UTSR Summer Internship Report Summer Internship Report . Document Status: Preliminary . Final. Proj Code UA Content Code . UNID-Nr . ... PG GT LGT EN MT 1 2 Page: ...

Hard Copies of this procedure are uncontrolled.

Document type No.: XXXXXXXXX

GT Engineering Manual Revision/Date: 8 2015 Issued by: MT12 Engineering

Title UTSR Summer Internship Report

Document Status: Preliminary Final

Proj Code UA Content Code

UNID-Nr

Handling: Confidential Page: 1 of 70 Appendices: 1 Table of Contents (see Page 3 for expanded TOC) Revision Sheet 1 Purpose

2 Scope / Applicability

3 Abbreviations and Definitions

4 Procedure

5 References

6 Appendices

Appendix 1: References etc Released by: signed

Reviewed by: signed

Reviewed by: signed

Reviewed by: signed

Reviewed by: signed

Prepared by: signed Name Org. Unit Signature Date Siemens AG Power Generation Weitergabe sowie Vervielfältigung, Verbreitung und/oder Bearbeitung dieses Dokumentes, Verwertung und Mitteilung seines Inhalts sind verboten, soweit nicht ausdrücklich gestattet.

Zuwiderhandlungen verpflichten zu Schadenersatz. Alle Rechte für den Fall der Patenterteilung, Gebrauchsmuster-oder Geschmacksmustereintragung vorbehalten. Transmittal, reproduction, dissemination and/or editing of this document as well as utilization of its contents and communication thereof to others without express authorization are prohibited. Offenders will be held liable for payment of damages. All rights created by patent grant or registration of a utility model or design patent are reserved.

David Billups Procedure No.: XXXXXXXXX Rev. PG GT LGT EN MT 1 2 Page: 2 of 20

Siemens AG Power Generation Diese Unterlage unterliegt den Handhabungsvorgaben wie auf der Titelseite angegeben. This document is subject to the handling restrictions set forth on the title page.

BILLUPS_DAVID_UTSR_INTERN_REPORT


Revision Sheet

Revision Reissue Date

Section Description Of Change





Table of Contents 1. Objective ...................................................................................................................................... 4 2. Abbreviations and Definitions ....................................................................................................... 4 3. StarCCM+ Version 10.04.009 Validation ...................................................................................... 5 3.1 Han ’84 Case Study ..................................................................................................................... 5 3.1.1 Han ’84 Case Study Results .................................................................................................. 6 3.2 MidFrame Case Study .................................................................. Error! Bookmark not defined. 3.2.1 MidFrame Case Study Results ................................................ Error! Bookmark not defined. 3.3 StarCCM+ Validation Conclusion ................................................................................................. 8 4. CFD Study of IDC Cavities ........................................................................................................... 8 4.1 Purpose of Study ......................................................................................................................... 8 4.2 Summary of GT2015-43499 ......................................................................................................... 8 4.2.1 Heat Transfer Results ................................................................................................................ 9 4.2.2 Flow Characteristics Results ................................................................................................ 10 4.2.3 CFD Modeling Issues .......................................................................................................... 11 4.3 S1 Ribbed Model .......................................................................... Error! Bookmark not defined. 4.3.1 S1 Geometry ........................................................................... Error! Bookmark not defined. 4.3.2 Results .................................................................................... Error! Bookmark not defined. 4.4 V-Shaped In-Line Dimple Array .................................................................................................. 11 4.4.1 V-Shaped In-Line Dimple Geometry ..................................................................................... 11 4.4.2 Results ................................................................................................................................. 12 4.5 V-Shaped Staggered Array ........................................................................................................ 13 4.5.1 V-Shaped Staggered Geometry ........................................................................................... 13 4.5.2 Results ................................................................................................................................. 14 4.6 Case 1: Cavities of the Same Dimensions as S1 Ribs .................. Error! Bookmark not defined. 4.6.1 Case 1 Geometry .................................................................... Error! Bookmark not defined. 4.6.2 Results .................................................................................... Error! Bookmark not defined. 4.7 Case 2: Double-Width Cavities ..................................................... Error! Bookmark not defined. 4.7.2 Results .................................................................................... Error! Bookmark not defined. 4.8 Case 3: Double-Width, Rounded-Edge Cavities ............................ Error! Bookmark not defined. 4.8.2 Results .................................................................................... Error! Bookmark not defined. 4.9 Case 4: Alternating Ribs and Cavities ........................................... Error! Bookmark not defined. 4.9.2 Results .................................................................................... Error! Bookmark not defined. 4.10 Compiled Results ....................................................................................................................... 14 4.11 Succeeding Work ....................................................................................................................... 15 5. Experimental Jet Impingement ................................................................................................... 15 5.1 Experimental Rig Setup ............................................................................................................. 15 5.2 Purpose of Work ........................................................................................................................ 16 5.3 Jet Exit Temperature at Orifice ................................................................................................... 16 5.3.1 Jet Exit Temperature at Orifice Results ................................................................................ 16 5.4 Vertical Thermocouple Rake ...................................................................................................... 17 5.4.1 Vertical Thermocouple Rake Results ................................................................................... 18 5.5 Horizontal Thermocouple Rake .................................................................................................. 18 5.5.1 Horizontal Thermocouple Rake Results ............................................................................... 19 6. Internship Takeaways ................................................................................................................ 20 7. Acknowledgements .................................................................................................................... 20 8. References ................................................................................................................................ 20





1. Objective

The purpose of this internship was to expose interns to gas turbine research opportunities in industry and offer experience in the field. The internship is funded by the Southwest Research Institute (SwRI) and allows prospective students the opportunity for work with one of several industry leaders. Work outlined in the following report was done with Siemens Energy in Orlando, FL. 2. Abbreviations and Definitions

Symbols Unit Descriptions X m Streawise direction Y m Spanwise direction Z m Channel height D m Diameter of the jet Re Reynolds Number r/D Radial Location z/D Jet height Tb K Bulk temperature Q watt Heat transfer rate h w/m^2-K Heat transfer coefficient Tw K Wall Temperature k w/m-K Thermal Conductivity Dh m Hydraulic Diameter Abbreviations SEC Siemens Energy Center UCF University of Central Florida IDC Internal Duct Cooling CFD Computational Fluid Dynamics SwRI Southwest Research Institute HTE Heat Transfer Enhancement





3. StarCCM+ Version 10.04.009 Validation

StarCCM+ is an important tool used at Siemens Energy. All potential new technologies are evaluated through experimental and numerical avenues. The numerical evaluations are done strictly in StarCCM+. Each quarter, StarCCM+ releases a new version of their product. If Siemens Energy believes the features in the new version are beneficial to them, the version is validated to ensure that it provides meaningful and accurate results. To validate, the new version of StarCCM+ is used to solve a case which has a known and trusted solution. In this way, it is possible to compare the known solution to the new solution to discover any changes in the new version. The case study is an internal duct cooling case involving horizontal ribs. To simulate an infinitely long duct, which ensures that a fully developed flow condition exists at the inlet, a periodic interface condition is applied to the duct section modelled in StarCMM+. This allows for faster and more accurate results. The results of the case are compared to results obtained using the existing StarCCM+ code. Historically, the parameters of interest are the residuals, velocity magnitude contour plot, temperature magnitude contour plot, and Nusselt number contour plot for the Han ’84. In the previous version of StarCCM+, the case was run to 30,000 iterations so that was done for this validation as well. 3.1 Han ’84 Case Study

The case study investigated for the validation was the Han ’84 case. Figure 3.1.1 shows the geometry for the case; a rectangular duct with horizontal ribs equally spaced in the streamwise direction.

Figure 3.1.1 Han ’84 Geometry

For each validation, the mesh stays the same so that direct comparisons of the numerical solvers can be evaluated. Below are the meshing statistics for this case. The polyhedral mesher and prism layer mesher are used. No volumetric mesh controls exist. In figure 3.1.2, the mesh quality at the walls is visually presented.





Figure 3.1.2 Han ’84 Mesh Quality

The prism layer mesher adequately captures the features along the bottom wall as well as the boundary layer. According to Table 3.1.1, this mesh is sufficient for validation purposes, but would likely need refinement in an actual investigation. The size of the mesh near the upper surface is quite large and the refinement upward of the ribs may not capture the flow characteristics accurately.

Table 3.1.1 Han ’84 Mesh Quality

3.1.1 Han ’84 Case Study Results

Below, Figure 3.1.3 depicts the residuals obtained in the two versions of StarCCM+ are on the same orders of magnitude, however the fluctuation is much less, almost non-existent, in the solution of the new version. This suggests that the new version has less numerical uncertainty.

Version 10.02.010 Version 10.04.009

Figure 3.1.3 Han ’84 Solution Residuals





The remainder of the parameters show little to no change between the two versions. This is apparent in the comparisons made in Figures 3.1.4, 3.1.5, and 3.1.6 below.


Figure 3.1.4 Han ’84 Velocity Scalar

Version 10.02.010 Version 10.04.009 Figure 3.1.5 Han ’84 Temperature Scalar


Figure 3.1.6 Han ’84 Nusselt Number The important metrics from the above three sets of scenes are summarized below in Table 3.1.2.

Table 3.1.2 Han ’84 Summarized Metrics

Version 10.02.010 10.04.009 Percent Difference

Smooth Wall Temp (K) 342.875 343.042 0.049





Ribbed Wall Temp (K) 337.397 337.256 0.042

Friction Factor 0.0347 0.0348 0.415

Average Nu for Ribs 252.989 254.183 0.472

Average Nu 206.283 206.565 0.137 From the right most column in Table 3.1.2, percent difference, it is clear that the two versions produced nearly identical solutions. With this observation, it is concluded that through the Han ’84 case study, version 10.04.009 is valid. 3.3 StarCCM+ Validation Conclusion

Based on the results obtained with the new version of StarCCM+ for the Han ’84 case study, StarCCM+ Version 10.04.009 was deemed suitable for production use. All results were identical to previously obtained solutions. The lack of change in these solutions is good. Only the residuals showed any deviation from the version 10.02.010 solution. The lack of unsteadiness in the version 10.04.009 solution residuals suggests that the numerical solvers are more accurate than in past versions, thus the numerical error is less in this newest version. 4. CFD Study of IDC Cavities

4.1 Purpose of Study

In experimental literature, dimples have been found to produce 2-3 times the heat transfer seen in a smooth channel, while only incurring a friction factor of 1-4 times that of a smooth channel [1]. Comparatively, ribs produce 2-5 times the heat transfer, but 10-20 times the friction factor. One study in particular experimentally investigated an interesting cavity geometry. It is the intention of this study to numerically model this proposed geometry. 4.2 Summary of GT2015-43499

Researchers at Baylor University created a new dimple geometry that is based on a staggered array of hemispherical dimples. Sets of three dimples were conjoined to produce a rounded V-shaped dimple. This new feature was tested in two arrangements; in-line and staggered. The study performed a steady state heat transfer test and an S-PIV flow visualization test to capture heat transfer and secondary flow characteristics. Figure 4.2.1 shows how the staggered hemispherical dimple array was used in the development of the V-shaped dimple arrays. In Figure 4.2.2, the experimental setup can be seen. The blue diamonds represent thermocouple locations and the green lines represent planes where PIV data was captured.





Figure 4.2.1 Dimple Construction

Figure 4.2.2 Experimental Setup

Both cases were run at Reynolds numbers of 10k, 20k, 30k, and 37k. The bottom surface, containing the dimples, is aluminum. Beneath the aluminum is a heater. 4.2.1 Heat Transfer Results

The steady state heat transfer test involved determining the heat leakage before testing. During testing, the heater was supplied with constant voltage to produce a constant heat flux boundary. The experimental rig would run for up to five hours before data acquisition began to ensure that the conditions within the test section had reached steady state. All heat transfer coefficients are channel averaged and are accompanied by pressure loss measurements through the channel. The HTCs were calculated using an average of the thermocouples as the wall temperature and a linear interpolation of the inlet and exit thermocouple measurements for the bulk temperature. The equation can be seen below.

ℎ =𝑄𝑄𝑖𝑖𝑖𝑖−𝑄𝑄𝑜𝑜𝑜𝑜𝑜𝑜𝐴𝐴𝑆𝑆(𝑇𝑇𝑤𝑤 − 𝑇𝑇𝑏𝑏) (1)

From the HTCs, the Nusselt number can easily be found by equation 2. The Nusselt number results are presented in Figure 4.2.3.





𝑁𝑁𝑁𝑁 =ℎ𝐷𝐷ℎ𝑘𝑘

(2)

Figure 4.2.3 Nusselt Number Results

The trends in Nusselt number show that the heat transfer enhancement for the in-line case increases with Reynolds number, while the HTE decreases above Reynolds number of 20k for the staggered case. The two geometries produce an increase in heat transfer of 1.45-1.75 times the heat transfer of a smooth channel. 4.2.2 Flow Characteristics Results

The flow characteristics results were obtained entirely through PIV data. Secondary flow data was collected for each Reynolds number and at each plane labeled in Figure 4.2.2. The vectors in the PIV results are normalized by the average velocity in the core of the channel. The velocity magnitude shown in the contour plots has been normalized by the average velocity in the x-direction. The secondary flow characteristics for Reynolds number of 37k are shown below.





Figure 4.2.4 Secondary Flow Distributions

4.2.3 CFD Modeling Issues

In the experimental report, there were several key items that were left undefined by the authors. Bulk temperature, wall temperature, heat in, heat out, and average core velocity are not specified. Without this information, it was impossible to replicate the same physical conditions that were observed during experimentation as well as organize the results in the same manner. For these reasons, it was decided to model these geometries in the same conditions as a previously solved model. The model that was chosen for comparison is Siemens proprietary information and cannot be shared in this report. By making this comparison, it is still possible to get meaningful results comparing ribs to cavities. The comparisons of the V-Shaped dimple solutions, at the same conditions as prior internal duct cooling studies at Siemens, are somewhat abstract because the experimental results were obtained at a Reynolds number of 37k while the modeling was analyzed at a Reynolds number of 70k. 4.3 V-Shaped In-Line Dimple Array

4.3.1 V-Shaped In-Line Dimple Geometry

The model geometry is shown below in Figure 4.3.1. This geometry follows exactly the structure of the experimental facility.





4.3.2 Results

This model produces an average Nusselt number of 145.929. The velocity streamlines and scalar plots in Figure 4.3.2 show that the dimples do not create significant mixing with the cooler core of the flow.

Average Nusselt number is relatively low compared to that of typical ribbed models. The velocity streamlines agree with the findings of the experiment which is that the upstream dimples redirect flow to the outer edges of the downstream dimples. From the velocity magnitude profile, it is clear that there is little pressure penalty and not much flow interaction with the dimples. Figure 4.3.3 presents the secondary flow characteristic results. As mentioned previously, the comparison is crippled by the fact that it is impossible to exactly replicate the experimental data reduction because the area considered for the average core velocity by which the data is normalized is unknown. For this reason, the scalar bar ranges from 0 to about 1.3 in the CFD results instead of 0:1 for the experimental results. With this in mind along with the fact that the Reynolds numbers are significantly different, it is still valid to compare the general flow patterns. The CFD results seem to replicate the experimental results pretty well in terms of flow characteristics.

Figure 4.3.1 V-Shaped In-Line Dimple Geometry

Figure 4.3.2 S1 Broken V Ribbed Results





Figure 4.3.3 CFD Secondary Flow Results – In-Line Case 4.4 V-Shaped Staggered Array

4.4.1 V-Shaped Staggered Geometry

The model geometry is shown below in Figure 4.5.1. This geometry also follows exactly the structure of the experimental facility.





4.4.2 Results

The average Nusselt number for this case is 128.454. The velocity streamlines and scalar plots in Figure 4.4.2 show that there is not much interaction between the dimples and the flow.

The velocity streamlines agree with the findings of the experiment which is that the upstream dimples redirect flow to the center of the downstream dimples. From the velocity magnitude profile, it is clear that there is little pressure penalty and not much flow interaction with the dimples. Figure 4.4.3 presents the flow characteristic results. Again, keep in mind the discrepancies between the experimentally obtained results and the CFD results. Here, the contours are much less similar than in the in-line case. 4.5 Conclusions

No valuable conclusions can be drawn here for the V-Shaped dimples. The flow parameters are not similar to the experimental work and the geometrical parameters are not similar enough to the previously completed ribbed channel studies. In the experimental geometry, the channel dimensions are half that of the ribbed channels that were previously evaluated. Also, for both the V-shaped and ribbed cases, the Reynolds number characteristic length is defined as half of the width of the channel, so with a decreased characteristic length in the V-Shaped dimple cases, the mainstream velocity is significantly higher than that of the other cases. Due to these

Figure 4.4.1 V-Shaped Staggered Dimple Geometry

Figure 4.4.2 V-Shaped Staggered Dimple Results





two modeling discrepancies, the conclusions drawn from the V-Shaped dimples model are invalid. 4.6 Succeeding Work

Due to the differences in the V-Shaped modeling compared to the other cases, no valid understanding of the performance of the V-Shaped dimples was obtained. These cases should be adjusted to match the dimensions of the ribbed model and rerun. The experimentally replicated geometry should be run at a few of the Reynolds numbers that were investigated experimentally, without heat transfer effects. This would validate the flow characteristics seen in the experimental work. 5. Experimental Jet Impingement

5.1 Experimental Rig Setup

At the Siemens Energy Center at UCF, a contained single jet rig has been built to investigate the heat transfer capabilities of several geometries. The rig uses an air compressor to feed a plenum. Within the plenum, there are two plates perpendicular to the two inlets to help slow down the flow in order to obtain a true plenum. The bottom of the plenum is an aluminum plate with a hole in, creating a jet. The jet enters the test section and impinges upon a copper plate which is heated from below and insulated on the sides. The test section is enclosed. Around the circumference of the test section, steel mesh is in place to help break up the wall jet after impingement. Thermocouples are in place throughout the rig. The plenum temperature, recirculation temperature, copper temperature at various depths, and wall jet exit temperature are all recorded for each test. The pressure drop from the plenum to the wall jet exit is also recorded. The rig is set up to investigate impingement on a flat plate. Figure 5.1.1 shows the black hose from the air compressor attaching to the plenum at the top of the rig. The graphic also shows the various locations where instruments were used to record pressure and temperature data.

Figure 5.1.1 Impingement Rig





5.2 Purpose of Work

Working with the impingement research team at the SEC involved mostly trouble shooting. All of the data had already been collected. The issues that were addressed focused on the uncertainty calculations for the rig and further understanding of entrainment effects. To investigate entrainment, recirculation temperature needed to be better understood along with the jet behavior. If entrainment was being under represented, the experimental results would under predict Nusselt number. The same is true of the jet temperature. If the jet temperature had previously been under represented, Nusselt number predictions would be low. To investigate these questions, thermocouples were added around the jet exit orifice to determine if there was a difference in jet exit and plenum temperatures. A vertical thermocouple rake was also constructed to capture a temperature profile away from the jet. Lastly, a horizontal thermocouple rake was used to capture the core jet temperature just before impingement and the temperature profile radially outward from the center of the copper plate. 5.3 Jet Exit Temperature at Orifice

To better understand the jet exit behavior, six thermocouples were added near the jet exit orifice. The relative location of each thermocouple can be seen in the graphic, Figure 5.3.1.

5.3.1 Jet Exit Temperature at Orifice Results

This investigation found that there was a difference in plenum and jet exit temperatures. A difference of about 2°C was recorded for several testing cases. Table 5.3.1 shows the cases that were run and the percent change in Nusselt number resulting from the findings.

Figure 5.3.1 Jet Exit Temperature Investigation Setup





Table 5.3.1 Jet Exit Temperature Investigation Results Re=20k Difference [%]

z/D=1 0.78

z/D=6 1

Re=40k

z/D=1 2.25

z/D=6 0.89

Re=60k

z/D=1 1.11

z/D=6 4.08

Re=80k

z/D=1 2.73

z/D=6 5.19

5.4 Vertical Thermocouple Rake

The next test that was performed involved constructing a vertical thermocouple rake to capture a temperature profile at a radial location away from the jet. Figure 5.4.1 shows the experimental setup for this investigation.

Ten thermocouples were equally spaced vertically up to a height of three diameters. The top view shows radially where the thermocouples were located and how they were oriented in the flow.

Figure 5.4.1 Vertical TC Rake Setup





5.4.1 Vertical Thermocouple Rake Results

This investigation was run twice due to a faulty thermocouple halfway through the first set of tests. The results were somewhat inconclusive. The variation in temperature along the thermocouple rake was within the error of the thermocouples; however the difference between the mean temperature and jet temperature was significant. At a difference of about 7°C, this gives good insight into how significant entrainment is for this rig. Below are the temperature data presented in percent difference from the jet exit temperature.

The 2.25% difference seen here is synonymous with the 7°C difference mentioned previously. 5.5 Horizontal Thermocouple Rake

Lastly, a horizontal thermocouple rake was employed to capture the jet temperature just before impingement as well as the temperature gradient radially outward. This will give the temperature difference between jet temperature at exit and impingement, thus revealing the effects of entrainment. Figure 5.5.1 shows the experimental setup for the horizontal rake. Ten thermocouples were spaced one third of a diameter, or 1 centimeter apart. The thermocouple beads were 2 centimeters off the surface of the copper and shielded from radiation by the wooden dowel that was used for the rake. These tests were run at Reynolds numbers of 20k and 80k, and z/D of 3 and 6.

Figure 5.4.2 Vertical TC Rake Temperature Profile





5.5.1 Horizontal Thermocouple Rake Results

The tests captured the jet temperature just before impingement, but the results were inconclusive generally. Since the thermocouple beads were at a height just above the wall jet, the temperature readings radially outward consisted of the core jet temperature, outer jet temperature, and recirculation temperatures, thus the shape of the data is insignificant. It can be seen, however, that at larger z/D, the core jet temperature is higher. This is expected and verifies that entrainment effects are present. The general temperature difference between the Reynolds number of 20k data and the Reynolds number of 80k data can be explained by consecutive testing. The 80k data was run first, thus the chamber temperature increased during testing and was not given sufficient time to cool before the 20k data was recorded. These results can be drawn from the data plots below.

Figure 5.5.1 Horizontal TC Rake Setup

Figure 5.5.2 Horizontal TC Rake Results





6. Internship Takeaways

Through this opportunity, I was able to learn how to use a valuable software tool, gain experience in CFD, and work on new (to me) research topics. Learning StarCCM+ was difficult, but I see the power of it. Using it to validate experimental data gives you more confidence in the conclusions drawn from experimentation. Although I do not think that CFD will entirely replace experimentation, I understand its place in this process and the benefits that it offers. Working on new research topics was enjoyable for me. I like to be involved in a variety of work rather than focusing on just one thing. I think learning and working this way allows me to understand things from various perspectives. One of the most tangible takeaways that I have from this internship is the realization of the need to document everything. Some of the work that I did was insignificant in my mind, but I found that it is still important to document the things you learn in everything you do. In this way, you might save someone else from wasting their time looking into an experiment that is not worthwhile. 7. Acknowledgements

I would like to thank Dr. Andrew C. Nix for exposing me to this program and his help navigating the application process. Thank you to Marco Brunelli and Jose Rodriguez who brought me onto their team and gave me meaningful work to do. Lastly, I would like to thank all of the people that I worked with over the summer; whether you helped me learn StarCCM+ or I worked with you on the impingement rig, I appreciated your help, sharing of knowledge, and willingness to include me in projects. 8. References

[1] Brown, C. P., Wright, L. M., & McClain, S. T. (2015). Comparison of Staggered and In-Line V-Shaped Dimple Array Using S-PIV. Proceedings of ASME Turbo Expo 2015: Turbine Technical Conference and Exposition.

UTSR Summer Internship Report Summer Internship Report . Document Status: Preliminary . Final. Proj Code UA Content Code . UNID-Nr . ... PG GT LGT EN MT 1 2 Page: ...

Documents