P13452: Dresser-Rand Compressor Characterization
/public/

# Build, Test, Document

## I. System Vibration Isolation

Pre-Fab Design

Construction Phase

Final Product

Resulting Vibration Reduction

## II. Thermosyphoning Characterization

• Thermosyphoning is a process occurring in a system in which a coolant is circulated by free convection and buoyancy induced flow, due to some temperature differential.
• Need to quantify the amount of cooling and flow occurring through the previously designed thermosyphoning system, and develop an estimating tool for industry application.

Previous system, below.

The mathematical model that was developed uses the derivation of the convective heat transfer coefficient, h, heat rate, q_t, and the volumetric flow rate, Q_dot, based on thermal physics, natural circulation principles, heat transfer and fluid dynamics.

The assumptions made for this analysis are as follows:

• Itemized text
• Incompressible
• Negligible viscous dissipation
• Uniform thermal-fluid properties (k, h, etc..)
• Closed loop
• Constant diameter
• Uniform heat flux at heat source
• Fully vertical, internal flow through a pipe
• Fully-developed, Laminar flow
• One dimensional radial conduction through fins
• Negligible radiation exchange with surroundings

The resulting schematic for the vertically oriented theoretical system appears as follows:

The average convective heat transfer coefficient is then, from the dimensionless parameters Gr, Pr, Ra, and Nu:

where k is the conductivity, D is the pipe diameter, beta is the volumetric thermal expansion coefficient, Ts is the mean surface temperature, Tinf is the ambient temperature, nu is the kinematic viscosity, and alpha is the thermal diffusivity and Gr, Pr, Ra, and Nu are defined below.

where n=0.25 for a vertically oriented pipe with Laminar flow.

Next, the heat rate equation was derived using extended surfaces analysis and employing the newly calculated heat transfer coefficient:

where h is the average heat transfer coefficient, N is the number of fins in the arrangement, Af is the fin area, At is the total exposed area, eta is the fin efficiency, and Ts and Tinf remain the mean surface and ambient temperature respectively. Af and At are defined below.

Finally, to determine the volumetric flow rate, we first solve for the mass flow rate using simple heat transfer principles:

We then use the density of the cold fluid to find the volumetric flow rate:

where Cp is the specific heat capacity at constant pressure, and rho is the density of the fluid

• Following the development of a successfully working model, to provide an estimating tool for use in industry, a macro-enabled Microsoft Excel spreadsheet was developed and coded using Visual Basic (VBA) programming, using the user interface (UI), displayed below:

• In VBA, inputs were called, calculations were performed, and results were output behind the scenes, and the user has but to click a button on the UI.

The fully functional document is stored in the index for the home node.

## III. Thermosyphoning Bench Experiment and Reconstruct

--Flow Meter

• The thermosyphoning flow meter, implimented by P12452 outputs a flow of zero ml/min while system is attempting to reject the systems heat.
• This implies the system is not flowing, thereby not effectively transferring energy from the cylinder head to the atmosphere.
• To verify the thermosyphing system is producing no flow, the pre-existing flow meter is verified through experimentation.

--Bench Experiment

• Using the heat exchanger elements from the compressor a bench thermosyphing unit is fabricated. To replicate the heat source an OMEGA heat element wrap is used. The heating element is capable of regulating temperature through the use of a built in thermostat. This is used in different experiments to see the heat rejection response of the system.
• The system is constructed of black iron NPT pipe and consists of 4 thermocouples strategically placed. Figure 3.1 shows the location of the thermocouples, which are read through an OMEGA thermometer. The system layout is flexible with the heat source able to be moved easily. The hand held thermometer data is captured in real-time by the OMEGA SE374 program

Figure 3.1

## IV. Continuous Health Monitoring

-- Health Monitoring is the process of continuously measuring and assessing the overall wellbeing of a system.

• Prior to the proposal of our system a system had not been proposed or built for the Dresser Rand compressor here at RIT.

--Our design was made keeping a few key points in mind:

• Must continuously monitor.
• The data capturing system must be simple enough to be understood by future groups and researchers.
• The design must be simple to build.
• The system must be easy to disassemble and reassemble without having to take apart the compressor extensively.
• As stated in MSD1 the system which the decision was made to go forward with was the Bentley Nevada proximity sensor design. Bentley Nevada donated the sensors to RIT. The system came with not only the sensor but the conditioning monitoring software.
• The Final Mount Design:

• It was proposed that a program should be written in Labview which could be integrated into the Labview program which monitors the rest of the compressors vitals other that the rod drop.
• Multiple mounting designs were considered to mount our proximity sensors.
• The final design was simple and easy and rather than being a complicated design, it utilized three of the four studs which mount the compressor packing.

--The studs were machined and tapped for ¼ NPT threads.

• This was done using a lathe and was extremely simple to do. Once they were all machined the studs were inserted back into the compressor casing and torqued back down to the proper torque.

## II. Thermosyphoning Bench Experiment and Reconstruct

--Flow Meter Calibration

• The flow meter was removed from the compressor system and assembled in line with a submergible water pump wired directly to a DC power supply.
• The pump runs at a steady state condition so a volumetric flow rate is attainable by filling a known volume over a given time. Forcing the flow meter with the DC pump generates an output signal, which is acquired by the preexisting National Instruments DAQ.
• The data collected is compared to the experimentally measured volumetric flow rate to verify accuracy.
• Each flow rate experiment shows the flow meter to be accurate within 2% error.

--Bench Experiment

• As the thermosyphoning system was under testing, the paddle of the flow meter was not rotating. The flow was not strong enough to generate movement of the paddle wheel, so an ultrasonic flow meter was borrowed from Dr. Day to measure a frequency output which then would be conditioned to a flow rate.

- Experiment 1

• The heat wrap was on the top of the left column, which generated the following results, shown in Figure ****. As seen, the upper left temperature, where the heat wrap was located, had no significant cooling throughout the entire experiment, which means in a real scenario, the cylinder would over heat. The flow rate was negligible as well

Figure 3.2

- Experiment 2

• The heat wrap was placed on the lower left corner of the system and as a result, Figure 3.3 was generated.

Figure 3.3

• This displays the transient performance of both the temperature and the flow rate as steady state is reached. 45 minutes into the test, the thermostat was reduced to 250 degrees Fahrenheit. As seen in the figure, the flow correlates appropriately with the heat generation temperature. The flow decreases from approximately 280 ml/min to 260 ml/min while the temperature of the heat wrap goes from 140 degrees Fahrenheit to 125 degrees Fahrenheit.

The plot also demonstrates the system under shut down. About 65 minutes into the test, after steady state at 150 degrees Fahrenheit was reached, the thermostat was turned off. Once again the flow correlates appropriately with the shutdown trend temperatures of the test rig.

- Experiment 3

• Figure 3.4 represents the identical test setup as experiment 3, except temperature and flow were varied from a preexisting steady state temperature of the heat wrap being set at 400 degrees Fahrenheit.

Figure 3.4

• The temperatures and flow rate decrease appropriately when the heat wrap thermostat is adjusted to 250 degrees Fahrenheit, after 45 minutes, and 150 degrees Fahrenheit, after 70 minutes. After 105 minutes, the thermostat was adjusted back to 400 degrees Fahrenheit where steady state was reached and held until 180 minutes. At 180 minutes, flow was purposely restricted, consequently observing a decrease in flow and a spike in all of the temperatures. After 20 minutes of restricting flow, the restriction was free and flow was attained throughout the system, ultimately returning the temperatures back down to steady state.

--Compressor Thermosyphoning Rebuild

• From the collected emperical data, the heat source must be below the finned heat exchanger to perform correctly. a quick and inexpensive redesign of the system on the compressor was necessary. With the spare heat exchangers from the group that had built the previous thermosyphoning system, the construction of a new system on the compressor began.
• The thermosyphoning can be seen in Figure 3.5. Figure 3.6 displays an eight hour experiment with the compressor running under 100 percent load. The system was restricted of flow around 350 minutes and as a result, a spike in cylinder head temperature and cylinder temperature were observed. After about 15 minutes the system restriction was removed and the temperatures returned to steady state. Overall the new thermosyphoning system was a success in cooling the cylinder and cylinder head to the appropriate temperatures.

,

Figure 3.5, Figure 3.6

- Concluding Observations

• It is easily observed that the bench experiment is a true functioning thermosyphoning system, and cools the heat source well.

## III. Thermosyphoning Characterization

- Concluding Observations

• In case 1, TH=115.8°F, Tc=99.9°F, and the flow rate is ~290 mL/min,
• Case 2, TH=111.3°F, Tc=98.0°F, and the flow rate is ~264.8 mL/min,
• In case 3, TH=118.0°F, Tc=102.0°F, and the flow rate is ~217 mL/min.
• For these values of temperature, the model yielded flow rates of 271.89, 263.38, and 276.65 mL/min respectively. That is an average relative percent error of only 6.303%. For this low margin of error, this model is deemed successful. Likely sources of error are outlined below.

- Sources of Error

The slight variations in flow rate between theoretical and empirical values can be attributed to a number of factors.

• First, several assumptions are made in the analysis that are vastly untrue. Notably, the system is not composed of entirely vertical piping. Not only would the gravitational effects differ, but also some minor head losses are associated with the changes in geometry due to entrance flow, fittings and reductions or expansions in flow path.
• The possibility of transition or turbulent flow at areas of impulse, such as the heat flux at the heat source, may occur altering actual values.
• Approximate averages were used in comparison of results, so some human error may be attributed when reading the graph.
• The model uses the full length of pipe in the extended surface portion of the model versus the length of actual finned tubing. This discrepancy should have negligible effect since, ultimately, it just assumes the fins are extremely generously spaced; however some error could result from this.

## IV. Continuous Health Monitoring

• Compressor results while not under load:

• Compressor results while under load:

• The Mount and Probes Installed