Unnecessary process shutdowns can occur when a temperature measurement reading is outside of process alarm limits but the actual process variable is within the acceptable range. This can lead to lost revenue and safety concerns associated with the shutdown process. Another issue, on-scale failure, occurs when the output of a temperature transmitter is within process alarm limits yet does not accurately indicate actual process information. This can lead to energy inefficiency, low quality product, low yield/throughput and safety issues.

The 848T temperature transmitter's Measurement Validation Diagnostic can detect data abnormalities before a shutdown condition occurs and can identify on-scale failures as they happen. Additionally, the Measurement Validation Diagnostic has the capability to detect abnormally fast process temperature changes. The onset of a runaway reaction can be difficult to detect when it is at a manageable level and near impossible to control once an alarm state is reached. With the ability to detect abnormally fast process temperature changes before alarm conditions are met, action can be taken sooner.

By evaluating sample to sample temperature deviations over an historical trend, Measurement Validation can detect abnormalities in measurements related to degraded conditions such as:

• Failing thermocouples or RTDs
• Electronic interference
• Excessive vibration
• Corroded termination points
• Loose electrical connections
• Unexpected process variation

It also makes measurement qualification before control possible. By providing validity of measurements points, users can implement a process control strategy based on valid measurements while discarding invalid points. Once an invalid temperature is detected, preventative action can be performed before shutdown is necessary or before the process is compromised.

Measurement Validation calculates the average sample to sample deviation of measurements by comparing each individual temperature measurement to an averaged trend consisting of current and previous measurements. Users can select a sample rate of 1- 10 seconds per sample. A higher frequency of samples (lower number seconds per sample) will put more emphasis on temperature sensor issues, while lower sampling frequency will put a greater emphasis on process conditions.

The deviation output provides a user with a representation of how much change is detected in the transmitter measurement in relation to a moving average of previous data points. Deviation processing is performed before any filtering has been applied to the measurement, allowing the transmitter to capture the noise signature while still providing a stable output to the DCS. The noise signature may be used to identify a variety of concerns in the plant including sensor degradation, sensor wiring connectivity, high vibration (affecting the measurement), and abnormal process variations.

This output is then compared to a user selectable alert limit. If the output exceeds the alert limit, the severity setting indicated during configuration will produce the following statuses:

• "Not used", there is no status produced
• "Advisory", an advisory plant web alert will be generated
• "Warning", a maintenance plant web alert will be generated as well as setting the PV status to "degraded"
• "Failure", a failed plant web alert will be generated and sets the PV status to "bad"

The Measurement Validation diagnostic can also determine process rate of change. This calculation is based off of the damped transmitter measurement. Rate of change is calculated by measuring the time and difference between samples and applying a moving filter. As with the deviation output, the process rate of change is then compared to a user selectable alert limit. If the output exceeds the customer alert limit, a status will be produced according to what severity limit was selected.