Definition
A water quality sensor is a sensor that measures physical and chemical properties of water. Typical parameters are turbidity, pH, residual chlorine, dissolved oxygen, conductivity, COD, and ammonia nitrogen. It is used in smart water systems, environmental monitoring, agricultural product traceability, fertigation, and pollution tracing. Readings go to the cloud platform over 4G, WiFi, or LoRa, and support water-quality alarms, irrigation decisions, and the traceability record.
Parameters
Physical and chemical readings include pH, conductivity (close to EC), turbidity, and dissolved oxygen. Disinfectants include residual chlorine and ozone. Nutrients include ammonia nitrogen, nitrate, and phosphate. Bulk indices include COD, BOD, and TDS. In fertigation, pH and EC sensors watch the nutrient solution and guide acid and fertilizer dosing. In pollution tracing, a multi-parameter network is read with remote sensing and big data to follow a discharge back to its source.
Where it is used
In smart water it covers water plants, wastewater plants, and the pipe network. In environmental protection it covers rivers, lakes, groundwater, and pollution tracing. In agriculture it covers source water for fertigation and drip irrigation, plus the water record for agricultural product traceability and organic certification. In aquaculture it watches dissolved oxygen, ammonia nitrogen, and the other parameters that decide whether the water is usable. With blockchain, the readings can be stored as evidence for traceability and compliance.
Architecture and selection
The system is the sensor (one parameter or several), a transmitter, a data logger or gateway, and the cloud platform. Probes need calibration and cleaning on a schedule. Selection should cover range and accuracy, fouling resistance and service life, the radio, fit with the cloud platform and API, and the standards of environmental protection and the water industry. SimtoLink can supply the sensors, the gateway, the cloud platform, and the API for a water-quality network.
Implementation
Name the water and the parameters, choose the sensor and where it sits, install the gateway and the network, connect the cloud platform to the business system, and write the calibration and maintenance routine. For pollution tracing, read the probes with remote sensing, big data, and AI. For agricultural product traceability, connect them to fertigation and organic certification. Also check that the API is open and that readings can be filed on a blockchain.
The Medium a Water-Quality Probe Needs
A water-quality sensor measures the physical and chemical indices of the small body of water around the probe, not the average of a whole river or a whole pond. Acidity, dissolved oxygen, turbidity, and conductivity each have their own sensing membrane or optical window. Mounting them on one pole does not mean they are valid at the same time. Dissolved oxygen needs the water outside the membrane to be renewed. A reading in a dead zone falls slowly and looks like the water is short of oxygen, when the probe has used up the oxygen around itself.
Before use, confirm the flow represents the purpose of the intake. An irrigation intake, a discharge point, and the center of a pond are not the same point. The probe should be immersed where water moves, away from silt and from corners where bubbles collect. The optical window must be reachable for wiping. Turbidity of a probe sunk in mud stays high. Calibration solution must be within its validity and close in temperature to the field sample. "Calibrating" with tap water in place of a standard defines tap water as the standard point, and every later reading is arranged around that wrong zero.
Biofilm is the most common false change in online water quality. The membrane is still there and communication is still there, while the value climbs or falls slowly, and the curve looks very much like a real water-quality process. Without a record of cleaning time, the step before and after cleaning cannot be separated from a pollution event. After a conductivity probe scales, its range shrinks and high salinity is flattened. Using temperature in the air to compensate oxygen in the water turns the air conditioner on and off into a false dissolved-oxygen curve.
Do not use moisture in soil in place of a water sample. Soil moisture describes the soil body. A water-quality probe describes free water. The two cannot fill each other's missing measurements. Temperature compensation should use a temperature sensor immersed in the same water. If a probe still reports the last value after it leaves the water, dosing control keeps acting while the measurement is in air.
Different indices cannot share one maintenance interval. Dissolved-oxygen membranes, turbidity windows, and conductivity electrodes foul at different speeds. Cleaning on the slowest one leaves the fastest one drifting for a long time. After cleaning, wait until the reading is stable before returning it to control. A window just wiped has a stretch of false improvement. If calibration is done on a laboratory sample while the field is hot or highly saline, the curve has to be checked under field conditions. The laboratory slope cannot be assumed still valid.
The sampling point should avoid the instant mixing zone of a dosing point. Too close downstream of a chlorine inlet, the reading jumps hard as the dosing pump starts and stops. That jump is not river water quality. Oxygen at the surface and at the bottom of a pond can differ a great deal. Sampling only the surface overestimates the whole water body. If an irrigation source turns muddy after rain and the probe is buried in silt, turbidity stays high and does not recover when the rain stops. Check the probe first, rather than reporting pollution the whole time.
Before data are used, they should carry a quality flag: calibrating, cleaning, out of the water, over range. A continuous curve with no flag is treated as valid throughout. If an over-range value is recorded as the top of the range, a high concentration is underestimated and the average still looks acceptable. Before several probes are averaged, remove the one that has already failed. Otherwise one drift pulls the conclusion for the whole cross-section aside.
Comparing cross-sections should also use similar sampling times. Dissolved oxygen upstream at noon and downstream at dawn are a daylight cycle apart and cannot be written directly as upstream being better. Time, water depth, and whether rain has just fallen should be kept with the reading. Without those three, a later reader sees only a rising curve and does not know whether the rise came from sunlight, from flow, or from a probe just wiped clean.