المخثرة لديك خاطئة (وكيف يساعد التيار المتدفق)

High-traffic-inspired · Technical · Published: October 14, 2026

Key Takeaways

• Coagulant dosing is the most consequential chemical decision in a conventional treatment train. When the dose is wrong, plants pay for it downstream in higher turbidity, heavier filter loading, elevated disinfection by-product formation and compliance risk [1].
• The jar test — still the default dose-setting tool — is a manual, batch-wise lab procedure run outside the plant, so it always reacts to raw-water changes after the fact [2].
• A streaming current detector (SCD) measures the net surface charge of coagulated colloids in real time and closes the loop on the dosing pump automatically, trimming the dose before turbidity drifts [1].
• One published retrofit record tells the story in operational terms: daily jar tests dropped out of the routine, effluent turbidity and color stopped swinging, alum consumption fell sharply, and the instrument paid for itself out of chemical savings alone [1].
• Overdosing is not “safe”: it burns chemical budget, multiplies sludge, and pushes residual aluminium upward in finished water — WHO guidance is to avoid excessive aluminium dosage and optimize the coagulation process [3].
• SCDs are not magic. Probe fouling, poor rapid mixing, pH excursions and low-turbidity/high-color waters all demand engineering attention, and some plants save more with feed-forward control than with a streaming current loop [4].

Why Coagulant Dosing Goes Wrong

Every conventional surface-water plant makes the same bet many times a day: that the coagulant dose leaving the chemical room matches the water arriving from the source. The bet fails for three structural reasons.

Raw water never sits still. Storms, seasonal turnover, algal blooms and upstream discharges move turbidity, natural organic matter, pH and alkalinity on timescales of minutes to hours. The required coagulant dose has to move with them, because the optimum dose for removing suspended, colloidal and dissolved matter shifts as the raw water shifts [2]. A dose that was correct at 6 a.m. can be wrong by noon.

The jar test looks backward. Jar testing is manually conducted in batches outside the treatment plant: an operator grabs samples, runs a series of beakers, reads the floc, and applies the “best” dose to the full-scale process [2]. Practitioners list the shortcomings bluntly — non-continuous testing, variability between operators, visual interpretation of floc characteristics, and low correlation with full-scale operation, especially at low turbidity in high-color waters [5]. Between two jar tests, the plant is effectively flying blind, and a plant that runs jar tests only once or twice per shift is dosing yesterday’s water.

Overdosing becomes a habit. When measurement is sparse, operators add a margin. A margin on top of a margin becomes chronic overfeed. The costs compound quietly: more chemical spend, more sludge to dewater, more aluminium carried into the filters — and, with alum, a rising residual aluminium concentration in finished water. WHO’s background document on aluminium is explicit that residual aluminium is minimized by optimizing the coagulation process, using optimum pH, and avoiding excessive aluminium dosage [3]. In other words, the cheapest and healthiest dose is the right dose, not the biggest one.

What a Streaming Current Detector Actually Measures

A streaming current detector is an on-line electrokinetic charge analyzer. Natural colloids — clay, silt, organic particles — carry negative surface charges that repel each other and keep them suspended. Coagulants work by neutralizing that charge so particles can collide, bond and settle. The controlled variable for coagulation is therefore charge, and charge is exactly what an SCD measures continuously.

Inside the sensing chamber, a piston reciprocates in a narrow annular gap, forcing the sample back and forth across the walls. Counter-ions in the electrical double layer get dragged along, separating charge and inducing a small alternating current at the electrodes. The processed signal — reported in dimensionless streaming current units, typically on a −100 to +100 display scale — tracks the net particle surface charge of the sample. As coagulant dose rises past charge neutralization, the signal crosses zero; underdosing reads negative, overdosing reads positive [1].

That single property is what makes closed-loop control possible: the SCD output is a direct, continuous indicator of “has this water reached its ideal point of destabilization,” available every second instead of every shift.

Closing the Loop: From Charge Signal to Dosing Pump

In closed-loop service, the SCD feeds a process controller that compares the streaming current with a setpoint — established once by jar testing at commissioning — and outputs a 4–20 mA correction signal to the coagulant metering pump. Industrial SCD packages ship with PID control, auto-tune, alarm contacts and remote mounting options, so the loop can live inside the existing DCS or PLC architecture [1].

The controller response matters as much as the sensor. The output should ramp the dose gradually rather than spike it, because abrupt dose swings upset flocculation and can shear fragile floc. Well-tuned streaming current loops respond to storm-front pH and turbidity excursions in near-real time, holding the charge setpoint while the raw water swings underneath.

Plants that pair the SCD with complementary raw-water instrumentation get the most out of the loop: raw turbidity, pH and temperature provide the context, while the SCD provides the verdict. A compact multi-parameter probe covering pH, conductivity, ORP and temperature in one body keeps the inlet picture synchronized — “Four parameters. One probe. Zero alignment error.” — and modern digital outputs let the same devices feed SCADA or machine-learning dose models, sensors that feed your AI water model, not just your dashboard [6].

Dosing Strategies Compared

Strategy	Principle	Strengths	Watch-outs
Manual jar test	Bench-scale floc test on grab samples; operator sets the plant dose	Low cost; chemist-friendly; needed when switching coagulants	Batch-wise and hours behind the plant; operator-to-operator variability; weak low-turbidity correlation [2][5]
Fixed / calendar dose	Dose adjusted only by season or flow	Zero instrumentation to maintain	Blind to storm events and blooms; invites chronic overdose as a safety margin
Feed-forward control	Raw-water signals (turbidity, pH, UV254) predict the dose before it is needed	Reacts before quality drift reaches the filters	Model drift over seasons; depends on reliable raw-water instruments [4]
Streaming current closed loop	SCD reads net colloid charge; PID trims the pump continuously	Continuous control on the actual variable that matters; near-real-time response	Probe fouling and mixing quality must be managed; setpoint still set by jar test [1]
Model-based / AI dosing	Historical plant data trains a dose predictor	Can fold in many variables at once	Requires clean, synchronized data streams; verify against jar tests before trust

The Cost of Getting It Wrong

The penalty table has two columns.

Too much coagulant. Chemical spend rises directly with dose, and every excess milligram becomes hydroxide sludge that must be settled, thickened and dewatered. With aluminium salts, overdosing pushes residual aluminium into the finished water, which regulators and the WHO treat as an optimization indicator to be minimized rather than a number to be traded away [3]. Overdosed water can also carry positively charged colloids and excess polymer into filters, shortening runs.

Too little coagulant. Under-coagulated particles pass the clarifier and load onto the filters, driving faster headloss, more frequent backwashes, and — in the worst case — particle breakthrough that raises filtered turbidity and carries disinfection-by-product precursors into the clearwell. Elevated DBP formation and regulatory non-compliance are the documented downstream consequences of a dose that is wrong on the low side [1].

The upside of closing the loop is equally concrete. One vendor-published before/after record from a plant retrofit describes the typical pattern: two or more daily jar tests fell out of the routine, effluent turbidity that had been oscillating settled into a tight band at a consistently lower level, effluent color stopped spiking during storms, and alum cost dropped substantially — with the instrument paying for itself from chemical savings alone [1]. Feed-forward projects report savings of the same order: a New Zealand plant replaced its streaming current device with a model-based feed-forward scheme at a highly variable DAFF works and reported meaningful reductions in chemical costs [4].

Implementation Notes and Limitations

A streaming current loop earns its keep only if the surrounding process is honest about its weaknesses.

• Set the setpoint with chemistry, not folklore. The target streaming current value is determined by jar testing at commissioning and re-checked whenever the coagulant product or source chemistry changes [1]. The SCD does not replace the jar test; it replaces the daily dependence on it.
• Mixing decides everything. Charge measurement assumes the coagulant is actually dispersed. A starved or short-circuited rapid-mix basin will feed the SCD a fiction. Verify flash-mix performance before blaming the instrument.
• pH interacts with charge. Aluminium and iron coagulants change the charge balance differently at different pH. A drifting raw-water pH will masquerade as a dosing problem, so pH belongs in the loop’s instrument set [3].
• Keep the probe clean. Sensing walls accumulate deposit, and fouled probes read sluggish or biased values; schedule the sensor’s cleaning cycle and alarm on signal flatness.
• Know when the loop is the wrong tool. Very low turbidity with high color, waters dominated by dissolved organics, or plants with unstable alkalinity can challenge a charge-only controller; feed-forward and hybrid schemes sometimes outperform, as the New Zealand case shows [4].
• Buy on lifecycle, not on sticker price. The dose-control payoff comes from years of trimmed chemical spend — buy the sensor, own the outcome — so weigh probe longevity, spares availability and support response alongside the instrument price.

For plants still dosing by habit, the sequence is simple: instrument the raw water, jar-test once to set the charge target, let the streaming current loop hold it, and audit the chemical invoices quarterly. The dose that was “wrong” is usually wrong because nobody was watching the variable that actually controls coagulation — surface charge.

References

1. Milton Roy — “Streaming Current Detectors: Automated Coagulant Control for Water & Wastewater Treatment.” Vendor process page describing the SCD principle, PID control architecture, and published before/after performance data (jar tests eliminated, turbidity, color, alum cost, ROI). https://www.miltonroy.com/en/industries/water-treatment/process-overview/streaming-current-detection/
2. PUB, Singapore — “Seamless Coagulation Control” Global Innovation Challenge statement. Explains that the optimum coagulant dosage changes with fluctuating source-water quality and that jar testing is a manual, batch-wise procedure run outside the treatment plant. https://www.pub.gov.sg/Industry/Events%20and%20Programmes/Global%20Innovation%20Challenge/Challenges/2020/Seamless%20Coagulation%20Control
3. World Health Organization — “Aluminium in Drinking-water,” background document for the WHO Guidelines for Drinking-water Quality. Lists optimization of coagulation, optimum pH and avoidance of excessive aluminium dosage as the methods for minimizing residual aluminium. https://www.who.int/docs/default-source/wash-documents/wash-chemicals/aluminium.pdf
4. Lutra Ltd — “Replacing a streaming current device with feed-forward control yields big savings” (Wainuiomata Water Treatment Plant, Wellington Water, New Zealand). Documents chemical-cost savings with feed-forward control on a highly variable DAFF plant. https://www.lutra.com.au/case-studies/wainuiomata-water-treatment-plant
5. Water & Wastes Digest — “Quicker than Routine” (Chemtrac). Enumerates the known shortcomings of jar-test-based dose control: non-continuous testing, operator variability, visual interpretation of floc, and low correlation with plant operation. https://www.wwdmag.com/quicker-routine
6. Shanghai ChiMay — Water Quality Sensor Buying Guide 2026 (technical resource page). Portfolio overview of inline pH, conductivity and multi-parameter sensors with Modbus RTU/TCP and 4–20 mA outputs for synchronized raw-water monitoring alongside charge-based dose control. https://www.chimaycorp.com/Water_Quality_Sensor_Buying_Guide_2026/223.html

About the author: Written by the Shanghai ChiMay Technical Editorial Team — process engineers who specify and support in-line water quality analyzers, multi-parameter sensors and control instrumentation for municipal and industrial water treatment plants.