The measurements below were conducted at the German ABB Corporate Research Center in Ladenburg. A realistic industrial installation was stimulated for the measurement setup. This was comprised of a pump, a tank and piping with shut-off valves at the beginning and end. The measuring medium used was water and it had a constant temperature near the ambient temperature at the beginning of the measurement.
The water in the tank was heated to a pre-determined temperature using closed valves and the pump was then switched on. At this point, both valves were opened to achieve the fastest possible temperature jump in the piping, which was made of stainless steel (material 1.4307) and had a diameter of DN 80.
As a result, the heated water flowed through the piping with a flow rate of 4 liters per second.
The medium velocity was therefore v ~75 cm/s and guaranteed turbulent flow in the pipe and good heat transfer of the measuring medium to the piping. This speed is well above the speed of v > 10 cm/s, which was determined above to be the speed at which good measuring accuracy is achieved for water during a non-invasive measurement.
In order to compare the measurement, a fast responding SensyTemp TSP321 with thermowell for classical temperature measurement was employed as well as the TSP341-N mounted on the pipe surface. A surface sensor by a competitor was also used in addition to a simpler ABB surface sensor without calculation algorithms.
The TSP321 sensor has these critical features:
- Fast-responding Pt100 Sensor as a thin film resistor, accuracy class AA in accordance with IEC 60751, four-wire circuit
- Welded stainless steel thermowell 1.4571/316Ti which is screwed in, with a 12 mm diameter, tip tapered 9 mm
Measurement Results
The difference between the non-invasive (TSP341-N) and classic (TSP321) temperature measurement before and after the temperature jump are so small that they will not be discussed in further detail here. Only the response times of each sensor will be discussed.
The measurement setup without insulation of the piping and measuring point is shown in Figure 16 on page 15. Measurements were recorded with no insulation and with insulation up to the maximum permissible height of 100 mm at the measuring point (on the extension tube of the TSP341-N).
.jpg)
Measurement setup for comparison of noninvasive and classic temperature measurement.
Insulation has a positive effect on the non-invasive measurement, although this is quite small because the non-invasivetemperature sensor calculation algorithms consider the existing insulation on the measuring point.
The diagrams below show temperature jumps of ~50o from the ambient temperature on the measurement setup.
A temperature jump from 29 oC to 77 oC is shown in Figure 17. No differences between the non-invasive and classic temperature measurement can be detected in this presentation with regards to responsiveness and accuracy.
.jpg)
Temperature jump from 29 °C to 77 °C. Measuring medium: water, piping not insulated.
The same measurement as in Figure 17 on page 16 is shown with a significantly reduced time window in Figure 18. This leaves no doubt surrounding the suitability of non-invasive temperature measurement with a TSP341-N, even with heavily increased temporal resolution.
The response times T90 of the TSP321 and TSP341-N are effectively identical and the measured response time T90 ~0.5 min is consistent with the reaction time of the sensor to a sudden temperature jump (unit step), as determined in the previous chapter.
The TSP341-N runs ahead of the TSP321 by a few seconds during the rise of the temperature up to the time T90.
However, this should not be justified by the fact that the TSP341-N is mounted along the piping in front of the TSP321 and so can react sooner to the temperature change of the measuring medium. The time difference here is less than 0.5 seconds.
Surface measurement on the metal pipe can react more quickly to the beginning of the temperature jump than the sensor that is installed in the thermowell and this highlights the advantage of the non-invasive measuring principle.
Lastly, it is also demonstrated that the effect of insulation at the measuring point is on non-invasive temperature measurement with a TSP341-N is very small. This is because the presence of insulation is taken into account by the non-invasive calculation algorithms as a result of the appropriate parameterization of the sensor.
The piping was fully fitted with ~40 mm thick thermal insulation for the measurement shown below. Both the mounted TSP341-N and the TSP321 were also fully wrapped, up to the maximum permissible height of 100 mm, with insulation material.
Figure 19 shows the measurement with insulation and the same temporal measurement behavior can be detected as in Figure 18 where no insulation was used. As the measurement starts, the TSP341-N reacts slightly quicker than the TSP321. The response times T90 are practically the same for both sensors and tend to be slightly shorter when insulated.
.jpg)
Temperature jump from 29 °C to 77 °C. Measuring medium: water, piping not insulated. Elongated time axis compared with Figure 17.
.jpg)
Temperature jump from 22 °C to 80 °C. Measuring medium: water, piping insulated (~40 mm), measuring points insulated up to maximum height of 100 mm.
Summary
The non-invasive measuring TSP341-N with the non-invasive sensing calculation algorithms integrated into the firmware of the transmitter stands out with a remarkable performance. This is still true when directly compared to the classic temperature sensor which measures in a medium. No significant difference during measurement of a constant temperature were detected.
The TSP341-N is also on par with the sensor with a fast-responding thermowell in terms of reaction time. It can also respond even faster than the sensor built into the thermowell of the classic device to the beginning of a quick change in temperature.
The TSP341-N is equally well suited for applications with or without insulation of the measuring point. The measurement setup used in this example is a realistic stimulation of an in-house installation and the effect of the insulation was positive but minor.
Generally, the non-invasive temperature sensor is suited to any application area in which temperature sensors with thermowells are used. This includes all sectors of industry and heavy industry, for example oil and gas, energy, paper and pulp, petrochemical or chemical sectors. The sensor is especially suited to the chemical industry because it takes into account NAMUR recommendations for example NE107, NE89 and NE24. Global approvals for explosion protection also permit use in atmospheres that are potentially explosive*.
The TSP341-N is particularly suitable for all applications in which an intervention in the process or a thermowell in the process or measuring medium is critical or even undesirable. For example, due to the risk of a possible thermowell break or if a thermowell complicates the regularly required cleaning work.
The sensor is perfectly suited for subsequent and even temporary ‘on-the-fly’ measuring point extensions because of the simple integration into the existing infrastructure of a system (two-wire technology and HART protocol). This is because the system does not need to be shut down and opened for assembly.
A very high level of measuring accuracy is achieved using the TSP341-N with:
- Processes with turbulent flow (often favored by high medium velocity)
- Processes with high medium velocities
- Measuring media with high thermal conductivity
- Measuring media with low viscosity
Examples
Water, water-base liquids and watery solutions as well as saturated steam or fast flowing oil.
However, as was shown in Example 3 on page 10, the temperatures of other measuring media can often be determined with very high accuracy.
*The temperature sensor belongs to ABB’s product family SensyTemp TSP. It is listed in the related type examination certificated for explosion protection as SensyTemp TSP341-N.
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.