In the June/July 2001 issue of

Control Engineering Europe, we have taken the discussion of intelligent sensors to the next level.

Smart instruments were created to provide more precise measurements and allow wider ranges of operation. Em-bedding a microprocessor in the transmitter made it possible to use digitally stored data to provide precise corrections for all the nonlinearities of indiviual sensors. By automatically adjusting zero and gain in software, an analogue temperature transmitter's typical drift of 0.025% of span per degree Celsius was improved dramatically to 0.006%. Manufacturers could just as easily store ranging data in their instruments, so they could be re-ranged by simply telling the microprocessor to look up a different set of values in the tables. Operators could now "zoom in" on the portion that better represented the process operating conditions. Honeywell's ST-3000 transmitter claimed a turndown ratio of 400:1, a figure considered so fantastic many didn't believe it.

But going beyond these benefits, what else can smart sensors do for us?

Most of us who have ever had to deal with a faulty sensor—or one that we think is faulty—know that the most perplexing question about a sensor is simply: is it working properly or not? Can we trust the data from the sensor? If sensor data are bad, what can we do? Take drastic countermeasures?

In our article, we give readers a glimpse of what highly intelligent, self-evaluating (SEVA) sensors will do for us. As an example, consider the situation in a typical wastewater plant as depicted in the diagram above. After one hour of operation, fouling occurs on the membrane of a dissolved oxygen sensor, which results in the drop in the permeability of the membrane, and the raw dissolved oxygen measurement slowly drops to less than 2 ppm.

However, the SEVA sensor is capable of detecting the membrane fouling and compensating for the measurement. The validated measurement stays near 8 ppm for the interval. An independent measurement of the true level is also shown, to demonstrate that the corrected value provides a reasonable estimate of the true process value.

The SEVA sensor also provides its own uncertainty analysis, represented by the blue shading. After one hour it increases to indicate the reduced quality of the corrected measurement. The operator might increase the dissolved oxygen level to 8.5 or 9 ppm to ensure that the true measur-and does indeed stay above 7 ppm despite the increase in measurement uncertainty.

If your career depends on receiving good quality data from field sensors you should have a look at this new development from the UK, which may become an European standard. Our article is at www.controleng.com under CE Europe, June 2001.