Within process manufacturing, temperature is one of the primary considerations in reactions and product separation. There are not many processes where temperature is not a significant factor in controlling product quality. As important as temperature is, measuring and controlling it can prove difficult due to characteristics separating it from other process variables such as pressure and flow. Also, due to the cost and complexity of installing temperature transmitters, often only the bare minimum of temperature measurements required to control the process are installed. This results in an incomplete understanding of the true temperature profile of the process.
This article will look at these difficulties and how to overcome them by installing the correct temperature instrumentation. Adding strategically placed temperature measurements can improve production rates and product quality.
Changing the temperature of a process requires transferring heat into or out of the system. There are few situations where a temperature change can be immediate. For example, heating a pot of cold water on a stove requires time for heat from the burner to transfer through the pot wall and into the liquid. Many factors —burner flame temperature, heat transfer surface area, temperature differential and so forth — affect the speed of the process.
Flow or pressure changes, by contrast, can be realized and measured virtually instantaneously. With temperature, heat has to move through the process and, depending on the heat transfer characteristics, heat may not circulate easily through the mass of fluid.
Thermal processes typically are designed to move temperature in a single direction — either cooling or heating the process medium. Some processes may need both actions at different times. For example, an exothermic batch process may need heating initially to initiate the reaction. Then, it may switch to cooling to avoid excessive heat buildup when the reaction gets going on its own. There is equipment that can be used to heat or cool. For example, a jacketed cooker in a food processing plant can circulate steam or chilled water. In some applications, it may have to do both: to heat a product for cooking and then cool it for packaging. But, relatively few processes are designed to switch back and forth between heating and cooling to maintain a setpoint.
So, in most applications, if a tank’s contents are being heated, the control system must avoid overheating. There may be no mechanism to reverse the action — other than waiting for heat to dissipate naturally, or by adding colder feedstock.
FIGURE 2. This hypothetical continuous thermal process could be controlled with just TT 01, but the addition of data from TT 02 through TT 05 provides mechanisms to optimize steam flow and monitor product mixing.
Heating or Cooling Distribution
There are many ways — ranging from jacketed vessels to electric heating elements — to add or remove heat from a process. A jacketed vessel or internal coil of pipe can heat or cool in a way that spreads out the effort as evenly as possible. This can help speed up the process and minimize stratification with products that do not circulate naturally. By contrast, an immersed electric heating element can add an enormous amount of heat in a very small area, with internal convection or mechanical agitation used to transfer heat through the thermal mass.
Typically, there may be only one or two temperature measurements performed on an asset, with the assumption that the temperature profile is consistent. However, these assumptions can be incorrect. Changes in a vessel such as fouling, blockages or agitator damage can interfere with accurate temperature measurements. Not-yet validated design assumptions can cause unexpected results. Adding extra temporary or permanent temperature monitoring devices to various types of vessels can provide valuable insight into the process. The temperature measuring devices can help identify issues that may be affecting product quality or increasing energy consumption.
FIGURE 3. Temperature transmitters are often packaged as complete units including thermowell, sensor and transmitter, which can be for wired or WirelessHART communication.
Understanding Temperature Sensors
Temperature sensors exhibit their own type of thermal inertia — just like the process itself. A sensing element usually is encased in a stainless-steel sheath (figure 1). The length varies, but the diameter is typically 0.25” (6 mm). Heat from the process must be transferred through the sheath — and whatever insulation may be packed inside the sheath — before it reaches the sensor.
Putting a sensor in a thermowell adds to the delay. The thermowell material and thickness, together with the sensor, will determine the assembly’s response time. The more material, the more time it takes for the heat to migrate to the sensor. This characteristic extends the effect of thermal inertia by slowing the recognition of a temperature change. So, choosing the right sensor configuration and its accompanying thermowell has a major effect on response time.
One type of sensor and transmitter combination forgoes the traditional thermowell and uses an algorithm to determine the process temperature from four datapoints:
- The pipe skin temperature.
- The ambient temperature.
- The thermal properties of the pipe.
- The thermal properties of the process.
Because this does not require penetrating the process to install, it is suitable for adding temperature measurement to running processes without interruption. Thermal inertia is still a factor but, in some cases, the pipe wall might be thinner than a thermowell.
FIGURE 4. Some WirelessHART temperature transmitters can send data from up to four individual sensors on one wireless signal.
Location, Location, Location
The number and location of temperature sensors can have a major influence on the success of an application. Some may be able to function well with a simple setup. Monitoring the temperature of a fluid flowing through a pipe can be accomplished with a single sensor placed in a thermowell if it is located in the right spot and designed to meet the demands of the process conditions.
A classic example of this is the inlet and outlet of a single heat exchanger. Many heat exchangers, however, are installed in a bank of four or more units, where only the inlet and the outlet of the combined bank of heat exchangers are measured. This temperature measurement is not as accurate as readings from the inlet and outlet of each individual heat exchanger.
More complex situations — where temperature profiles within a reactor or other vessel are not so clear — may call for more elaborate solutions with multiple sensors. Consider a batch process where a tank is filled with liquid that must be heated and held at a specific temperature for some period of time. If heated from the bottom and the liquid flows easily enough to allow free convection, where should the sensor be placed? If there is just one measurement point, choosing the location could be a challenge. The natural flow currents will reflect the differentials: Near the heater, the fluid will be hot while at the top, it will be cold.
A continuous process — one where two feedstocks are introduced into a tank and must be mixed and heated to create a new product, for instance — typically will use a tank large enough to provide sufficient residence time for the reaction to take place (figure 2). This type of process also will require a heating element sized to deliver the energy fast enough to keep up with throughput. If there is only one temperature transmitter at the outlet (TT 01), the process will be difficult to control for all the reasons mentioned. But if the unit has multiple sensors (TT 01 through TT 05), it will be much easier to maintain the desired outlet temperature and optimize the mixing process.
Mitigating Cost and Complexity
A complex distillation column might have a temperature sensor at every tray so operators can optimize separation and extraction of specific products. Facilities operating such processes understand the importance of multiple temperature measurements and, therefore, build these capabilities into the control strategy.
Other situations may look at a temperature application and conclude that more sensors would indeed help create a more detailed profile. But, the expense of adding wired infrastructure and more tags to the process automation system is too high, or there simply is no available I/O capacity on the control system.
FIGURE 5. Externally mounted sensors can infer interior temperature by measuring heat transfer through a pipe wall and compensating for ambient conditions.
For these situations, WirelessHART may provide a practical alternative. A WirelessHART network can communicate with many process instruments, but it is particularly advantageous for temperature due to the low amount of data produced by temperature sensors. For example, a temperature transmitter might be configured as a complete assembly with thermowell, sensor and transmitter (figure 3).
For complex situations with multiple sensors, there are transmitters that can support up to four individual sensors on one wireless signal (figure 4). The system reads the sensors sequentially and sends the data to the automation host system, where it is separated into individual readings. This reduces installation and integration costs while providing a high degree of flexibility. The ease of implementing multi-point temperature measurement wirelessly also opens up the potential of temporary installations to troubleshoot or optimize the process.
The WirelessHART gateway can send signals from multiple instruments to the automation host system via an Ethernet or Modbus digital communications link. So, it is not necessary to increase the number of I/O slots on the control system. The growing degree of digital transformation going on in many plants is supported and facilitated by WirelessHART. The ability to add instruments at will via the WirelessHART network can provide data to characterize a process in ways that were difficult to imagine just a few years ago.
One available technology infers the process temperature from the pipe-wall temperature corrected by other factors (figure 5). The sensor is clamped to the pipe and insulated without a process penetration. The pipe size, material and schedule are configured into the transmitter, which then infers the process temperature. Such an instrument can be left in place permanently or moved as needed to take readings in new locations. Once the transmitter is joined to a WirelessHART network, it can be placed anywhere within the overall coverage area.
The ability to place sensors in the most strategic locations to characterize complex temperature profiles is becoming easier as the digital transformation provided by advanced instrumentation and wireless networks advances more completely into process manufacturing at every level. Having critical data makes process characterization and optimization easier and more effective. The ability to pull apart complex issues such as temperature measurement and control is just the beginning.