Example 12: corr_rinkoTB - correction of Rinko dissolved oxygen using logger temperature sensor

RINKO formula revision

In November 2013, JFE Advantech introduced a revised version of the equation relating sensor voltage output to dissolved oxygen saturation. The revised equation is referred to as the (B) version, and is described in this example. For a description of the older, original equation, refer to Example 6.

Sensors calibrated according to the (B) version of the equation should be labelled as such, so if the sensor has no label it probably uses the original equation. However, there is always the possibility that this label was not applied or is now missing; for example, older sensors returned to the manufacturer to be recalibrated will be processed using the (B) equation, and may not be re-labelled. The most reliable way to determine which equation applies is to check the most recent calibration certificate from the manufacturer.

It is obviously important that the correct equation is used by the logger, which requires the correct channel type to be set in the logger's configuration (see the channel command, and the Section Supported channel types). The (B) version of the equation described in this example requires channel type "doxy09". If you believe that the logger configuration is not correct for your sensor, please contact RBR Ltd for assistance.

Consider an RBRconcerto³ C.T.D.DO logger, where the DO channel is a Rinko-III dissolved oxygen sensor from JFE Advantech. For a number of reasons, it may be desirable to use the logger's measured temperature to compensate the temperature dependence of the raw DO output, rather than the built-in Rinko-III temperature sensor. This can be done, but requires some additional calculation, because as pointed out earlier (see the section corr_rinko - Correction of Rinko Dissolved Oxygen using Rinko Temperature sensor), the time response of the logger's temperature sensor does not match that of the Rinko-III DO sensor.

There is no change to the primary equation for the temperature corrected output of the DO sensor:

as before, but in this case Tr is a 'delayed' version of the monitored temperature, computed by the following simple filter operation:

where

T is the monitored temperature,
Tr(n) is the delayed temperature output,
Tr(n-1) is the previous delayed temperature output,
K is a term relating the measurement interval and the time constants of the sensors as follows:

where

TCdo is the time constant of the Rinko-III Do sensor,
TCt is the time constant of the logger's temperature sensor,
P is the logger's measurement interval set by the "sampling period" command.

The time constants are specified in seconds using the additional auxiliary coefficients x6 (TCdo) and x7 (TCt); the logger correctly accounts for the fact that the measurement interval P is specified in milliseconds.

Often the logger's temperature sensor has a much faster response than the Rinko-III DO sensor, in which case it is acceptable to set TCt to zero. Also, in cases where the measurement interval P is long compared to the sensor time constants, the logger limits the computed value of K to 1, in which case the 'delayed' temperature Tr(n) follows the input temperature T exactly.

The form of the delayed temperature equation used by the logger becomes:

In the example discussed here, the value of n0 is 2, and value(n0) is the temperature reported by the logger.

After calculating DOtcomp using this 'delayed' temperature value, it then becomes the input to a simple linear equation, which according to JFE Advantech can be further compensated for pressure effects as follows:

where

E, G, H are coefficients provided by JFE Advantech,
P is pressure in Mpa,
DOcorr is the fully corrected dissolved oxygen output in percentage saturation.

The G, H coefficients can be modified by the user to update the calibration of the sensor; the remaining coefficients should not need to be modified.

Casting the equation into the form used by the logger gives:

where

c0, c1 are G, H,
x4 is E,
n1 is the index of the pressure channel, in this example 3,
value(n1) is the final output value of the pressure channel in dbar.

The logger correctly accounts for the fact that value(n1) is in dbar but the value of coefficient x4 (E) is determined assuming the pressure to be in MPa.

Examples

CODE

>> calibration 4 type
<< calibration 4 type = doxy09

Confirm the channel type.

CODE

>> calibration 4 datetime = 20171201000000, c0 = 0.346, c1 = 1.08873

Set the "user" coefficients.

CODE

>> calibration 4 datetime = 20171201000000, x0 = -41.7148, x1 = 25.425, x2 = -0.08097, x3 = 0.0021, x4 = 4.5e-5, x5 = 0.0, x6 = 4.2, x7 = 0.0

Set the coefficients provided by JFE Advantech, and the time constants.

CODE

>> calibration 4
<< calibration 4 type = doxy09, datetime = 20171201000000, c0 = 0.346, c1 = 1.08873, x0 = -41.7148, x1 = 25.425, x2 = -0.08097, x3 = 0.0021, x4 = 4.5e-5, x5 = 0.0, x6 = 4.2, x7 = 0.0, n0 = 5, n1 = 3

Request confirmation of all calibration coefficients.