Five commercial electronic relative humidity (RH) sensors (A to E), each with three replicates, were evaluated in a swine barn for one year. The sensors were calibrated in a laboratory before the in-barn trial and were brought back to the laboratory periodically for four intermediate and final calibrations. All of the sensors were affected by the corrosive barn environment. None of the sensors could maintain their stated accuracy after one year. Two sensors, A and C, each had one failed replicate. Sensor A also showed a malfunction at 15% RH. After one year, the sensor errors varied from 63% RH (sensor B) to 17.3% RH (sensor A) over the 15 to 85% RH range. Sensor B was significantly more accurate than sensors A and E (PO.05) However, over the 55 to85% RH range, sensor A had the lowest error of 2.8% RH while sensor B had the largest error of 8.9% RH. Sensor A was significantly better in accuracy than sensors B, D, and E (P<0.05) over this reduced range. The linearity of all sensors was impaired while sensor static sensitivity and hysteresis were relatively stable. One year appears to be the minimum period for evaluating RH sensors for use in barns; a shorter period would fail to determine sensor malfunction or failure. The sensor should be calibrated periodically after one month in-barn and once every three to four months thereafter. Because variations existed among individual sensor units of the same sensor type, three replicates are suggested to get reliable evaluation results. Because sensor accuracy varied widely at different humidity levels, a humidity sensor used for control purposes should be chosen according to its accuracy over the whole measuring range (15 to 85% RH) as well as the control range (55 to 85% RH).
展开▼