Closed-loop chemical control is often presented as a simple diagram: an analyser measures water quality, a controller compares the reading with a setpoint, and a metering pump adds chemical. Real systems are less tidy. Sensors foul, samples arrive late, reactions take time, pumps deliver pulses, process demand changes, and one chemical can influence several measured variables. A dependable loop is therefore an engineered combination of chemistry, sampling, analytics, dosing hydraulics, and control logic.
At Jams (Pvt.) Ltd, we assess these loops from the process outward. Our analytical-instrument work establishes what can be measured reliably, while our water treatment and disinfection capability connects that measurement to a practical treatment action. This is particularly important in Pakistani textile, cement, chemical, and utility plants, where water source, load, temperature, and operating schedules may vary significantly.
What “closed loop” should mean
A loop is closed only when a measured process response influences the dosing command. A pump receiving a fixed manual percentage is open-loop control. A pump paced from water flow is feed-forward control: it responds to throughput but not directly to the resulting chemistry. An analyser trimming a flow-paced base dose creates a combined strategy that can handle both predictable load and slower chemical variation.
Not every application benefits from direct feedback. If the measured response occurs long after injection, an aggressive controller may chase old information and create oscillation. If the sensor is placed where fresh chemical reaches it before mixing with the process, the loop can shut the pump off early and under-treat the main volume. The correct question is not “Can this analyser control a pump?” but “Does this measurement represent the process soon enough to support safe corrective action?”
- Define the treatment objective and the consequence of exceeding either limit.
- Identify the variable that best indicates that objective, not merely the easiest variable to measure.
- Measure process delay from injection through mixing, reaction, transport, and sample conditioning.
- Provide an independent maximum dose, run-time limit, or permissive where overfeed could be harmful.
- State the safe control state for sensor fault, loss of sample, low chemical level, and loss of process flow.
Select analytics that fit the chemistry
pH measurement
pH control is strongly nonlinear. Near a neutralisation endpoint, a small reagent addition can create a large pH change, while buffered water may require much more chemical earlier in the curve. A controller tuned at one load may behave poorly at another. Acid and alkali feeds should be separated physically and logically, with suitable mixing volume and interlocks. The electrode needs a representative, continuously renewed sample and maintenance appropriate to coating, temperature, and process contamination.
ORP measurement
Oxidation-reduction potential reflects the combined oxidising and reducing condition of the water; it is not a direct concentration reading for one chemical. ORP can be useful for monitoring or controlling certain disinfection and oxidation processes when site chemistry establishes a meaningful relationship. That relationship can shift with pH, temperature, contaminants, and oxidant demand. A copied ORP setpoint from another plant is not a substitute for process validation.
Disinfectant residual and conductivity
A suitable residual analyser can offer more specific information than ORP for a selected disinfectant, but sample flow, pH range, chemistry, and calibration requirements remain important. Conductivity is useful for dissolved-ion trends and cooling-water blowdown, yet it does not prove inhibitor concentration or microbial control. Each analyser must be selected for its actual sample conditions and maintained as a measuring system, including flow cell, filters, valves, drains, and reagents where applicable.
Engineer the sample before tuning the controller
A sophisticated instrument cannot correct a poor sample. The take-off should represent the treated stream after adequate mixing and reaction. Sample tubing should minimise unnecessary delay and avoid dead legs, gas pockets, sediment traps, excessive heat gain, or pressure beyond the analyser assembly’s limits. Flow should remain within the cell requirement and be monitored when loss of sample could leave a plausible but frozen reading.
Fast bypass flow can reduce transport delay, with only the required branch passing through the measurement cell. The disposal or return of sample water must be safe and hydraulically sound. In dirty water, automatic or manual cleaning provisions may be needed, but filtration must not remove the very constituent being measured. Accessibility matters: if operators cannot safely inspect, clean, calibrate, and replace consumables, data quality will decline regardless of instrument capability.
Connect the analyser and metering pump correctly
Common command methods include pulse pacing, analogue current, relay on/off, and digital communications. The signal must be scaled consistently at both ends. For a 4–20 mA command, document what pump output corresponds to 4 mA and 20 mA, what happens below range, and how a broken wire is detected. The pump should report meaningful status such as running, fault, low level, or stroke feedback where available, rather than leaving the PLC to assume that a command equals chemical delivery.
Flow feed-forward often provides the stable backbone. The required dose ratio converts measured process flow into a base pump demand. The analytical controller then applies a bounded trim. Limiting the trim prevents a drifting sensor from demanding unlimited chemical. For slow tanks or loops, proportional-integral settings should respect residence time; derivative action is rarely a cure for a noisy or delayed sample. On/off control can work for suitable buffered volumes when hysteresis and minimum run times prevent rapid cycling.
Verify the final control element
The “final control element” includes the entire dosing path, not only the pump motor. A blocked injection valve, empty tank, gas-bound liquid end, siphoning line, or failed check valve can break the loop. Calibration at representative back pressure links command to actual delivery. Relief, back-pressure, and anti-siphon accessories should be selected as a hydraulic set. Chemical compatibility must cover every wetted component from tank connection to injection quill.
Pitfalls that create unstable or unsafe loops
- Placing the sensor too close to injection and measuring an unmixed chemical plume.
- Tuning around a dirty electrode instead of restoring the measurement.
- Using feedback alone when process flow changes faster than the analytical response.
- Failing to define pump behaviour during analyser maintenance or sample loss.
- Scaling the analogue signal differently in the controller, PLC, and pump.
- Allowing integral wind-up while an interlock has stopped chemical delivery.
Another trap is treating a stable trend as proof of accuracy. A coated sensor can produce a smooth but biased value. Good practice combines scheduled inspection, calibration or verification, process cross-checks, and alarm review. Operators should be able to place a loop in a documented manual mode for maintenance without defeating essential maximum limits or forgetting to return it to automatic operation.
When to involve JAMS
Call JAMS when a new measurement is intended to drive dosing, an existing loop oscillates, sensors require excessive attention, or separate instrument and pump suppliers are each blaming the other. Useful information includes a process and instrumentation diagram, water analysis, chemical data, flow range, tank volumes, injection and sampling locations, pressure, expected process delay, existing signal list, and maintenance history.
JAMS can review the measurement point, analyser arrangement, pump duty, accessories, signal scaling, alarms, and commissioning checks as one application. We are an authorized ProMinent partner in Pakistan; our ProMinent authorization information explains the local relationship where ProMinent analytical and metering equipment is appropriate. We do not replace site chemistry validation with a product claim: the loop must be proven against the real process.
FAQ
Is flow-paced dosing considered closed-loop control?
Flow pacing is feed-forward, because it changes dose with throughput without measuring the resulting chemistry. It is often an excellent base strategy. Adding a representative analytical measurement as a bounded trim creates a more complete feedback arrangement.
Why does a pH loop keep overshooting?
Likely causes include excessive process delay, poor mixing, an oversized pump, strong reagent, nonlinear titration behaviour, incorrect controller tuning, or a sensor located too near injection. Measure the timing and verify delivered flow before repeatedly changing tuning values.
What should happen if the analyser loses sample?
The design should detect sample loss and move to a defined safe state. Depending on the process, that may mean stopping the pump, holding a limited flow-paced dose, or alarming for operator action. Continuing unrestricted feedback from a frozen value is not acceptable.
Can one analyser control two chemical pumps?
It can be engineered in specific processes, but interacting chemicals can create unstable or unsafe behaviour. Acid and alkali, for example, should not fight each other from one pH error. Sequencing, deadband, independent limits, mixing time, and failure states need a deliberate control review.
