Calibration of level sensors

Calibration of level sensors

The sophistication of calibration procedures for level sensors depends on the degree of accuracy required. If the accuracy demands are not too high and a tank is relatively shallow, a simple dipstick inserted into a tank will suffice to verify the output reading of any other form of level sensor that is being used for monitoring the liquid level in the tank. However, this only provides one calibration point. Other calibration points can only be obtained by putting more liquid into the tank or by emptying some liquid from the tank. Such variation of the liquid level may or may not be convenient. However, even if it can be done without too much disturbance to the normal use of the tank, the reading from the dipstick is of very limited accuracy because of the ambiguity in determining the exact point of contact between the dipstick and the meniscus of the liquid.

If the dipstick method is not accurate enough or is otherwise unsuitable, the alternative method of calibrating level is to use a calibration tank that has vertical sides and a flat bottom of known cross-sectional area. Tanks with circular bottoms and rectangular bottoms are both commonly used. With the level sensor in situ, measured quantities of liquid are emptied into the tank. This increases the level of liquid in the tank in steps, and each step creates a separate calibration point. The quantity of liquid added at each stage of the calibration process can be measured either in terms of its volume or in terms of its mass. If the volume of each quantity of liquid added is measured, knowledge of the cross-sectional area of the tank bottom allows the liquid level to be calculated directly. If the mass of each quantity of liquid added is measured, the specific gravity of the liquid has to be known in order to calculate its volume and hence the liquid level. In this case, use of water as the calibration liquid is beneficial since its specific gravity is unity and therefore the calculation of level is simplified.

To measure added water in terms of its volume, calibrated volumetric measures are used. If a 1 liter measure is used, this has a typical inaccuracy of ±0.1%. Unfortunately, the errors in the measurement of each quantity of water added are cumulative, and therefore the possible error after 10 quantities of water have been added increases ten-fold to ±1.0%. If 20 quantities are added to create 20 calibration points, the possible error is ±2.0% and so on.

Better accuracy can be obtained in the calibration process if the added water is measured in terms of its mass. This can be done conveniently by mounting the calibration tank on an electronic load cell. The typical inaccuracy of such a load cell is ±0.05% of its full-scale reading. This means that the inaccuracy of the level measurement when the tank is full is ±0.05% if the load cell is chosen such that it is giving its maximum output mass reading when the tank is full. Since the total mass of water in the tank is measured at each point in the calibration process, the measurement errors are not cumulative. However, the errors do increase for smaller volumes of water in the tank because the measurement uncertainty is expressed as a percentage of the full-scale reading of the load cell. Therefore, when the tank is only 10% full, the possible measurement error is ±0.5%. This means that calibration inaccuracy increases for smaller quantities of water in the tank but the measurement uncertainty is always less than the case where measured volumes of water are added to the tank even for low levels.

Wherever possible, the liquid used in the calibration tank is water, since this avoids the cost involved in using any other liquid and it also makes the calculation of level simpler when the quantities of water added to the tank are measured in terms of their volume. Unfortunately, the liquid used in the tank often has to be the same as that which the sensor being calibrated normally measures. For example, the specific gravity of the measured liquid is crucial to the operation of both hydrostatic systems and capacitive level sensors. Another example is level measurement using a radiation source, since the passage of radiation through the liquid between the source and detector is affected by the nature of the liquid.