The Real-Time Production Monitoring system is based on active sensors which are used to detect the manufacturing objects during production. The effectiveness of production monitoring system is highly dependent on the accuracy of these sensors. The inherent unreliability of RFID devices abstain their deployment in such systems. We have developed a RFID based data consistency model which ensures a consistent production monitoring. This model uses a probabilistic sequence detection algorithm which detects the sequences of manufacturing objects in a production environments. The detected sequences are then assigned a probability which is based on the accuracy or the probability of detection of RFID devices which are used to detect the objects in the sequence. The previous consistency model assumes a fix probability of detection for each RFID device. This might induce some discrepancy in probabilities of detected sequences, because some RFID devices might not have the same accuracy as we have assumed. Since the system is unaware of actual accuracy of the RFID devices, hence to estimate the accuracy of RFID there must be constant calibration of the probability of detection for each RFID device. This Calibration is done by maintaining a history of detected objects for each RFID device. The historical information is then compared with detected sequences to determine the accuracy for each RFID device.