May/June 2010

Special Section: Temperature/Pressure

The temperature tango

Temperature instrumentation problems and solutions in industrial processes

FAST FORWARD

  • The methods for measuring temperature in industrial processes are relatively simple and based on long-standing principles.
  • The calibration and response time of temperature sensors are subject to a variety of degrading influences including long-term exposure to heat, humidity, and other taxing environments.
  • The LCSR method, cross-calibration, and the noise analysis technique make it possible to verify the calibration and response time of temperature sensors while they are installed in an operating process.
 
By H.M. Hashemian

Despite many advances in electronics and computer technologies, industrial facilities still measure most temperature with conventional sensors, such as thermocouples and resistance temperature detectors (RTDs)-technologies whose basic designs are more than 50 years old. Though so-called smart sensors are now available, the "smartness" of the sensor lies mostly in its electronics and memory and in its ability to adjust its output remotely using digital technology. The same is true for wireless sensors. They usually measure temperature with a conventional sensing device but transmit the information to a remote location with wireless technology.

ThermometerThough manufacturers have made great advances in producing essentially drift-free electronics for temperature sensors, no new sensor technology lies on the horizon promising significantly new drift-free, sturdy sensors that can readily tolerate the temperature, humidity, and vibration environments found in industrial processes. For this reason, today's instrumentation or sensor problems, for example sensor drift, are similar to those of a generation ago. Moreover, questions still linger over how plants can objectively assess the accuracy, response time, residual life, and other characteristics of their installed instrumentation. This article examines some problems involving temperature sensors and offers practical ways to identify, assess, and resolve them.

Temperature instrumentation

Temperature sensors are the most common form of process instrumentation. Plants most often use RTDs and thermocouples, followed distantly by thermistors, to measure industrial temperature. Today, 50 to 60% of all industrial applications use RTDs, 30 to 40% use thermocouples, and the remaining applications use thermistors and other temperature sensors.

Temperature sensor problems

Though RTDs and thermocouples are simple devices, they still suffer problems such as calibration drift and response-time degradation in industrial applications.  For example, a typical problem in temperature sensors is the failure of the seal that keeps moisture from entering the RTD and thermocouple. When this happens, the sensor's insulation material will often degrade or fail, causing temperature indication errors, as well as noise, at the output of the sensor. Furthermore, the moisture in the sensor can accelerate the deterioration of the sensor material. For example, moisture can cause a chemical interaction between an RTD sensing element and the sensor's insulation material. This reduces the diameter of the sensing wire, which increases its resistance and causes the RTD to show a higher-than-true temperature. Instrument and control engineers test for moisture in a temperature sensor by measuring its insulation resistance (IR) using a megohmeter set at 100 volts DC (VDC). The results should indicate IR values in tens of megohms (or higher) at room temperature for a sensor that is dry. If there is moisture in the sensor, the IR could be as low as a few kilo-ohms. Similarly, when there is moisture in the sensor, the IR value will fluctuate so much operators would struggle to register the reading.

fast2

RTDs
More accurate than thermocouples, but not as effective in poor heat-transfer media. Also, not as good for vibration environments, but better than thermocouples in noisy environments.

Thermocouples
Wider temperature range than RTDs, but less accurate and cannot be calibrated after use. Survive better than RTDs in vibration environments, but not as good in noisy environments.


Long-term exposure to heat, humidity, and other taxing environments in an industrial process causes a sensor's seal to fail. Heat can cause the seal to dry out and crack, allowing moisture to enter the sensor. Moisture can also enter sensors through microscopic cracks in the sheath, though this problem is rare in high-quality sensors.

Flow-induced vibration can also cause temperature sensors to fail. Such vibration can cause RTDs and thermocouples or their thermowells to crack, allowing moisture to enter the sensor. Vibration can cause ruptures and shearing in sensors or thermowells, resulting in catastrophic failures.

Problems with RTDs

RTDs can suffer from a wide range of problems affecting their accuracy and response time. These include calibration drift, response-time degradation, reduced insulation resistance, erratic output, and wiring problems, among others.

Almost all cases of dynamic response problems in RTDs are caused by problems at the point where the RTD and its thermowell interface, at the sensing tip of the assembly, specifically, dirty RTDs, dirty thermowells, residue left from using thermal coupling compounds in the thermowell, and dimensional tolerance issues involving the RTD and/or thermowell. Similarly, the extension lead wires that carry the RTD's signal from the sensing element to the transmitter can fail if defective silver soldering has been used to attach the RTD leads to the extension wires protruding from the sensor.

 fast3

A high and stable insulation resistance is important for an RTD's accuracy. Most RTD manufacturers accept industrial RTDs must have an insulation resistance of at least 100 megohm at room temperature (20°C, 68°F) when measured with an applied voltage of 100 VDC. However, if moisture enters the RTD, the IR value can drop to as low as a few kilo-ohms.

In an RTD circuit, there are several transition points from the field to the instrument cabinets in the control room area. Along this path are terminal blocks, weld/solder joints, or splices where loose or bad connections can be found.

EMF, or electromotive force, is a voltage signal that may develop in an RTD circuit if an RTD is designed poorly or contains dissimilar metals. EMF interferes with resistance measurement and can cause as much as 0.5°C error depending on the RTD, the application, and the temperature in which it is used. If EMF occurs, the resistance of the RTD will depend on the measurement polarity. That is, if the resistance is measured with one polarity, then the result will be slightly different than when the resistance measurement is repeated with reverse polarity.

The sensing elements of industrial RTDs are almost always made from fine platinum wire because platinum has superior properties. The platinum element in RTDs is very fragile and can crack or open as a result of vibration, stress, and interaction with other material in the RTD. Usually, the weak points where RTD elements fail are in weld points and places where the element is bent. RTD failures due to open elements are sometimes preceded by erratic behavior whereby the RTD indication experiences large swings, spikes, and random shift.

Platinum wire is more linear than metals like copper and nickel, and it has a much wider temperature range. The sensing elements of some RTDs have experienced corrosive thinning caused by the chemicals that were used to clean the elements when they or the RTD were manufactured. This causes the cross section area of the sensing wire to decrease and its resistance to increase. Thinning of the RTD element can also result from the chemical interaction between the element and the RTD insulation material.

Imbalanced lead wires are a potential problem in three-wire RTDs that are connected to three-wire Wheatstone bridges for measuring temperature. The two wires from across the RTD element that run from the sensing element to the resistance measuring equipment must have equal resistances.

Engineers sometimes dissect RTDs that have failed to determine the root cause of the failure. Reviewing some of these analyses, experts discovered RTD sensing elements usually become open at their weakest points, as one might expect. The weak points are usually where the sensing element is welded to extension wires or where the sensing element is bent, especially when it is bent and passed through openings in the mandrel that supports it. Experts have determined the most common cause of such failures is vibration, especially that caused by flow. As such, RTDs located downstream of pumps or in a turbulent flow path are more likely to fail from flow-induced vibration than those in a calm location. To avoid vibration issues, engineers should use thermocouples rather than RTDs in applications for which vibration is a problem.

 fast5

Sometimes, cracks in the weld joints of RTD sensing elements cause the RTD to fail intermittently. Sudden and erratic changes in RTD resistance are one manifestation of this problem. Engineers can correct this problem, at least temporarily, by remotely applying a few amperes of electric current to the sensor extension leads, a process known as in-situ annealing.

Industrial plants have reported numerous incidents of thermowell-mounted RTDs and thermocouples being plagued by response-time problems. The air gap in the thermowell between the sensing tip of the sensor and the inside wall of the thermowell allows the sensor to expand without contacting the thermowell. This ensures the sensing area is not stressed by the expansion and contraction of the sensor material. Although the air gap is necessary, it has a negative impact on the dynamic response of the sensor. An air gap of even a few micrometers in the tip of a sensor/thermowell assembly can significantly affect the response time of thermowell-mounted RTDs and thermocouples.

Problems with thermocouples

In general, thermocouples are not as accurate as RTDs. This is partly because thermocouples are not normally calibrated individually. Rather, thermocouple wires or a representative sample in a large batch of thermocouples are calibrated, and that calibration is used for all thermocouples in the batch.

Thermocouples can also be accidentally reverse-connected. In such cases, at room temperature, the thermocouple indication could appear to be normal, but as the temperature is increased, the thermocouple will show a negative reading.

Thermocouples, even those properly connected, can have good indication at room temperature but diverge from true temperature as the plant heats up. Thermocouples can also suffer response-time degradation as they age.

 fast7

Temperature sensor solutions

To improve the response time of temperature sensors, plants sometimes apply a thermal compound in the sensor thermowell. This approach is effective, but it has one drawback-long-term exposure to heat can degrade the thermal compound, greatly increasing the sensor's response time. Plants that have older sensors or older thermo-wells have achieved their desired response times by plating the sensor tip with silver or gold and by custom fitting the thermowell. These techniques enable them to keep older sensors or thermowells in service, but they have two main drawbacks. One is the sensor may seize in the thermowell. The other is plating and force-fitting the sensor into the thermowell stresses the sensing element, causing calibration shift or premature failure of the sensor.

Industry has also developed a range of techniques, collectively referred to as in-situ and/or online testing, to verify the calibration and response time of temperature sensors while they remain installed in an operating process. For example, the loop current step response (LCSR) method can easily identify problems such as the inadequate insertion of sensors in thermowells. To indicate the temperature of the intended area and to achieve a reasonable dynamic response, an RTD or thermocouple must reach the end of its thermowell. The LCSR method made it possible for engineers to verify that a thermocouple had indeed reached the bottom of (or "bottomed out") its thermowell.

In applications in which thermocouples are attached to solid surfaces, engineers need to know whether the thermocouple is still in good contact with the body whose temperature it is measuring. Engineers can use the LCSR test in such applications to verify that thermocouples are bonding to solid surfaces.

Like thermocouples, strap-on RTDs on pipes and solid surfaces may lose their bond with the solid surface. The LCSR method detects this bonding problem.

Engineers using thermocouples must avoid inhomogeneity, which occurs when the thermoelectric properties of part of the thermocouple wire changes. The LCSR method can help engineers identify gross inhomogeneities in thermocouples installed in a process. Moving a heat gun slowly along the thermocouple and its extension wires while monitoring its output is an alternative method that plants can use to identify any inhomogeneous section in a thermocouple circuit. The advantage of the LCSR test over the heat-gun approach is it enables engineers to make in-situ tests for homogeneity remotely from the end of the extension leads. Its disadvantage is it may not identify subtle inhomogeneities, and engineers need special expertise to interpret the results.

The industry developed the LCSR method in the late 1970s primarily to measure in-situ the response time of RTDs and thermocouples installed in operating nuclear power plants. Since then, other successful applications have emerged, such as determining the quality of bonding between sensors and solids, verifying sensors are properly installed in thermowells, testing for thermocouple inhomogeneity, and identifying cable and connector problems or moisture in temperature sensors.

The industry also developed a method referred to as cross calibration to verify the calibration and monitor the drift of redundant temperature sensors. Using another method, known as online monitoring, plants can identify drift and calibration problems in other sensors. Processes, such as nuclear power plants, that use several sensors to measure the same process parameter also use redundant temperature sensors to ensure the safety of the plant and to provide better availability and increased efficiency.

When the response time of thermocouples is tested, the noise analysis technique is preferred over the LCSR test, which though more accurate, requires high heating currents (> 500 mA) that can degrade the seal of the thermocouples and/or insulation material. The noise analysis technique is based on the fact that the output of process sensors normally contains fluctuations due to random heat transfer, turbulence, vibration, and other mechanical and thermal hydraulic phenomenon. These fluctuations (noise) can be extracted from the sensor output and analyzed to yield the response time of the sensor. The test involves three steps: data acquisition, data qualification, and data analysis.

Whether through the LCSR, cross-calibration, or noise analysis techniques, appropriate online testing can avoid the problems that affect the accuracy and response time of RTDs and thermocouples.

ABOUT THE AUTHOR

Dr. H.M. Hashemian (hash@ams-corp.com), since founding AMS Corp. in 1977, has worked for almost all of the 104 nuclear power plants currently operating in the U.S. and many in Europe and Asia. He has written two books (translated into Chinese, Japanese, Korean, and Russian) and is the author of 10 book chapters, 10 U.S. patents, and over 250 published papers.

RESOURCES