1 June 2006
Temperature sensor diagnostics
Troubleshooting old technology devices takes high-tech tilt
By H.M. Hashemian
Here is a review of practical techniques for in-situ diagnostics of problems in temperature measurements with resistance temperature detectors (RTDs) and thermocouples in industrial processes.
These are diagnostic techniques based on the Loop Current Step Response (LCSR) method for the following applications:
1. How to verify the attachment of temperature sensors to solid surfaces in aerospace applications;
2. Detecting secondary junctions in thermocouples as installed in operating processes;
3. Remote testing to identify reverse-connected thermocouples;
4. Verifying adequate insertion of RTDs and thermocouples in thermowells;
5. Separating temperature sensor problems from cable problems;
6. Using RTDs to verify water level in pipes and vessels;
7. In-situ detection of gross inhomogeneity in thermocouples.
We will describe these applications and show how they serve in aerospace and other industries.
The LCSR test involves heating the sensing element of the temperature sensor with an electric current.
For RTDs, a DC current goes to the RTD extension leads to cause internal heating in the RTD sensing element.
A Wheatstone bridge is a part of the RTD test and the electrical test current steps from a few milliamps to 40 to 80 mA to perform the LCSR test.
The transient that results from the internal heating registers at the output of the Wheatstone bridge and then, after digitization, provides the response time of the RTD and also for RTD diagnostics purposes.
For thermocouples, an AC current of about 0.2 to 2 amps works depending on the thermocouple and its installation. The current wires to the end of the thermocouple extension wire for a few seconds and then switched off as the thermocouple output records.
This output undergoes analysis to obtain the thermocouple response time or diagnose thermocouple installation problems.
The LCSR test causes the sensing element of the RTD or thermocouple to heat up about 10°C to 20°C above the ambient temperature depending on the amount of applied current and the sensor ability to dissipate the heat.
Verifying the attachment
Temperature sensors attach to and embed within solid material in a variety of applications and by a variety of means. In many of these applications, there is a fear the sensor can come loose or detach from the solid material, resulting in temperature measurement errors and long response times. Fortunately, the LCSR method works to help determine whether or not a sensor is in good contact with a solid material. The method is useful for RTDs, thermocouples, and strain gauges. The figure below shows LCSR transients from laboratory testing of a thin-film RTD with varying degrees of bonding. It is clear the LCSR signal is sensitive to the degree of bonding between each sensor and the solid material. Here are two examples of this application.
1. In the National Aeronautics and Space Administration's (NASA's) space shuttle, detecting leakage in the fuel lines of the shuttle's main engine uses temperature measurements from surface-mounted RTDs on the fuel lines. During the launch or operation of the shuttle, these RTDs can come loose and render the temperature measurement results useless. After adaptation, the LCSR method verified proper attachment of the RTDs on a shuttle test engine.
2. The lining of the nozzles of solid rocket motors (SRM) for aerospace vehicles is made of a composite material designed to withstand very high temperatures during the firing of the SRM. To verify the performance of a new composite material for SRM nozzles, thermocouples embed in the material to provide transient temperature data during SRM firing. To verify these thermocouples remain installed during SRM firing, LCSR measurements showed they could do that. Here are representative test results before and after firing. This is for a thermocouple that did not remain intact during SRM firing.
Detecting in thermocouples
When welding or other methods involving high heat form the measuring junction of thermocouples, secondary junctions may result from where the two thermocouple wires come together at a location other than at the measuring tip. Thermocouple manufacturing procedures usually include steps to avoid this problem. Nevertheless, secondary junctions occasionally turn up in thermocouples. LCSR can identify these junctions.
See the response-time-measurement results from LCSR testing of 27 redundant thermocouples installed in an engine testing for an aerospace vehicle at a NASA facility. The response time value for one of the thermocouples is nearly an order of magnitude larger than some of the other thermocouples. This observation resulted in an investigation that revealed a false junction in the slow thermocouple. More specifically, this thermocouple had two junctions, one at its tip, as intended, and another a few centimeters above the tip. A secondary junction is typically a problem because it produces an erroneous temperature indication from the affected thermocouple. In the case described above, it also resulted in very poor dynamic performance.
There have been many reports in industrial situations where thermocouples were reverse-connected, resulting in false temperature indications and creating unsafe conditions. In one instance, a reverse-connected thermocouple caused an industrial fire in a nuclear fuel fabrication facility.
It is possible to identify reverse-connected thermocouples with the LCSR method when there is no other way to pinpoint the polarity.
This application of the LCSR test is important because polarities of bare thermocouple wires are not always easy to identify during or after installation at ambient temperatures. In thermocouples such as Chromel-Alumel (Type K), the alumel wire is magnetic. Therefore, for this type of thermocouple, a magnet works to verify the polarity of the wires. In other thermocouples, and where bare wires are not accessible or visible, the LCSR provides a useful tool to identify the polarity of thermocouples.
Technicians identified reverse-connected thermocouples in an assembly on its way to installation in an irradiation test bed at a government laboratory. In this application, the LCSR test was measuring the response times of thermocouples when it also showed the thermocouples were reverse-connected. This discovery was very significant; it saved substantial cost and effort that would have been required to repeat the experiment.
Adequate sensor insertion
The sensing element of an RTD or a thermocouple is typically located as close as possible to the tip of the sensor. Therefore, it is crucial in thermowell-mounted RTDs and thermocouples for the sensor to reach the end of its thermowell. Any air gap, obstruction, or dirt at the sensor/thermowell interface at the tip can cause a significant difference in the dynamic response of the sensor. If the sensor strays from proximity to the tip, the resulting temperature indication from the sensor can also be in error. Therefore, a means to verify the insertion of a sensor in its thermowell is very useful. The LCSR method provides this means. That is, the LCSR test can determine if an RTD or a thermocouple has bottomed out in its thermowell and whether or not there is any obstruction or dirt in the RTD/thermowell interface. Here are the results from three examples of this application.
1. This table shows LCSR test results performed to identify and resolve RTD-in-thermowell insertion problems. These results are from several instances in nuclear power plants where personnel used the LCSR method at cold shutdown conditions to verify proper RTD-in-thermowell installation. The table shows two response time values for each RTD at cold shutdown conditions. The first value—As Found—was the measurement during the initial test of the RTD to verify proper installation. For those RTDs that had an installation problem, a second LCSR test occurred after the problem was resolved. The second value—As Left—is the response time after correcting the problem. The cause of the problem is also in the table along with the corrective action. In some cases, the problem was resolved by simply cleaning the thermowell, and in other cases, cleaning alone would not restore the response time to an acceptable level. In the latter case, it was necessary to replace either the RTD or the thermowell. Note the response times shown at cold shutdown conditions are not the same as response times obtained during normal plant operation. This is due to the effect of process conditions on RTD response time.
2. In some processes, long thermocouples go in long thermowells to reach hot spots or high temperature regions of the process. In such applications, LCSR tests can ensure the thermocouples bottom out in their thermowell.
There are additional results where the LCSR method verified thermocouple installation. A group of identical thermocouples in identical thermowells tested under the same process conditions. These thermocouples should have the same, or very close, LCSR transients, since they are identical sensors tested in identical process conditions. However the LCSR transients for these thermocouples were different. That is, the thermocouples had vastly different response times. The variations in response times were due to differences in sensor/thermowell dimensions and different sizes of air gap at the tip of the thermocouple/ thermowell assemblies.
RTD or cable problems?
When trouble shooting installed RTDs, performing both Time Domain Reflectometry (TDR) and LCSR measurements together can be helpful in determining if a problem is in the RTD or the RTD cable. In particular, the LCSR test can supplement TDR results to show loose connections, moisture in the RTD, moisture in connectors, or vibration of the RTD assembly. An example of how well the LCSR measurement and TDR test work together is in detecting moisture in RTDs. If moisture enters an RTD, both the TDR signature and the LCSR transient are affected. In particular, the response time of an RTD is typically less when there is moisture intrusion and the LCSR test results become noisy or erratic. The TDR test involves sending a pulse through a cable to locate changes in impedance along the cable. If the problem is not in the cable but in the sensor, the LCSR test can help to distinguish between cable problems and sensor problems.
Water level in vessels
In processes where RTDs or thermocouples are in a fluid, LCSR tests can help determine the presence or absence of fluid in a pipe or a vessel. This application of the LCSR method became useful in the recovery efforts during the accident at the Three Mile Island (TMI) nuclear power plant in the U.S. in 1979. More specifically, technicians used the LCSR test on existing RTDs in the primary coolant system of TMI to determine whether there was air or water in the pipes.
In a pipe with three RTDs spaced evenly—120° apart—around the pipe, the average readings of the three RTDs serve as the process temperature. The RTD at the top of the pipe that is in the air will show a much slower response time when tested using the LCSR method. This is just one example of how the LCSR method can work with existing sensors in a plant to determine ambient conditions when there is no other way to make this determination.
Flaws in thermocouples
Inhomogeneities can occur in thermocouples if the thermocouple wire is mechanically or thermally stressed, which would cause the Seebeck coefficient of the thermocouple material to change.
This problem is detectable using several methods. For thermocouples that are not yet installed, the simplest way is to slowly move a heat gun along the length of the thermocouple while monitoring its output. The spike occurs when the inhomogeneous section of the thermocouple is heated. Alternatively, a hot liquid bath works in place of a heat gun. For this test, the thermocouple would slowly submerge into the bath while monitoring the output. There are more sophisticated means of performing these tests on a thermocouple including equipment that produces a large temperature gradient over a short distance through which a thermocouple can transit to detect inhomogeneity.
For installed thermocouples, the LCSR method may have the potential to provide an in-situ means to reveal gross inhomogeneities. It did not work so well for subtle inhomogeneities, and so we may consider the LCSR test as a screening test for inhomogeneity only.
ABOUT THE AUTHOR
H.M. Hashemian (firstname.lastname@example.org) is an ISA fellow and a current or past member of numerous professional societies in the U.S. and Europe. He is the president of AMS (Analysis and Measurement Services Corporation). His recent book is Sensor Performance and Reliability, ISA Press, 2005.
Thermocouples: Two dissimilar wires joined that generate a voltage proportional to temperature when their junction is heated
Thermowell is a closed-end tube designed to protect a temperature sensor from harsh process conditions.
Time-domain reflectometry is a technique for measuring cable lengths by timing the period between a test pulse and the reflection of the pulse from an impedance discontinuity on the cable. The returned waveform reveals many undesired cable conditions, including shorts, opens, and transmission anomalies due to excessive bends or crushing.
LCSR: loop current step response
RTD (resistance temperature detector) is a temperature-sensing device that changes resistance at a predetermined rate in response to changes in temperature.
Seebeck coefficient is the thermo power, thermoelectric power, or Seebeck coefficient of a material that describes how it functions thermoelectrically. It is the derivative (rate of change) of thermal electro motive force (EMF) or voltage with respect to temperature normally expressed as millivolts per degree.
Sensor Performance and Reliability, by H.M. Hashemian
Sensor Performance and Reliability (SP20P), ISA Training
Hazard analysis and risk assessment impact each device: ANSI, ISA, and IEC standards provide few guidelines for choosing sensors.