Low-cost heat exchanger benches for remote operation

Abstract

The COVID-19 pandemic created significant challenges in operating the lab component of undergraduate courses and promoting active learning, with only a short time available to implement alternative teaching methods. In this work a low-cost platform for distance operation and assessment of replaceable bench-scale heat exchangers was developed to provide students an opportunity to observe the transient and steady-state behavior of heat exchangers while unable to access lab facilities. Each workbench had a new material cost of approximately C$5 000. Operation of physical equipment provided students the opportunity to observe non-ideal behavior and compare various heat transfer correlations which may not be seen in process simulators. The developed platform implemented an Arduino microcontroller for low-cost process control. Equipment was seamlessly slotted in to the existing course upon the return to on-campus learning and provided a more stable system when compared to previously existing lab experiments. Most learning outcomes were observed in remote and in-lab experiments and challenges faced in remote operation are highlighted. No statistically significant difference was observed in student performance between students completing lab experiments remotely and students completing experiments in-lab.

1. Introduction

In March 2020 the COVID-19 pandemic resulted in sudden shutdowns of many university facilities and a sudden transition to a fully-remote learning environment was required. In response to this, the Department of Process Engineering and Applied Science at Dalhousie University looked for novel solutions to provide students with an engaging lab environment with the opportunity to collect and analyze real experimental data while knowing classes beginning in September 2020 would be held remotely, with campus closed to all undergraduate students. Additionally, many students were located internationally and were unable to come to Canada because of government restrictions. The Thermal Unit Operations course was selected as a course to pilot the development of remote-delivery lab experiments to engage students and apply knowledge in a realistic scenario. Prior to the pandemic the course covered the fundamentals of heat transfer by conduction, convection, and radiation, as well as multi-phase flow, boiling, condensation, and heat transfer in furnaces and heat exchangers in conjunction with a transport phenomena course. The objective of the lab component of the course was to demonstrate the operation and behavior of heat exchangers and to introduce students to the calculation of heat transfer coefficients.

Prior to the COVID-19 pandemic, many approaches had been taken to delivering unit operation lab experiments remotely. These include distillation columns (Klein and Wozny, 2006), heat exchangers (Selmer et al., 2005, Selmer et al., 2007), and tank level control (Abdulwahed and Nagy, 2013). Of interest here is the work completed by Selmer et al. (2007) which demonstrated the use of a LabVIEW web application and VNC software to control the Armfield HT30 platform with one of three heat exchangers to service multiple institutions with one shared set of lab equipment. This work represents an early approach to remote control and data collection; however, the cost is prohibitive when accommodating multiple student groups simultaneously. The work of Selmer et al. has also been used to develop a guide to the Armfield platform for remote operation (Fridman and Mahajan, 2014). Similar work has been conducted on other commercially-available platforms (Ennetta and Nasri, 2015). As noted by Lang et al. (2007) there is no significant decrease in learning with remote lab experiences in an engineering education setting, however the time investment for development of systems is significant.

An alternative cost-reduction measure for labs employed prior to the pandemic is the use of simulation software developed from fundamentals. Simulations allow investigation by students with reduced resource and time requirements, while also providing a safer alternative (Seifan et al., 2020). Additionally, Ayatollahi et al. (2008) found that simulation from fundamentals can provide a better tool to demonstrate transients with a visual interface. In a heat exchanger specific context, Lona et al. (2000) have created a steady-state shell-and-tube heat exchanger analysis tool as a teaching aide, and Cartaxo et al. (2014) have produced a similar tool for transients in shell-and-tube heat exchangers.

In response to lab shutdowns, several novel approaches were applied to provide students with an engaging experience, such as the use of mixed-reality and in-lab teaching assistants at Imperial College London (Bhute et al., 2022). A novel approach taken in Australia involved purchasing heat transfer equipment and mailing packages to remote students at which point the equipment could be assembled and students could analyze the system, improving learning (Honig et al., 2022). These experiments, however, require a significant time investment for preparation, discussion, and in-lab activities, as well as capital investment. As a low-cost alternative, other institutions have turned to 3D-printing components to investigate the transient heat transfer in non-flowing systems (Larriba et al., 2021), or transitioning to dataset analysis (Bangert et al., 2022). However, these systems do not necessarily give a clear demonstration of the dynamics of a system as would be observed in an operating process plant.

The combination of the idealized dynamic responses seen in process simulators and steady-state responses available to students remotely provides a knowledge gap in real dynamic responses. The thermal mass of a heat exchanger is a factor which can stabilize a process and affect startup and shutdown, however this behavior is not captured in the transient simulator of Cartaxo et al. (2014). Recordings of equipment where students may not see the real-time response of flow rates and temperature profiles also miss this transient behavior. The steady-state simulator of Lona et al. (2000) misses these key transients altogether, along with variables such as fouling, and relies on a selection of one of two correlations which may not be suitable in all cases.

This experiment was prompted by a sudden need for a substitute experiment and project in response to the COVID-19 pandemic. The primary intended learning outcome of these lab experiments was to have students develop an experimental plan in heat transfer, meeting local requirements for engineering accreditation. Secondary learning outcomes of these experiments were to analyze the impact of total flow rates and the ratio of flow rates on the overall heat transfer coefficient and effectiveness of a heat exchanger using the log-mean temperature difference (LMTD) method. Using the calculated heat exchanger effectiveness values students were tasked with identifying unknowns labeled as “Direction 1” and “Direction 2” as co-current or counter-current flow. Students were also expected to locate relevant design parameters from heat exchanger diagrams and datasheets and select appropriate heat transfer correlations for the conditions present and compare their experimental results against theory.

In this work we discuss the design and operation of heat exchanger platforms for a third-year undergraduate unit operations course which can be operated by students remotely or in-person. The platform developed includes flow metering and control, flow reversal, and temperature measurement, with the use of quick-connects for heat exchanger replacement. This work aims to provide a low-cost platform with interchangeable equipment to give students an opportunity to interact with process equipment and record individual experimental data when in-person access is not viable (space limitations, health concerns, available facilities, or others). The use of the Arduino platform and its integration with LabVIEW is also introduced as a cost-saving measure for data acquisition systems which are input-only. After students had interacted with the equipment in 2020 an additional analysis was added where the performance of students who completed experiments remotely was compared against the performance of students who conducted experiments in-lab to assess the impact of remote learning through lab experiments.

2. Methodology and implementation

2.1. Software

Student control of equipment was managed using LabVIEW 2015 with a sample interface shown in Fig. 1. Students were presented with an interface with toggles for opening water supply shutoff valves, setting the direction of flow through the heat exchangers (co-current or counter-current), and setting desired flow rates. Two charts were presented to students showing the current and desired flow rates for the hot- and cold-water streams, and a third chart for the temperature readings at both inlets and outlets. Students were required to identify when the flow rates and temperatures were stable across the system.

Fig. 1.

Control screen presented to remote students showing controls for flow rate and flow direction with profiles for several cases given. Temperatures are shown on a longer period with flow rates over the short term. Note that cold water temperatures are labeled as “Side 1” and “Side 2” due to the flow reversal mechanism which is demonstrated in the red and white lines on the temperature reading chart.

Due to the budgetary constraints a low-cost system with input/output capabilities was required. Prior to the COVID-19 pandemic the available lab DAQ systems installed at each lab station were input-only (thermocouple, analog, and digital input). The Arduino platform, intended for educational environments, can emulate an analog output signal, and can provide a stable 5 V output. Using the LINX package LabVIEW can interface with an Arduino microcontroller to read and write analog and digital signals, allowing communication with process equipment (National Instruments, 2020). The emulated analog output from the Arduino can be smoothed with a low-pass (R-C) filter. Communication between the Arduino and LabVIEW took place over a USB connection at approximately 180 instructions per second, which was deemed suitable for the application.

Remote access was provided to students using a commercial VNC tool, allowing students to use the cursor, keyboard, and display of their personal device (computer, tablet, or smartphone) as if they were in the lab themselves. VNC was previously used by Selmer et al. (2007) for this purpose The VNC tool selected allowed for multiple simultaneous connections, including optional view-only access, to enable group work. Log files were produced to record student access to equipment to address safety and security concerns.

2.2. Hardware

The platform developed for the heat exchanger benches is shown in Fig. 2, with its equivalent process diagram in Fig. 3. Major components are described and approximate costs are given in Table 1. In this system, hot and cold water are supplied via hoses connected to building hot/cold water utilities. Remote shutoffs for flow are provided by solenoid diaphragm valves controlled using solid state relays. Pressure regulators were installed following the initial design to manage a high building water pressure. Flow was metered using Hall-effect turbine flow meters. Proportional solenoid valves were used to regulate flow with control by LabVIEW’s PI controller. The proportional valves each required a controller to use a 0–5 V control signal with a 24 V power supply, and one programmer was required for initial setup of the controllers. A pair of three-way motor-driven L-pattern ball valves were operated in tandem using an H-bridge circuit to set the flow direction for the cold-water stream to be co-current or counter-current. J-type thermocouples were installed at the process connections on the heat exchanger. All components were mounted to a 13 mm HDPE sheet. Each heat exchanger was affixed to a sheet designed to be quickly mounted on the platform. All non-electric components were purchased at local hardware stores, and electric components were purchased from online retailers. The total cost of each bench was approximately C$5 000 for new components, with C$3 500 for major components and C$1 500 for electronics, tubing, and fittings.

Fig. 2.

Integrated system with double-pipe extended surface heat exchanger in 2023 with key components highlighted. The bench measures 102 cm by 71 cm.

Fig. 3.

Process and instrumentation for the developed platform.

Table 1.

Major component P&ID identifiers and descriptions.

Component ID	Manufacturer	Description	Approximate Individual Component Cost (C$)
E-1		Heat exchanger with mount sheet	100–500
FT-1, FT-2	Uniquers	Hall effect water flow meters, 1–30 L/min, 2.25 mL/pulse, 1/2-in. MPT thread, 5–24 VDC	15
FV-1, FV-2, and controllers	Burkert	Proportional solenoid valve with control electronics, 1/2-in. FPT, 24 VDC, Cv = 2.5 m3/h⋅bar	704
FV programmer	Burkert	Proportional solenoid valve programmer	95 (one total)
TT-1 to TT-4		J-type thermocouples, 1/8-in. OD	50
V-1 to V-4		Ball valves/globe valves for manual shutoff, compression fit, 1/2-in. tube	10
V-5, V-6	Granzow	Solenoid shutoff valves, 1/2-in. FPT, 24 VDC, Cv = 9.3 m3/h⋅bar	150
V-7, V-8	Cash Acme	Pressure reducing valve, 10–70 psi, compression fit, 3/4-in. tube	88
V-9, V-10	Valworx	Three-way L-pattern motor actuated ball valve, 24 VDC, 1/2-in. FPT, 0.59-in. orifice, Cv = 15.0 m3/h⋅bar	445
All major components			∼3500

The heat exchangers installed with the lab equipment consisted of a plate heat exchanger with a pressed chevron pattern (Guoli B3–012–10–3.0-H), a double-pipe heat exchanger with annular fins (Alfa Heating FES-70), and a 1–1 configuration shell-and-tube heat exchanger (Alfa Heating SP-55KTi-S), however other bench-scale heat exchangers could be easily installed with appropriate fittings. The plate heat exchanger was used to first introduce students to the online platform, because of its simple construction, and to give a clear demonstration of concepts. The double-pipe heat exchanger introduced flow in an annulus and computational methods to approximate fin efficiency. The shell-and-tube heat exchanger was introduced concurrently with a term project on shell-and-tube heat exchanger design. Students were provided with the model numbers and dimensions of the heat exchangers in a lab manual which detailed calculation procedures for heat transfer coefficients. At the beginning of the semester, students were also provided with photographs of the equipment identifying each component.

3. Results and discussion

3.1. Remote operation results

Over the course of the semester of remote operation, students gained familiarity with the remote lab system and required less frequent assistance from the course teaching assistant. Student groups individually scheduled two-hour time sessions via a shared class calendar for equipment access without supervision and were successful at installing the VNC software and accessing the remote desktop. Several students located in other countries and time zones were still able to access the equipment through the term with an extended day. The first experiment for each group often required additional guidance from a teaching assistant on software and controls however students were quick to learn how the systems operated and full datasets could often be acquired in under one hour. A sample dataset collected by a student group for each of the remote experiments is included as supplementary information.

Sessions were available to students Monday through Friday from 08:30 to 16:30 lab local time (AST/ADT) for 20 sessions per week and two weeks per experiment. This format was enabled by the asynchronous teaching method employed at Dalhousie during the Fall 2020 semester, however it may not be a viable option with regularly scheduled classes. Collaboration with international institutions is a possible option to reduce these challenges. By partnering with a second institution with a comparable enrolment one-half of the needed equipment could be installed in each facility and shared with a morning-afternoon arrangement or across time zones. This arrangement would allow for comparable utility use with reduced capital expenses and storage requirements while allowing students to see the equipment they would be interacting with remotely. This method could also be beneficial in developing countries or locations with limited utilities by implementing a subscription-type system to cover operating expenses with equipment hosted by a partner institution or regional equipment cluster.

Periodic checks were required to ensure students were not accessing equipment outside of typical working hours for safety reasons, which was completed via the software logging. At the beginning of each day, a technician located on the university campus would toggle a power switch located on the electronics housing. The equipment was also powered down overnight to prevent excess water usage. Biweekly access to the equipment by teaching staff was needed to change the heat exchangers over at the end of each experiment cycle and check for leaks. No major or unresolved issues were noted over nine weeks of operation regarding equipment outages or scheduling conflicts. Only about 11% of student groups required teaching staff intervention to manage conflicts due to time zone differences and students facing significant challenges due to the pandemic.

3.2. Experimental analysis

To evaluate the unknown direction of flow the effectiveness for each data point is necessary, however this is a relatively simple calculation in a spreadsheet. Graphically, this becomes a simple test of which case has the higher effectiveness for a given condition or set of conditions, as is shown in Fig. 4. Evaluation by this method also shows the impact of the ratio of flow rates on effectiveness, where in co-current flow a ratio of 1 gives a maximum effectiveness of 0.5, increasing towards an effectiveness of 1 as the ratio of flow rates approaches 0. If different fluids are used the ratio of volumetric flows should be replaced with the ratio of the products of mass flow rates and specific heat capacities.

Fig. 4.

Effectiveness data for the plate heat exchanger showing the use of experimental data to identify flow as co-current or counter-current.

With the direction of flow known the LMTD can be properly determined and used to evaluate the overall heat transfer coefficient. The overall heat transfer coefficient, if examined in a well-planned experiment, can be shown to be dependent on the flow rate of each fluid, as in Fig. 5. Examining the overall heat transfer coefficient in this way also gives a clear demonstration of the sum of resistances to heat transfer, where increasing the flow rate of one fluid while keeping the other constant has a diminished impact on the overall heat transfer coefficient.

Fig. 5.

Experimental overall heat transfer coefficients for the plate heat exchanger in the co-current setup (left) and counter-current setup (right) demonstrating the impact of the flow rates of both streams.

To determine a theoretical convective heat transfer coefficient for the plate heat exchanger, several correlations exist. In this work students were provided with the correlation developed by the APV Corporation (2008), however they were also provided with a more recent work by Nilpueng et al. (2018) which examined plate heat exchangers with pressed chevron patterns, albeit this was simplified so students would not need to measure the angle themselves. For the extended surface heat exchanger the heat transfer coefficient in the annulus was evaluated using a correlation by Schmidt (2010), and students were provided with figures for the efficiency of the fins based on the work of Kraus et al. (2001). To examine heat transfer in the shell and tube heat exchanger students were provided with the correlation developed by Žukauskas (1972).

3.3. Return to the classroom

In the 2020 cohort, consisting of 28 pairs of students, a total of four remotely operated labs were completed by each group (one experiment demonstrating energy balances with conduction, convection, and radiation, and the three heat exchanger experiments). The 2021 cohort, with 14 groups of three students, completed the four remote lab experiments in-lab as well as a previously existing experiment with manual control of flow rates. All grading was completed by the same individual for both cohorts using the same evaluation criteria. Between the two cohorts, the grades were determined to be not significantly different (p > 0.05) for an unpaired two-sided t-test, as is shown in Fig. 6. This result is the same as observed by Lang et al. (2007).

Fig. 6.

Mark distributions for the 2020 and 2021 cohorts. Experiment 1 the plate heat exchanger, experiment 2 is the extended surface double-pipe heat exchanger, and experiment 3 is the shell and tube heat exchanger.

Setup and teardown of equipment was required daily due to the shared nature of the space after returning to campus. The time requirement was approximately 10 min for each station daily to confirm all connections were working as intended. Teardown was completed as student groups finished with the equipment each day. The benches were sufficiently light that they could be carried by one individual between storage spaces and student benches. The three computer-controller experiments were run simultaneously with the manually-operated experiment with no noted decrease in flow rates.

3.4. Learning outcome performance

The primary learning outcome for these experiments was on providing students an opportunity to plan an experiment, collect data, and analyze their data. When given instructions for the experiments students were provided with the requirements that they collect a minimum of 32 data points at different steady states with three variables (two variable flow rates and co-current/counter-current operation). With these instructions student groups were successful in planning and executing full-factorial experiments. Students also were able to directly interact with lab equipment which would not have otherwise been the case in the remote learning environment.

Using experimental data to calculate parameters such as convective heat transfer coefficients, overall heat transfer coefficients, and heat exchanger effectiveness was generally successful for students in both cohorts. Students selected suitable Nusselt number correlations from a set provided and noted that the correlations may not be ideal for their system due to the pressed chevron pattern (plate heat exchanger) and spiral tube shape (shell and tube heat exchanger). Students also were able to identify that the overall heat transfer coefficient was directly related to the flow rates through the system. Identification that the ratio of flow rates was the key parameter in evaluating effectiveness was not observed in most cases. Many students noted that potential fouling and the low quality of the flow meters used created some deviation between their experimental and theoretical data.

One key benefit to using physical equipment over simulators, as we have noted, is the ability to observe transient behavior. It is difficult to qualify or quantify the impact of students observing these transients given the scope of the course and assessment methods, however in this case it served as a detriment to some groups. In periodic observations several lower-performing student groups were noted to wait only for the flow rates through the heat exchanger to stabilize, while neglecting the thermal inertia of the heat exchanger and not noting that the outlet temperatures were changing. Future implementations could include the time-derivatives of the outlet temperatures to show the stability of the outlets and reduce the likelihood that students capture transient temperature profiles.

The use of the heat exchanger effectiveness to identify if the heat exchanger was operating under co-current or counter-current flow was less successful. In a traditional lab environment, students directly interact with equipment and can identify co-current and counter-current flow from their experimental setup, however this is not the case with the reversing system implemented in this work. Many students simply assumed that the first condition case was always co-current and the second was counter-current rather than interpreting their experimental data. Additionally, the extended surface double pipe heat exchanger (Experiment 2) had very poor overall performance which made identification of the operating mode difficult, in particular with the online environment. In the lab students were able to connect or remove thermocouples from the DAQ system to identify the inlet and outlet of the heat exchanger. Given our results on this learning outcome it is recommended that with remote flow reversal the directions be clearly labeled as co-current or counter-current flow to support student learning when applying other methods for analysis. It is also recommended that heat exchangers with low overall heat transfer coefficients not be used as it is difficult to determine if the heat exchanger is operating under co-current or counter-current flow from experimental data with a low effectiveness.

3.5. Planned expansions

Future cohorts of students will continue to operate the computer-controller experiments however the platforms will serve a dual purpose. In future iterations students will have the option to design a shell-and-tube heat exchanger based on Kern’s method, or to design a gasketed plate heat exchanger in various configurations. The students will then model the heat exchangers in CAD/CAM software and plasma cut tubesheets, baffles, and plates, before constructing and testing their prototype to evaluate the performance of the heat exchanger against the design. Student-designed heat exchangers have previously been implemented at Hofstra University (Forsberg, 2004), and student-led construction was found to improve group and individual learning, as well as student construction skills.

The control systems developed through this experiment open new possibilities for low-cost and shared equipment across institutions, as was noted by Selmer et al. (2007). Installation of a limited number of experiments at multiple institutions will provide students the opportunity to observe equipment and its behavior but work independently to collect data on equipment located at their home institution or another facility located internationally with no detriment to learning. These experiments are not limited to heat transfer, but the control systems could be implemented in fluids courses, transport phenomena, and more. At Dalhousie the prospect of introducing remotely-operated experiments to thermo-fluids courses is being investigated.

The use of Arduino systems for process control has been further implemented at Dalhousie for operations such as pump speed control, blower control, and heater control, which historically were operated through interfaces on equipment. Integration with LabVIEW through Arduino has provided better logging and has shifted controls to be located with other control systems managed by NI-DAQ systems. The retrofit cost is much lower and allows for fast prototyping of systems before expensive upgrades are installed. Although the Arduino offers a reduced resolution in measurements compared to the NI-DAQ, this does not impact the quality of results and measurements in many teaching cases.

4. Conclusions

The COVID-19 pandemic created many new challenges in operating the lab component of a course, but also presented an opportunity to develop undergraduate lab experiments and test new technologies. The addition of an Arduino and its integration with LabVIEW significantly reduced the cost of new equipment compared to commercial data acquisition equipment. New designs for remotely-operable thermal unit operations experiments were produced and implemented in a short window of time and were successful at emulating their commercial counterparts at a fraction of the price, with commercially-available components. Each station was capable of metering and controlling flow rates, recording temperatures, and switching between co-current and counter-current flow remotely. The heat exchanger platform developed demonstrates energy balances in heat exchangers, as well as the fundamentals of conduction and convection with the opportunity to use any commercial bench-scale heat exchanger or an in-house design. When students returned to campus the new heat exchanger benches easily fit into the lab component of the course along with a manually-operated system. Students displayed the same level of knowledge of concepts regardless of if they completed the experiment remotely or in-person. The different times required for flow to stabilize and for temperatures to stabilize presented a challenge for some students. Students were able to meet most learning outcomes when completing experiments in either setting. The use of a double pipe heat exchanger with a low overall heat transfer coefficient created challenges in student analyses.

Future expansions on this work will provide students the opportunity to work with CAD/CAM/CNC tools and compare theoretical designs to physical equipment, while providing knowledge of heat exchanger assembly and practical skills. In addition to developing a platform for heat transfer the expansion of remote experiment capabilities offers the potential to increase enrolment by allowing more students to control equipment and collect data in limited lab space.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix

Supplementary data is provided in the form of the data remotely collected by a student group.