ArticlePDF Available

Thermal Flow Sensor for Non‐Invasive Measurements in HVAC Systems

Authors:

Abstract and Figures

We present a feasibility study on non‐invasive flow rate measurements in heating, ventilation, and air conditioning (HVAC) systems utilizing thermal transduction instead of commonly used ultrasonic techniques. The investigated thermal flow transduction comprises two temperature sensors and a heater, all mounted non‐invasively on the outer pipe surface and, therefore, not disturbing the fluid flow inside. One temperature sensor measures the heater temperature, whereas the other one, mounted upstream of the heater, follows the fluid temperature for reference. The temperature difference (i.e., the heater excess temperature) depends on the fluid velocity and can be used to derive the mean volume flow inside the pipe. In order to visualize and study the temperature field, a finite element method (FEM) model of the system was implemented. Measurements conducted with the sensor prototype confirm the feasibility of this approach.
Content may be subject to copyright.
Proceedings 2018, 2, 827; doi:10.3390/proceedings2130827 www.mdpi.com/journal/proceedings
Proceedings
Thermal Flow Sensor for NonInvasive Measurements
in HVAC Systems
Samir Cerimovic 1,*, Albert Treytl 1, Thomas Glatzl 1, Roman Beigelbeck 1, Franz Keplinger 2 and
Thilo Sauter 1
1 Department for Integrated Sensor Systems, Danube University Krems, A2700 Wiener Neustadt, Austria;
albert.treytl@donauuni.ac.at (A.T.); thomas.glatzl@donauuni.ac.at (T.G.);
roman.beigelbeck@donauuni.ac.at (R.B.); thilo.sauter@donauuni.ac.at (T.S.)
2 Institute of Sensor and Actuator Systems, Vienna University of Technology, A1040 Vienna, Austria;
franz.keplinger@tuwien.ac.at
* Correspondence: samir.cerimovic@donauuni.ac.at; Tel.: +4326222342014
Presented at the Eurosensors 2018 Conference, Graz, Austria, 9–12 September 2018.
Published: 24 December 2018
Abstract: We present a feasibility study on noninvasive flow rate measurements in heating,
ventilation, and air conditioning (HVAC) systems utilizing thermal transduction instead of
commonly used ultrasonic techniques. The investigated thermal flow transduction comprises two
temperature sensors and a heater, all mounted noninvasively on the outer pipe surface and,
therefore, not disturbing the fluid flow inside. One temperature sensor measures the heater
temperature, whereas the other one, mounted upstream of the heater, follows the fluid temperature
for reference. The temperature difference (i.e., the heater excess temperature) depends on the fluid
velocity and can be used to derive the mean volume flow inside the pipe. In order to visualize and
study the temperature field, a finite element method (FEM) model of the system was implemented.
Measurements conducted with the sensor prototype confirm the feasibility of this approach.
Keywords: thermal flow sensor; noninvasive flow rate measurement; HVAC systems
1. Introduction
Systems for heating, ventilation, and air conditioning (HVAC) contribute significantly to the
overall energy consumption of modern buildings. Some analyses show that up to 40% of energy
could be saved by improved control strategies of such systems [1]. In order to optimize the energy
consumption of a building or to detect atypical behaviors, information on the energy flow within the
HVAC system is required. For this purpose, fluid temperature and flow velocity must be obtained at
distinctive points in the hydraulic circuit. These data are usually collected using invasive temperature
and flow sensors. However, modifications of an existing hydraulic system are often not desirable or
possible (e.g., due to legal matters). In this case, noninvasive clampon temperature sensors as well
as ultrasonic (US) flow sensors can be used. Besides their high price, US flow sensors are less suitable
for smaller pipe diameters. Searching for a promising alternative, we have studied the utilization of
lowcost thermal flow sensors to hydraulic circuits with metal pipes.
According to the underlying physical principle, there are three different types of thermal flow
sensors [2,3]:
1. Hotwire or hotfilm flow sensors, which exploit directly the cooling effect of the passing fluid
on a heater. Here, the heater excess temperature (i.e., the difference between heater and fluid
temperature) corresponds to the flow velocity.
2. Calorimetric flow sensors utilize the flow dependent asymmetry of the temperature profile
Proceedings 2018, 2, 827 2 of 5
around the heater. In this case, temperature sensors arranged around the heater are needed for
a detection. Their temperature difference is a function of the flow velocity.
3. Timeofflight (TOF) flow sensors measure the propagation of a heat pulse over a known
distance between the heater and the temperature sensor located downstream of the heater.
Due to high thermal conductivity of metal pipes, the excess temperature along the pipe surface
drops significantly with increasing distance from the heater, even if very high heating power is
applied. Hence, calorimetric and TOF sensors are less suitable for noninvasive measurements on
metal pipes. In the course of this feasibility study, we therefore concentrated on hot wire transduction
and developed a prototype, which was tested on common copper pipes. In contrast to sophisticated
errorprone US setups, this transduction principle suits for arbitrary pipe diameters and allows for
an extremely simple design.
2. Sensor Design
Figure 1a depicts a schematic cross section of the sensor. It consists of a heater and two Pt100
elements, which measure the fluid (T
F
) and heater temperature (T
H
). Their difference (i.e., the heater
excess temperature T = T
H
T
F
) correlates with the mean flow velocity v. For the sensor prototype,
commercially available, miniaturized Pt100 elements were used (Figure 1b). The first Pt100 element
measuring the heater temperature (T
H
) was placed on the copper pipe surface and fixed with cable
ties. Subsequently, a thin copper wire was wound around it forming the heater, which was supplied
by a constant electrical current (Figure 1c). The second Pt100 element was positioned upstream of the
heater at a distance of 5 cm to acquire the fluid temperature (T
F
). Around this temperature sensor, the
same heater structure was built, however, without supplying it with electrical current. This ʺdummy
heaterʺ therefore does not influence a temperature field, but significantly reduces transient response
after sudden variations of fluid temperature, since both temperature sensors feature approximately
the same thermal mass. Finally, the whole pipe was insulated from the ambient.
(a) (b) (c)
Figure 1. (a) Schematic cross section of the sensor; (b) Pt100 elements used for the sensor prototype;
(c) A copper wire (0.2 mm diameter, about 10 m long) was coiled around the downstream Pt100
element serving as a heater. The heater resistance amounts to approximately 11 Ω.
3. FEM Simulations
In order to analyze the flow conversion, a finite element method (FEM) model of the sensor was
implemented (Figure 2a). The dimensions of the simulated Cupipe are the same as for the prototype
(16 mm diameter and 1 mm wall thickness). The sensor surface is considered to be ideally isolated
from the ambient (adiabatic boundary condition). A constant heating power was induced at the
heater and the temperature was recorded at the pipe surface below the heater (T
H
) as well as 5 cm
upstream of the heater (T
F
). Figure 2b shows the simulated heater excess temperature as a function of
the mean flow velocity in the pipe as well as the actual measurement results obtained by means of
the sensor prototype.
Proceedings 2018, 2, 827 3 of 5
(a) (b)
Figure 2. (a) FEM model of the sensor; (b) Heater excess temperature as a function of the mean flow
velocity (simulation and measurement results at 7 W heating power).
4. Sensor Electronics
The dynamic of the excess temperature in Figure 2b (i.e., the signal variation over the desired
measurement range) lies in a range of only 1 K indicating that a high amplification and subsequent
thorough signal conditioning must be applied. The signal conditioning circuit used for this purpose
is depicted in Figure 3. It consists of a Wheatstone bridge and two amplifiers with optional offset
correction. The electrical current through the Pt100 elements amounts to approximately 1 mA
ensuring that self heating effects are minimized. When the offset correction of the first
instrumentation amplifier is turned off, its output signal U
1
correlates to the heater excess
temperature depicted in Figure 2b. For proper signal conditioning, the offset (denoted as offset1 in
Figure 3) must be removed. The offset corrected signal is then normalized to the desired output range
by the second amplifier in order to obtain the best readout and sensitivity.
Typically, the output signal UOUT depends on the fluid temperature T
F
. The exact temperature
dependence is a function of many factors such as fluid parameters, pipe and heater material, or
overall amplification (A
1
A
2
) and must be determined experimentally. It can be accounted for, by
measuring the fluid temperature and applying an appropriate offset at the second amplifier (denoted
as offset (T
F
) in Figure 3). Preliminary evaluation using the sensor prototype (water as a fluid, copper
pipe and heater, A
1
= 1000, A
2
= 7.5, offset
1
4.4 V, heater supplied by a constant current, heating
power approximately 7 W) reveals a temperature coefficient in the range of a few percent per 1 K
fluid temperature change. Thus, a signal correction must be applied in most applications.
Figure 3. Schematic diagram of the signal conditioning circuit. R(T
H
) and R(T
F
) denote Pt100 elements
located under the heater and upstream of the heater, respectively. Adjusting the offset of the second
amplifier, the temperature dependence of the output signal can be taken into account.
5. Results
The sensor prototype was tested in a hydraulic system using water as test fluid. The deployed
water pump allows for mean flow velocities in the sensor pipe up to 1 m/s. The overall amplification
and the offset were chosen such that the sensor output signal fits between 0 V and 5 V in the flow
range of interest (between 0.2 m/s and 1 m/s). Hence, the output can easily be sampled with any
commercially available microcontroller.
Proceedings 2018, 2, 827 4 of 5
The obtained conversion characteristic is depicted in Figure 4a. Applying the inverse function
of the conversion characteristic from Figure 4a, the mean flow velocity inside the pipe can be
calculated. The highest sensitivity is reached in the lower velocity range. For high flow velocities, the
signal saturates and the sensitivity decreases.
In order to estimate the order of magnitude of the time constant, sudden changes of the volume
flow were induced by fast adjusting (approx. 1 s) of a pressure reduction valve. Figure 4b shows a
comparison of the sensor response for a typical clampon US sensor and the presented lowcost,
noninvasive thermal flow sensor. Our prototype of the thermal flow sensor shows good accuracy
but slower response time.
(a) (b)
Figure 4. (a) Measured output characteristic of the noninvasive thermal flow sensor; (b) Comparison
of the sensor response after a sudden change of the volume flow for a typical clampon US sensor and
the noninvasive thermal flow sensor.
6. Conclusions
We presented a feasibility study on noninvasive flow rate measurements in metal pipes
utilizing a thermal flow sensor. The proposed thermal flow transduction comprises two temperature
sensors and a heater, all mounted on the outer pipe surface. The temperature difference of the Pt100
elements (i.e., the heater excess temperature) depends on the fluid velocity. However, a thorough
signal conditioning must be applied to derive the mean volume flow inside the pipe.
The investigated sensor prototype consists of two miniaturized Pt100 elements with a copper
wire wound around them. One coil was operated as a heater, whereas another one just serves as a
“dummy” reducing the transient time. Practical realizations should comprise self adhesive Pt100
elements and heating stripes, which can be easily mounted on the pipe surface.
First measurement results demonstrate the feasibility of noninvasive measurements on metal
pipes in hydraulic circuits by means of lowcost thermal flow transduction instead of highpriced
clampon ultrasonic sensors. The sensor prototype was tested in a measurement range between 0.2
m/s and 1 m/s. By appropriately adjusting the sensor electronics lower or higher flow range can also
be achieved. The general drawback, however, is the saturation of the output characteristic with
increasing flow velocity. Moreover, the output signal is temperaturedependent, so the variations of
the fluid temperature during the measurements must be taken into account.
In comparison with a typical commercially available clampon US sensor, the presented non
invasive, lowcost thermal sensors features an extremely simple design and good accuracy, however,
at the cost of a slower response time.
Acknowledgments: The presented work has been conducted in the course of the research project OptiMAS
(FFG, research grant No. 854641), which is funded within the 3rd program “City of tomorrow” by the Austrian
Ministry for Transport, Innovation and Technology.
Proceedings 2018, 2, 827 5 of 5
References
1. PerezLombard, L.; Ortiz, J.; Pout, C. A review on buildings energy consumption information. Energy Build.
2008, 40, 394–398.
2. Elwenspoek, M. Thermal flow micro sensors. In Proceedings of the CAS ‘99 1999 International Semiconductor
Conference (Cat. No. 99TH8389), Sinaia, Romania, 5–9 October 1999; Volume 2, pp. 423–435.
3. Nguyen, N.T. Micromachined flow sensorsa review. Flow Meas. Instrum. 1997, 8, 7–16.
© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
... In the future, a PIN sensor will be installed in a chiller to further verify the energy savings of the AI assisted HVAC control. smart sensors incorporating a microprocessor chip [6], visual sensors [7], thermal flow sensors [8] and infra-red (IR) cameras [9] have been utilized to detect the thermal comfort, occupancy preference, and to eliminate measurement errors and reduce energy consumption. The detection of information by developed sensors will be fed back to the HVAC system for compensating the error between the set values and the measured ones. ...
Article
Full-text available
The study continues the theoretical derivation from Part 1, and the experiment is carried out at a bus station equipped with six water-cooled chillers. Between 2012 and 2017, historical data collected from temperature and humidity sensors, as well as the energy consumption data, were used to build artificial intelligence (AI) assisted heating ventilation and air conditioning (HVAC) control models. The AI control system, in conjunction with a specifically designed prior information notice (PIN) sensor, was used to improve the prediction accuracy. This data collected between 2012 and 2016 was used for AI training and PIN sensor testing. During the hottest week of 2017 in Taiwan, the PIN sensor was used to conduct temperature and humidity data predictions. A model-based predictive control was developed to obtain air conditioning energy consumption data. The comparative results between the predictive and actual data showed that the temperature and humidity prediction accuracies were between 95.5 and 96.6%, respectively. Additionally, energy savings amounting to 39.8% were achieved compared to the theoretical estimates of 44.6%, a difference of less than 5%. These results show that the experimental model supports the theoretical estimations. In the future, a PIN sensor will be installed in a chiller to further verify the energy savings of the AI assisted HVAC control.
Article
A noninvasive, thermal energy flowrate sensor based on a combination of heat flux and temperature measurements is developed for measuring the volume flowrate and the fluid temperature in a pipe. The sensor is covered by a thin-film heater and clamped onto the outer surface of the pipe. Two types of thin-film thermocouple elements are compared to minimize the thermal contact resistance R″ between the thermocouple and the surface of the pipe. A thin, flexible thermopile heat flux sensor (PHFS) is mounted over the thermocouples. A one-dimensional transient thermal model is applied before and during activation of the external heater to provide estimates of the fluid heat transfer coefficient h. The results are correlated with the volume flowrate Q and the fluid temperature Twc. Several different parameter estimation codes are used to estimate the optimal parameters by using the minimum root-mean-square (rms) error between the analytical and experimental sensor temperature values. The experiments are completed over a range of volume flowrates—1.3 gallons/min to 14.5 gallons/min. Encouraging measurement results give good correlation, repeatability, and sensitivity between the heat transfer coefficient h and the volume flowrate Q with an accurate estimation of the fluid temperature Twc. The resulting noninvasive thermal energy flowrate sensor can be used to estimate the volume flowrate and the fluid temperature in a variety of applications.
Conference Paper
Full-text available
A review is given on sensors fabricated by silicon micromachining technology using the thermal domain for the measurement of fluid flow. Attention is paid especially to performance and geometry of the sensors. Three basic types of thermal flow sensors are discussed: anemometers, calorimetric flow sensors and time of flight flow sensors. Anemometers may comprise several heaters and temperature sensors and from a geometric point of view are similar sometimes for calorimetric flow sensors. We find that depending on the Reynolds number to or three element anemometers may perform as calorimetric sensors (very small Re) while calorimetric flow sensors operate like anemometers in the high Re regime
Article
Micromachining technology has been developed very rapidly in recent years. This technology takes advantage of the benefits of semiconductor technology to address the manufacturing and performance requirements of the sensors industry. The compatibility of micromachining and microelectronics makes the integration of electronics and mechanical elements possible. This covers the need of low-cost, accurate and reliable sensors for industrial and consumer product applications. An important product of micromachining technology is the micro-mass flow sensor which has a history of over 20 yrs. This paper presents a review of the research and development of micromachined flow sensors which have been done in the last few years by international academic and industrial institutions. (C) 1997 Elsevier Science Ltd.
  • L Perez-Lombard
  • J Ortiz
  • C Pout
Perez-Lombard, L.; Ortiz, J.; Pout, C. A review on buildings energy consumption information. Energy Build. 2008, 40, 394-398.