When configuring a new process temperature measurement, many users like to begin by working through a classic engineering question: Should I use a thermocouple (TC) or resistance temperature detector (RTD)? This is a valid question in many situations — and has been the topic of many discussions and articles — but it should not be the starting point for a new application. Other factors should be approached first if an application is to move from simply doing the job to achieving outstanding performance throughout its lifetime.
It is not difficult to find manufacturing plants where some temperature measurement points look like they were cobbled together with whatever components were available in the maintenance shop. These types of designs often fulfill the simply-doing-the-job level, but rarely do they achieve outstanding performance.
So, how should a new installation be approached? What does outstanding performance look like in this context?
Working through a series of questions can help improve the final result of the installation and reduce its lifecycle cost.
FIGURE 2. Modern manufacturing methods allow equipment makers to deliver specialized parts with shorter delivery times.
Why Do We Need This Measurement?
Here is where the process really begins. Whether checking the temperature of bearings on a compressor, liquid flowing through a pipe or the interior of an oven, there has to be some understanding of why the measurement is necessary. Without proper justification to prove the new equipment provides value, it will be difficult to move forward with a purchase. Consider:
Who needs the information?
- Do operations depend on having the reading available?
- Is it a monitoring function?
- Is it driving a temperature control loop?
- Is it supplementing a different instrument and variable (e.g., being used to calculate a mass flow reading)?
- Is it controlling a safety function?
Questions like those can help decide whether a new measurement point is right for your application.
What Is Being Measured?
Knowing what needs to be measured is the next step in finding the best sensor for the application. What is being measured determines the mechanics of how the temperature sensor will be deployed. It also may determine the form factor of the sensor itself (figure 1). Possibilities may include:
- Liquid, gas or steam flowing through a pipe.
- Liquid in a vessel.
- A metallic surface (the outside of a pipe, bearing housing, etc.).
- Interior space (oven, freezer, etc.).
This determines what type of environment the sensor will be subjected to, and what equipment and form factor are needed to ensure safe and proper installation.
What Are the Measurement Requirements?
The individuals or department needing this measurement will have answers to specific questions, including:
- How accurate must it be?
- How wide a range must it cover?
- How corrosive or erosive is the environment?
- How long must the equipment last before replacement is possible?
- How quickly and how drastically can the temperature change?
- How quickly must the sensor respond to changes?
- How and where will the value be sent and displayed?
Once the individuals involved agree on the answers, a series of requirements quickly emerges. These define what has to happen and what hardware may be necessary to accomplish the task. Other external factors when selecting the supplier also may enter the equation such as the prospective supplier’s product quality, reputation, customer support and technical expertise.
Measuring the temperature of liquid flowing through a pipe is different than measuring bearing temperature on a compressor or the interior of a heat-treatment oven. All three would require much different hardware and sensor placement methods, yet it is surprising how often engineers tend to select sensors, mounting devices and signal transmission hardware from the same short list of items. The desire to minimize the number of SKUs in company stores is understandable, but this should not be the primary driver when approaching something as important as temperature measurement.
FIGURE 3. Software makes it easier to determine if a given profile will stand up to vibrations across a range of operating levels.
Matching Tool to Application
Temperature is an important process variable and diagnostic tool in almost every industry. If someone were to catalog all of the temperature sensors, accessories and other hardware available from all suppliers, there would probably be more line items for a unique application than any other type of device. The question is whether engineers and technicians want to make the effort to select the optimal tool for a given application. The answer can be that they are not willing to do so, for reasons such as:
- It is too much work.
- It is too expensive.
- It takes too long to get the parts.
Too Much Work. Some engineers and technicians simply do not have the bandwidth to carry out the necessary research to find products to solve the problem at hand. If an individual is left alone to do the hunting online with nothing but search engines to help, it can be a time-consuming and frustrating challenge. It is important to work with vendors that maintain easy-to-search online resources and provide human support via phone. Local salespeople can come and see applications firsthand and offer knowledgeable support.
Too Expensive. Buying the exact products to solve an out-of-the-ordinary application will invariably be pricey because they are specialized, and vendors do not make them in large enough quantities. While this often is the case it does not have to be.
Modern manufacturing methods offer a higher degree of flexibility for vendors needing to build a product in small quantities or as a special order. Manufacturing setup costs can be lower with sophisticated machining centers taking information from computerized design platforms. The vendor’s application engineer creates a design, and it can be fed to a machining center to create the part with less human intervention. All of these and other advances result in lower costs for end users.
Too Long to Get the Parts. Ordering something special, even if affordable, can take weeks or months for delivery. For many vendors, this is still the reality. Making a special profile thermowell out of non-standard material, if done in some far-flung factory in another part of the world, can indeed take a long time. For special requests, users should make sure the vendor working on the project has local design and manufacturing capacity (figure 2) to compress lead times and provide other benefits such as material certifications and specialized testing.
Working with suppliers who offer temperature measurement solutions designed to address the requirements of specific applications can minimize the effect of these constraints. Engaging a knowledgeable vendor early in the selection process can help to ensure that an appropriate sensor is chosen, which reduces product lifecycle costs and boosts operational performance.
FIGURE 4. Some temperature transmitters can provide diagnostic capabilities as well as alerts for sensor alerts and thermocouple degradation.
A Typical Example
To see how this might work, consider this example: Imagine a large reactor in a chemical plant where multiple feedstocks are introduced, and an exothermic reaction takes place. A critical temperature measurement is at the reactor exit. So, a sensor is placed in the outlet pipe several feet downstream.
The outlet temperature is critical for operators because it is a primary indicator of the reaction itself. The problem for operators is that the reaction is difficult to control effectively because the reading is not responsive enough. The thermowell inserted in the pipe is thick, which was thought necessary to avoid metal fatigue because of the vortex-induced vibration (VIV) due to wake-shedding effects of the flowing stream. Many operators are asking for a more responsive measurement so the process automation system can provide better control of the reaction.
The first issue to address is the thermowell dimensions. Can it be made thinner to reduce response time while still being able to withstand VIV effects? The way to find out is to use software based on the ASME PTC 19.3 TW-2016 standard (figure 3) to determine if a smaller profile thermowell will avoid fatigue failure at all the possible flows the process unit might experience. The results suggest a tapered thermowell with a small diameter tip will avoid VIV problems without the thermal inertia of the larger counterpart.
The sensor sheath and thermowell bore should be matched to ensure the most direct heat transfer from the process liquid through the thermowell and sensor sheath to the actual sensing element. The amount of empty space within the thermowell must be minimized to speed up heat transmission. The back of the sensor sheath can be supplied with a spring to push the sensor all the way to the end of the thermowell so it will make direct metal-to-metal contact.
At the head of the thermowell, it is possible to add a temperature transmitter to convert the output of the sensor to a 4 to 20 mA plus HART signal (figure 4). The sensor can then send diagnostic information to the automation system and warn operators if there is a problem developing with the sensor that might affect the validity of the reading. Maintenance managers might even want to use a dual-element sensor and enable a hot-backup capability in the transmitter so it is able to move from one sensor to the other automatically in the event of a problem. This can avoid shutting down the process to replace a faulty sensor.
The final question to consider involves the sensor itself: Should it be a thermocouple or RTD? By this point, it might not even matter. General guidelines suggest that RTDs are suited for applications where temperatures are moderate and high accuracy is necessary. Thermocouples are better for high temperatures and high vibration levels. In most cases, either does a good job, and if there is a transmitter at the thermowell head, either can probably be accommodated without any problem.
The result of all this effort is easy to see. The operators and automation system get the kind of responsive temperature data they need to control the reactor effectively, which could give them the ability to run closer to production limits, thereby increasing throughput. The cost of the hardware is likely higher initially, but it will require less maintenance and create fewer headaches in the long run. Getting the appropriate equipment for an application without the need for tedious research, long delivery times or excessively high costs can make all the difference for a successful installation.