Linearization of temperature measurements
Temperature is the most widely measured value in process control. Across a wide variety of industries from large-scale chemical manufacturing to small labs, accurate temperature measurement is relied upon to ensure such things as yields, quality, safety and compliance.
The most common methods of measuring temperature in an industrial process are by either thermocouple or RTD (Resistive Temperature Devices), with each having its own strengths and weaknesses, depending on the conditions being measured.
Neither thermocouple nor RTD sensors are linear in nature. This means that their signal output is not directly proportional to the temperature being measured, with distinct curves associated with different types as per example below.
In addition, no two devices are identical with each having their own unique characteristics. To assist with this, various standards exist, such as IEC60584 and IEC60751, to allow manufacturers to classify each device depending on how far it deviates from the nominal curve as defined in the standard. Thermocouples and RTDs can then be selected based on accuracy required.
In modern temperature transmitters, the ideal curves are stored and interpreted to enable a linear 4…20 mA output from a wide selection of thermocouples and RTD types.
These embedded linearizations are based on the relevant standards for each sensor type and allow for an acceptably linear output – especially across a small range and for applications, which do not require a very high accuracy temperature measurement.
However, there are a number of applications where high accuracy is required and where the standard linearization may not be accurate enough.
- Control of reaction temperatures in chemical manufacturing, critical in ensuring yields/quality and safety
- Custody transfer in the oil and gas industry, where flow compensation is required for accurate costing and billing
- Safety shutdown systems, where fast response to over-temperature is vital
To increase the accuracy and linearity of your measurements you must ensure that you install a high specification temperature transmitter, with the best long-term stability, sourced from a recognised manufacturer.
These transmitters will likely include additional linearization methods through programming, which enable the best sensor/transmitter matching possible for standard sensors, as well as the ability to linearize custom sensors.
One such method is using the Callender Van Dusen equation for RTDs. Sensor specific data is generated for each RTD during calibration, which is then entered during the programming of compatible transmitters resulting in a much-improved linearity.
Regardless of linearization, regular calibration is required to ensure ongoing accuracy throughout the lifespan of the sensor.