Are Thermowells Really Necessary?

Non-invasive temperature measurements are a simpler and safer approach to process sensing, but this approach can struggle to meet the performance requirements of many facilities.

Image Credit: ABB Measurement & Analytics

Conventional temperature measurements in the process industry will make use of an electronic transmitter that acquires temperature data from a process. This data is transferred into a controlled system that is connected to an insert, often a thermocouple or an RTD.

This insert is placed inside a series of protective sheets known as a ‘thermowell’ – a pipe inserted directly into a flow, allowing measurements to take place. This technology has existed for many years and has proven to be reliable, accurate and appropriate in a wide range of applications.

Thermowells have a number of distinct disadvantages, however, but the prevalence of the technology has prompted much of the process industry to simply accept and work with these downsides.

Disadvantages of Thermowells

One of the most prominent disadvantages of thermowells is that they must be engineered to exceptionally high standards. Thermowells are inserted into processes that utilize aggressive media, and this media is potentially hazardous in the case of a leakage or anything escaping the piping.

Thermowells are engineered against recognized standards that relate to known process conditions, but processes are prone to often unexpected changes.

For instance, when working with upstream oil and gas, there is an assumption of what is coming from the source, but over time this can change due to a wide range of external factors.

Sand could be inadvertently extracted, which would start eroding the thermowell, and because this is occurring underneath the pipe surface, organizations may not be aware of this until a catastrophic failure occurs. Careful design and precise engineering of thermowells cannot always accommodate these process changes.

The Untapped Potential of Accurate Surface Measurements

Surface measurements are widely considered as lacking the sensitivity to provide a viable option to the traditional thermowell-based approach to temperature measurement. Technology has existed for a long time.

However, that is more than capable of meeting the performance requirements of a wide range of processes in a non-invasive manner, provided this is properly implemented.

Traditional surface measurements using skin-type sensors are one such option. This approach sees skin-type sensors placed along the pipe, accurately measuring the surface temperature of the pipe.

This approach is not currently widely used, however, and many people consider it to be inaccurate, particularly due to its sensitivity to ambient temperature. Accuracy can be improved in a number of ways, therefore increasing the viability of this approach.

The use of heat conductive paste between the sensor and pipe is one option, or the sensor could be insulated to minimize heat losses that occur via connectors such as a heat antenna. Using proper planning, expertise and the right sensors, it is possible to overcome the barriers associated with surface measurement.

When done properly, non-invasive temperature measurement is a simpler, safer approach to measuring process temperature without compromising on performance.

This approach, therefore, removes the need for complex engineering of thermowells, the need for shutdowns to perform maintenance, and significantly reduces the risk of leaks due to there being no holes or weak points in the piping.

Employing a non-invasive approach to temperature measurement requires some understanding of the physics involved, but the application is relatively straightforward.

Ensuring the sensors are robust, sensitive, and efficient is essential to this process, but much of this sensor hardware is the same as the instrumentation used inside thermowells – the inserts have simply been removed and attached to the outside of the pipe.

The hardware of the thermowell has been adapted to use a physics- or a model-based approach.

Understanding the Physics of Process Measurements

Typical processes will involve a medium flowing through a pipe. Understanding this from a physics perspective first involves defining the boundary conditions, ensuring that equations are as accurate as possible.

The goal is to ascertain the bulk medium temperature, and a typical thermowell will take into account an insertion length that can vary based on certain parameters.

It is important to note that the traditional approach is not measuring the medium temperature – it measures the wall temperature of the thermowell. Instruments are designed to ensure that wall temperature is consistent with what is occurring in the medium.

It is also important to note that often, the thermowell is the only reference as to the temperature in the pipe - there is rarely another thermowell in place to verify these measurements. If the insert is too short or there is debris in the sensor, for example, there is no way of knowing that measurement is affected.

Predicting the medium temperature using a physics model involves determining density, viscosity, flow rate, pipe material, and pipe dimensions, then using these parameters to accurately calculate the medium temperature based on the acquired surface temperature.

An interesting phenomenon occurs when a mixture is flowing through a turbulent regime in a metal pipe – under these circumstances, there is virtually no temperature difference between the medium and the surface temperature.

A large number of applications in the process industry work with liquid-like mixtures flowing through a metal pipe in a turbulent regime, meaning that in these circumstances, there may be no need for a thermowell at all.

While skin temperature measurements of the outer pipe wall have previously been considered inaccurate and slow to respond, physics suggests that the outer surface temperature of a pipe is generally trustworthy.

Where this has not been the case, and surface temperature measurements have proven to be inaccurate in the past, it can reasonably be suggested that this was primarily a result of the sensors in use not being effective enough.

ABB’s development team have developed an innovative solution to this issue, essentially taking the insert from the thermowell and recontacting it on the surface of the pipe. A second sensor is placed in the vicinity of the first sensor to measure ambient temperature in the vicinity of the sensor.

A thermal model of the device has been developed, allowing the surface temperature of the pipe to be mapped against the ambient temperate, providing a much more robust and reliable measurement.

As mentioned, a liquid-like mixture flowing into a turbulent regime in a metal pipe will lead to there being virtually no difference between the surface temperature and the medium temperature.

ABB has conducted a number of tests to verify this assertion, for example, by heating up a tank and instantaneously pumping hot water into a pipe to simulate an instant step change of a turbulent flowing fluid.

Measurements were performed using a traditional thermal insert - PT100 sensors stripped down and glued to the surface of the pipe with a very thin layer of glue.

These sensors were then covered with epoxy to provide insulation. It should be noted that this was a temporary fixing for the purpose of the experiment rather than a permanent installation.

The water was heated to around 70 °C the tank before being pumped into the pipe, ensuring a turbulent regime. Measurements over time confirmed that this non-invasive approach was able to match the thermowell’s performance, comfortably outperforming existing surface measurements.

Application in Practice

As well as validating a range of theoretical models, simulations, and experiments, evaluations have also been conducted on existing industrial applications in order to verify the performance of this approach to temperature measurement.

Tests were conducted using pipe diameters ranging from DN 40 to DN 2500, and it was confirmed that one sensor variant could accommodate all these size ranges. This offers considerable benefits over traditional thermowells, which would generally need to be custom engineered to be mounted inside a specific size of pipe.

This is particularly useful when using more complex piping systems such as closed-loop cooling systems.

Metal pipes remain the most common, but sections of the industry are looking at piping made from other materials. ABB’s current system can work with medium temperatures between -40 °C and 400 °C, and with medium density above 50 kg per m3.

The additional flexibility afforded through the use of non-invasive measurements has been projected to result in a reduction in installation costs of up to 75%.

Almost every process could potentially benefit from non-invasive temperature measurement. In one real-world example from the upstream oil and gas industry, a customer was working within a highly aggressive environment, managing pipes that had to transport oil, gas, sand and water at very high velocities.

This medium had the potential to erode thermowells in a matter of weeks, but the company had investigated the use of skin surface temperature sensors before and discovered that these were too sensitive to ambient conditions.

It is virtually impossible to use models to provide accurate predictions for multi-phase flows, but the rules around liquid-like mixtures, metal pipes and a turbulent flow still apply. The customer in this example used a pressure-tested thermowell, specifically for this test.

The offset of temperature between the two measurements is around 1.7 °F due to a minimal layer of paint on the pipe. This offset was found to be fairly consistent, assuring the customer that repeatability was acceptable.

Ambient temperature around this installation varied between 55 °F and 90 °F due to the surrounding desert environment, with these day/night cycles being relatively clear and predictable in the data.

It is also possible to offset this by using appropriate insulation. Overall, the surface measurement solution was found to be sufficiently comparable to the traditional thermowell approach and much for suitable for working with this aggressive medium.

A global chemical producer in Germany has been working with ABB to test the company’s sensors under a range of different conditions, and the customer is now looking to replace a series of invasive sensors with non-invasive alternatives.

One example involved a metal DN 300 pipe with a liquid flowing through it. Surface sensors were tested alongside the existing invasive approach, and a difference in the accuracy of 0.14 °C was established, with good repeatability.

Insulation was also required in this instance. This example also showed that a thermowell was not required to acquire accurate process temperature measurements.

ABB has also developed models for steam that have been validated. A distributed energy supplier in Germany was looking at whether it was possible to measure temperature at distributed points without having to shut down their system while also potentially increasing the accuracy of these measurements.

This sort of system would require the use of multiple thermowells, which would be costly and inefficient. However, it could accommodate a potentially infinite number of non-invasive sensors, providing high levels of confidence in any measurement data collated.

The company evaluated a pipe that was transporting steam at approximately 250 °C. The accuracy difference between the two sensors was found to be 0.24 °C with good repeatability.

The most important thing to consider in each of these application examples is that even when using a thermowell, offsets or inaccurate data can present themselves. It is more important to explore the best way of capturing the dynamics of a process.

This is possible with robust surface measurement technology, meaning that thermowells could soon become a thing of the past.

This information has been sourced, reviewed and adapted from materials provided by ABB Measurement & Analytics.

For more information on this source, please visit ABB Measurement & Analytics.


Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    ABB Measurement & Analytics. (2023, April 06). Are Thermowells Really Necessary?. AZoSensors. Retrieved on April 23, 2024 from

  • MLA

    ABB Measurement & Analytics. "Are Thermowells Really Necessary?". AZoSensors. 23 April 2024. <>.

  • Chicago

    ABB Measurement & Analytics. "Are Thermowells Really Necessary?". AZoSensors. (accessed April 23, 2024).

  • Harvard

    ABB Measurement & Analytics. 2023. Are Thermowells Really Necessary?. AZoSensors, viewed 23 April 2024,

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Your comment type

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.