• By Jack Smith
  • August 31, 2023
  • Features

Starting with the sensor, measuring and controlling temperature is an essential building block of process industries.

Automatic control in continuous processes uses industrial control systems to achieve a production level of consistency, economy, and safety that could not be achieved by human manual control only. It is implemented widely in industries such as oil refining, pulp and paper manufacturing, chemical processing, and power generating plants, to name a few. The “big four” process control parameters are temperature, pressure, flow, and level. Other parameters, such as pH, conductivity, and composition, are indispensably important. However, this article focuses on temperature control only.

Temperature measurement

“You can’t manage what you can’t measure,” Peter Drucker, management consultant, educator, and author said more than a half century ago to apply to the business world. Since then, automation, instrumentation, and controls professionals have “borrowed” it to apply to their own needs: “You can’t control what you don’t measure.”

Temperature is arguably the most extensively measured variable in process industries. If a temperature measurement is not accurate, repeatable, and reliable, it can have a detrimental effect on process efficiency, energy consumption, product quality, and process safety. Each temperature measurement system must be evaluated carefully and designed to satisfy process requirements.

Measurement starts with the sensor

As with any control parameter, temperature measurement starts with the sensor. Typical sensors used for modern temperature controllers include thermocouples and resistance temperature detectors (RTDs). Although there are other sensors associated with temperature, thermocouples and RTDs are the most common.

Thermocouples. Thermocouples are simple sensors, rugged, relatively inexpensive, and easy to use (Figure 1). When certain selected metals (such as Chromel and Alumel in a type K thermocouple) of different composition come into contact, they form a junction that produces a voltage in the millivolt range. If the temperature changes, there will be a corresponding change in the millivoltage produced by the hot junction. To operate, thermocouples must form a closed circuit, or loop. Within this circuit are two primary junctions: the hot, or measuring junction, and the cold junction.

The cold junction must have compensation. Cold-junction compensation “offsets” the junction formed where the thermocouple attaches to the measurement device or controller, which is frequently called the “reference junction.” If the cold junction is not compensated, its presence introduces an error into the measurement that corresponds to the ambient temperature at that point. Cold-junction compensation “corrects” the temperature reading by compensating for the error introduced by the junction itself; it continues to compensate for errors introduced by the reference junction even when the ambient temperature changes.

RTDs. The electrical resistance of an RTD changes according to the temperature it senses (Figure 2). The electrical resistance increases in a predictable manner as temperature increases. Most RTD elements consist of a length of fine platinum, nickel, or copper wire wrapped around a ceramic or glass core. Because platinum is more stable, more linear, and covers wider temperature ranges than the other metals, it has become the industry standard. Platinum RTDs have a high degree of accuracy and repeatability.

According to the book, “Temperature Reference Guide,” by Moore Industries, high-quality RTDs are very stable and rarely drift. If a measurement error is suspected, the problem is typically caused by the extension wire. Error also can be caused by long lead wire runs where multiple junction points are made. It is important to ensure all junctions are tight, as loose connections can be another source of lead-wire resistance imbalance. To eliminate lead-wire induced error, the use of 4-wire RTDs is recommended.

Transducers/transmitters. A transducer converts a physical phenomenon into an electrical signal. In effect, thermocouples and RTDs are types of transducers. In temperature control, the word “transducer” is used infrequently. The use of the term is more common in flow and pressure control. In temperature control, most process control engineers just say, “thermocouple, RTD, or sensor.”

Transmitters convey a measured signal to a control device. The signal coming directly from the sensor is at a low level. The job of a transmitter (Figure 3) is to convert the sensor output into a strong standardized signal and transmit it to a control system. Most temperature transmitters have the ability to work with different types of sensors. They can handle various types of RTDs and thermocouples (Figure 4). Transmitters can perform the required cold-junction compensation when working with thermocouples. Sophisticated transmitters can perform diagnostics on the sensor to determine if there is degradation of the actual element. The transmitter connects to the control system to provide the process variable (PV) measurements.

According to “Temperature Reference Guide,” when using intelligent transmitters or remote input/output (I/O), additional accuracy is gained over transmitters that must use a 4-20 mA output. In addition, maintaining a digital signal to the control system maximizes accuracy. Digital communications avoid errors of converting the digital signal to analog 4-20 mA on both the transmitter end and the control system end. Digital options include HART, Modbus, Profibus, and FOUNDATION Fieldbus.

Temperature control

Accuracy and stability are considered fundamental traits of any process measurement. They are absolutely essential in temperature control. Although temperature control can be accomplished in many ways with many technologies, such as programmable logic controllers (PLCs) or distributed control systems (DCSs), for the purpose of this article, assume a stand-alone single loop controller. This theoretical controller includes a signal processing front end that converts the nonlinear millivolt input from the thermocouple (assuming that’s the type of sensor used) to a usable linear signal, which is compared with a setpoint. The resulting output depends on the amount of error between the measured temperature, or process variable PV, and the setpoint.

Single-loop temperature controllers (Figure 5) are used in small facilities, or for some isolated stand-alone processes. Large continuous process facilities, such as refineries and chemical plants, use DCS (Figure 6) to control pressure, temperature, flow, level, and how they affect the operation of the plant. In some cases, and in some industries, PLCs (Figure 7) instead of—or in addition to—DCSs are used. Sometimes, PLCs control subprocesses via signals obtained from a main DCS.

PLCs have been used to control temperature for decades. It should be noted that if only temperature control is required, a DCS or a PLC is gross overkill. These systems are designed to control entire process plants, or parts of plants. Either of these devices is capable of having hundreds of temperature control loops, as well as flow, pressure, and level.

Closing the loop

Regardless of the type of controller (PLC, DCS, or single-loop controller), the measured signal from the sensor and/or transmitter is compared to a setpoint. The resulting output depends on the amount of error between the measured temperature, or PV, and the setpoint.

In addition to accurately measuring a process, there must be a way to control the amount of heat or cooling applied to that process. The process itself “ties” the system together (Figure 8). The process materials absorb (or dissipate) the energy applied to the process. The sensor detects the temperature of the process and feeds this information back to the temperature controller, which affects its output by applying more or less heating or cooling to the process.

The output of the controller can be relay, voltage pulse, current, and/or linear voltage. It is applied to the heating actuation device. Assuming the temperature control system controls process heat rather than cooling, the process can be heated by gas burners or electrical heating elements, in most cases. Heating actuation devices for gas burners can include modulated gas valves or solenoid valves. There are a variety of heating actuation devices that enable electrical power to be applied to heating elements. These may be pulse-width modulation (PWM), relay, or silicon-controlled rectifier type units.

Control modes

On/off control. On-off control is the simplest form of temperature control. All temperature controllers use a setpoint, which establishes the temperature at which a process is maintained. For example, setting a temperature controller that maintains a food mixture in a vat at 275 degrees F should ensure the temperature of ingredients in that vat is at that temperature.

A controller that uses on-off control supplies an output to increase heat when the process temperature is lower than the setpoint, and no output when the process temperature is higher than the setpoint. It is 100 percent “on” when heating is called for; it is “off” when the process temperature is at or above the setpoint. This arrangement is reversed for cooling control.

Theoretically, the controller switches “on-off” states exactly at the setpoint. However, in reality, this is not practical. If this condition were allowed to exist, the output device would switch on and off so quickly that it would either make the process unstable or ineffective. Another reason is because rapid state changes would quickly wear out the output actuation device.

Rather than have an on-off temperature controller switch on and off exactly at the setpoint, manufacturers provide for an adjustable range around the setpoint. Introducing a small range above, and/or below the setpoint effectively desensitizes the controller to rapid on-off cycling around the setpoint. Some manufacturers call this adjustable range “deadband.” Others refer to it as “hysteresis.” Regardless of the name, it can be effective in stabilizing the operation of an on-off controller if adjusted properly.

In some applications, on-off control produces a cyclical temperature response. The actual temperature of a process could vary from a minimum temperature to a maximum temperature. If the process can tolerate this, an on-off controller may be a simple, inexpensive solution to a temperature control need.

In other applications, the thermal mass of the process may be large enough to resist rapid thermal changes. An example of this is die casting. Typically, die casting machines maintain a reservoir of molten metal at an optimum temperature to allow the machine to operate efficiently. Because of the volume of material and its resistance to thermal change, an on-off controller is adequate for maintaining precise temperature control of this process.

Proportional control. Proportional control takes on-off control a step further. A temperature controller can be proportional with respect to time, or it can be analog-proportional.

Time-proportional controllers apply power to the output as a percentage of a cycle time. If cycle time is adjustable, the time-proportional control divides this cycle time into a percentage of that time. If the cycle time is 10 seconds, and the controller output is at 60 percent, the outputs are energized for 6 seconds of the cycle time. For the remaining 4 seconds the outputs are deenergized. Time-proportional controller outputs can be relay, triac, solid-state relay (SSR), or dc pulse, which drives an external SSR.

Analog-proportional controllers can have voltage or current outputs. Popular output ranges are 0-5 Vdc and 4-20 mA. Analog-proportional controllers are used with SCR power controllers or valve positioning motors.
To set proportional control, the user selects a proportional band. A proportional band is a region above and below the setpoint within which the output of the controller is neither full on nor full off, but somewhere in between. The direction and deviation between the setpoint and process temperatures determines the exact output level.

PID. PID control combines proportional control with two other actions: integral and derivative. Integral action is also referred to as “reset.” It is introduced when a stable process does not coincide with the setpoint. Derivative action is also referred to as “rate.” It is introduced when abrupt or rapid changes in the load affect controller response.

Reset and rate are intended to compensate for temperature offsets and shifts. More often than not, heaters and burners do not match the application. Typically, systems are designed using the “if enough BTUs are good, then more are better” concept. Not true. In a perfect world, heater or burner output would be 50 percent when the process and the controller are at setpoint. In real life, there are usually many more BTUs available than are actually needed. Reset helps to minimize this mismatch.

Rate is used when the process or load changes. Extreme variances in load size and thermal mass necessitate the use of the rate parameter. Because processes behave differently with different loads, the controller must compensate for this difference as if there had been no load change. When used correctly, rate is effective only when there are rapid changes in process physics.

Final thoughts

All temperature measurements begin with the sensor. Measured temperature is compared to a desired temperature in a controller, which provides an output to an actuating device that provides heating energy to elevate the temperature of a process. The process is a necessary part of the temperature control loop. The controller mode can be on/off or some type of proportional control such as PID.

Although much of the temperature control in process industries is performed via DCS, a single-loop controller is sufficient to illustrate the fundamental operating principles of temperature control.

Reader Feedback

We want to hear from you! Please send us your comments and questions about this topic to InTechmagazine@isa.org.

Like This Article?

Subscribe Now!

About The Authors

Jack Smith is a senior contributing editor for Automation.com and ISA’s InTech magazine. He has been a trade journalist for 22 years.