Test Data AC2-10: Difference between revisions

From KBwiki
Jump to navigation Jump to search
m (Dave.Ellacott moved page Lib:Test Data AC2-10 to Test Data AC2-10 over redirect)
 
(8 intermediate revisions by 2 users not shown)
Line 14: Line 14:
|+'''Table 1:''' Engine operating parameters for the base case.  
|+'''Table 1:''' Engine operating parameters for the base case.  
|-
|-
|RPM||800 &plusmn;7 min<sup>-1</sup>
|RPM||800 &plusmn; 7 min<sup>-1</sup>
|-
|-
|Cyl. Head, coolant temp. ( <''T<sub>eng</sub>''> ) || 60 &plusmn;1&deg;C
|Cyl. Head, coolant temp. ( <''T<sub>eng</sub>''> ) || 60 &plusmn; 1&deg;C
|-
|-
|Avg. Press. Intake 1 ( ''p<sub>in,1</sub>'' ) || 0.95 &plusmn;0.002 bar
|Avg. Press. Intake 1 ( ''p<sub>in,1</sub>'' ) || 0.95 &plusmn; 0.002 bar
|-
|-
|Avg. Press. Intake 2 ( ''p<sub>in,2</sub>'' ) || 0.95 &plusmn;0.002 bar
|Avg. Press. Intake 2 ( ''p<sub>in,2</sub>'' ) || 0.95 &plusmn; 0.002 bar
|-
|-
|Avg. Press. Exhaust ( ''p<sub>out</sub>'' ) || 1.00 &plusmn;0.016 bar
|Avg. Press. Exhaust ( ''p<sub>out</sub>'' ) || 1.00 &plusmn; 0.016 bar
|-
|-
|Intake temp. 1 ( <''T<sub>in,1</sub>''> ) || 22.9 &plusmn;0.1&deg;C
|Intake temp. 1 ( <''T<sub>in,1</sub>''> ) || 22.9 &plusmn; 0.1&deg;C
|-
|-
|Intake temp. 2 ( <''T<sub>in,2</sub>''> ) || 23.2 &plusmn;0.1&deg;C
|Intake temp. 2 ( <''T<sub>in,2</sub>''> ) || 23.2 &plusmn; 0.1&deg;C
|-
|-
|Exhaust temp. ( <''T<sub>out</sub>''> ) || 33.2 &plusmn;0.5&deg;C
|Exhaust temp. ( <''T<sub>out</sub>''> ) || 33.2 &plusmn; 0.5&deg;C
|-
|-
|Mass flow air in ( <''m<sub>in</sub>''> ) || 11.4 kg/h &plusmn;2%
|Mass flow air in ( <''m<sub>in</sub>''> ) || 11.4 kg/h &plusmn; 2%
|-
|-
|Mass flow air out ( <''m<sub>out</sub>''> ) || 11.4 kg/h &plusmn;2%
|Mass flow air out ( <''m<sub>out</sub>''> ) || 11.4 kg/h &plusmn; 2%
|-
|-
|Humidity (&Phi;)||1.8% - RH
|Humidity (&Phi;)||1.8% - RH
Line 69: Line 69:


<div id="figure9"></div>
<div id="figure9"></div>
{|align="center" width=532
{|align="center" width=460
|[[Image:AC2-10_BC.png|532px]]
|[[Image:AC2-10_fov.png|460px]]
|-
|-
|align="left"|'''Figure 9:''' Phase-averaged flow field during intake at 270&deg; bTDC showing the FOV for PIV, high resolution PIV, high speed PIV, stereo-PIV and tomographic PIV. This figure is taken from [[Best_Practice_Advice_AC2-10#4|[4]]].
|align="left"|'''Figure 9:''' Phase-averaged flow field during intake at 270&deg; bTDC showing the FOV for PIV, high resolution PIV, high speed PIV, stereo-PIV and tomographic PIV. This figure is taken from [[Best_Practice_Advice_AC2-10#4|[4]]].
Line 163: Line 163:
===Uncertainty of mean velocity estimation===
===Uncertainty of mean velocity estimation===
Statistical analysis of the in-cylinder flow velocities requires a sufficient sample size to accurately represent the ensemble-mean values. Due to limited computational resources the number of simulated cycles is limited and varying cycle numbers have been reported to sufficiently reproduce the ensemble-mean and standard deviation to a certain extent as obtained in the experiment, i.e. [[Best_Practice_Advice_AC2-10#21|[21]],&nbsp;[[Best_Practice_Advice_AC2-10#18|18]]].  
Statistical analysis of the in-cylinder flow velocities requires a sufficient sample size to accurately represent the ensemble-mean values. Due to limited computational resources the number of simulated cycles is limited and varying cycle numbers have been reported to sufficiently reproduce the ensemble-mean and standard deviation to a certain extent as obtained in the experiment, i.e. [[Best_Practice_Advice_AC2-10#21|[21]],&nbsp;[[Best_Practice_Advice_AC2-10#18|18]]].  
The convergence of flow velocity statistics based on finite sample sizes are shown in Figure \ref{fig:convergence} by the standard deviation of the sample-means  for the x- and y-velocity components at 4 specific locations. It reveals that the convergence is spatially dependent. For a sample size (n) of 50 the uncertainties of  values are up to 30\,\% of the maximum as obtained for n\,=\,2 and sample sizes as large as 2700 samples are needed to provide minimum uncertainty of the mean values. At 270\degree bTDC uncertainties are highest for the y-velocity component in the stagnant flow region where the annular flow from the intake impinges on the reversing flow above the piston. The location of this stagnant flow region varies in size, shape, and location among the different sample-means, resulting in the high relative uncertainty for this region. During compression, at 90\degree bTDC, the largest uncertainty exists near the tumble vortex center, indicating variances in size, shape, and location of the tumble center among the different sample-means.
The convergence of flow velocity statistics based on finite sample sizes are shown in [[Test_Data_AC2-10#figure10|Figure 10]] by the standard deviation of the sample-means  for the x- and y-velocity components at 4 specific locations. It reveals that the convergence is spatially dependent. For a sample size (n) of 50 the uncertainties of  values are up to 30% of the maximum as obtained for n&nbsp;=&nbsp;2 and sample sizes as large as 2700 samples are needed to provide minimum uncertainty of the mean values. At 270&deg; bTDC uncertainties are highest for the y-velocity component in the stagnant flow region where the annular flow from the intake impinges on the reversing flow above the piston. The location of this stagnant flow region varies in size, shape, and location among the different sample-means, resulting in the high relative uncertainty for this region. During compression, at 90&deg; bTDC, the largest uncertainty exists near the tumble vortex center, indicating variances in size, shape, and location of the tumble center among the different sample-means.




Line 174: Line 174:


===Repeatability of engine operation===
===Repeatability of engine operation===
The use of complementary measurement techniques allows to identify errors and to improve the reliability of the data set. The comparison of all velocity measurements as obtained from the different PIV techniques emphasizes the reliability of the velocity data (Figure \ref{fig:veloprofile}). Statistics are based on the maximum number of phased-locked flow images available for each PIV technique. Good agreement is shown between each PIV measurement for ensemble-average velocity profiles, while slight differences are shown for the standard deviation. Profiles of standard deviation show overall good agreement, but specific locations show differences among the PIV data. Standard deviation profiles for the HS-PIV and high-resolution PIV data can deviate from the other profiles. It is likely that a larger sample size for the HS-PIV data (78 cycles) is needed to better resolve the standard deviation profile. The high-resolution PIV data reveal a higher standard deviation in some locations, which could result from the spatial filtering of PIV techniques.
The use of complementary measurement techniques allows to identify errors and to improve the reliability of the data set. The comparison of all velocity measurements as obtained from the different PIV techniques emphasizes the reliability of the velocity data ([[Test_Data_AC2-10#figure11|Figure 11]]). Statistics are based on the maximum number of phased-locked flow images available for each PIV technique. Good agreement is shown between each PIV measurement for ensemble-average velocity profiles, while slight differences are shown for the standard deviation. Profiles of standard deviation show overall good agreement, but specific locations show differences among the PIV data. Standard deviation profiles for the HS-PIV and high-resolution PIV data can deviate from the other profiles. It is likely that a larger sample size for the HS-PIV data (78 cycles) is needed to better resolve the standard deviation profile. The high-resolution PIV data reveal a higher standard deviation in some locations, which could result from the spatial filtering of PIV techniques.




Line 190: Line 190:
----
----
{{ACContribs
{{ACContribs
|authors=Carl Philip Ding,Rene Honza, Elias Baum, Andreas Dreizler
|authors=Carl Philip Ding, Rene Honza, Elias Baum, Benjamin Böhm, Andreas Dreizler
|organisation=Fachgebiet Reaktive Strömungen und Messtechnik (RSM),Technische Universit&auml;t Darmstadt, Germany
|organisation=Fachgebiet Reaktive Strömungen und Messtechnik (RSM),Technische Universit&auml;t Darmstadt, Germany
}}
}}
Line 198: Line 198:
}}
}}
{{ACContribs
{{ACContribs
|authors=Chao He , Wibke Leudesdorff, Guido Kuenne, Benjamin Böhm, Amsini Sadiki, Johannes Janicka
|authors=Chao He, Wibke Leudesdorff, Guido Kuenne, Amsini Sadiki, Johannes Janicka
|organisation=Fachgebiet Energie und Kraftwerkstechnik (EKT), Technische Universität Darmstadt, Germany
|organisation=Fachgebiet Energie und Kraftwerkstechnik (EKT), Technische Universität Darmstadt, Germany
}}
}}

Latest revision as of 15:55, 2 November 2018

Front Page

Description

Test Data

CFD Simulations

Evaluation

Best Practice Advice

Internal combustion engine flows for motored operation

Application Challenge AC2-10   © copyright ERCOFTAC 2024

Test data

Operational conditions

The engine was motored at 800 RPM, with 0.95 bar and 23° C intake air temperature as the base case. The cylinder head temperature was thermally controlled at 60° C. The engine operated with dry air with a relative humidity of 1.8%. Additionally, engine speed variations were performed (1500 and 2500 RPM). Unless otherwise stated the operational conditions are summarized in table 1.

Table 1: Engine operating parameters for the base case.
RPM 800 ± 7 min-1
Cyl. Head, coolant temp. ( <Teng> ) 60 ± 1°C
Avg. Press. Intake 1 ( pin,1 ) 0.95 ± 0.002 bar
Avg. Press. Intake 2 ( pin,2 ) 0.95 ± 0.002 bar
Avg. Press. Exhaust ( pout ) 1.00 ± 0.016 bar
Intake temp. 1 ( <Tin,1> ) 22.9 ± 0.1°C
Intake temp. 2 ( <Tin,2> ) 23.2 ± 0.1°C
Exhaust temp. ( <Tout> ) 33.2 ± 0.5°C
Mass flow air in ( <min> ) 11.4 kg/h ± 2%
Mass flow air out ( <mout> ) 11.4 kg/h ± 2%
Humidity (Φ) 1.8% - RH
Intake Valve Opening (IVO) 325° TDC
Intake Valve Closing (IVC) 125° bTDC
Exhaust Valve Opening (EVO) 105° aTDC
Exhaust Valve Closing (EVC) 345° bTDC

The spark plug was removed and replaced with a threaded plug and the injector was inactive for the experiments. In-cylinder pressure as well as pressure and temperature in the intake and exhaust manifolds were recorded simultaneously for all experiments to monitor the engine and to provide boundary conditions for simulations. The ensemble-average and standard deviation in-cylinder pressure trace and pressure boundary conditions (pin,1, pin,2, and pout) are shown in Figure 8. The blow-by past the piston rings was found to be below the precision of the rotary piston gas meter systems in the intake and exhaust (1.8% of intake flow).

AC2-10 BC.png
Figure 8: Engine transient boundary conditions: a) intake and exhaust valve lift, b) in-cylinder pressure, c) intake manifold pressures (pin,1, pout), d) exhaust manifold pressure. Statistical values are based on 600 engine cycles. Negative CAD values indicate bTDC. This figure is reproduced with permission from [4].

Overview

The available data set on the in-cylinder flow field of the motored engine is summarized in this section. Figure 9 summarizes the available field-of-views (FOV) for the variety of PIV measurements, which are all centered within the central tumble plane at z = 0mm.

The first two statistical moments (mean and rms) are available for the following data:

  • Both in-plane velocity components for representative phases of the engine cycle at selected crank angles within a large FOV (65x65 mm2) from the statistically independent PIV measurements (PIV) resolving the global in-cylinder flow (large database: up to 2700 cycles).
  • High resolution data of both in-plane velocity components within a small FOV (20x15 mm2) from the statistically independent PIV measurements (PIV (High Res.)) resolving the smaller scales.
  • All three velocity components in the measurement plane within a FOV (47x35 mm2) from stereo-PIV (Stereo-PIV).
  • All three velocity components within a 47×35×8 mm3 volume centered in the central tumble plane from tomographic-PIV (Tomo-PIV).
  • Time resolved data of both in-plane velocities within a 54×54 mm2 FOV over entire consecutive cycles (PIV (High-Speed)).

The imaged CADs together with the number of respective images are given in Table 2 for all operational conditions.


AC2-10 fov.png
Figure 9: Phase-averaged flow field during intake at 270° bTDC showing the FOV for PIV, high resolution PIV, high speed PIV, stereo-PIV and tomographic PIV. This figure is taken from [4].


Table 2: Overview of the available experimental flow data.
PIV Technique Engine Speed (RPM) Intake flow CAD imaged (bTDC, negative numbers refer to aTDC) Number of images
Low repetition rate 2D2C PIV 800 Tumble 270, 90, -90, -270 2700
800 Tumble 315, 260, 180, 45, -45, -160, -180, -315 600
1500 Tumble 270, 90 600
2500 Tumble 270, 90 600
2D2C High-resolution PIV 800 Tumble 270, 90, -90, -160, -180, -270 600
2D2C High-speed PIV 800 Tumble 360 to - 360 78
Stereoscopic PIV 800 Tumble 270, 225, 180, 135, 90, -90, -180, -270 600
800 Swirl 270, 225, 180, 135, 90, -90, -180, -270 600
Tomographic PIV (4mm light sheet thickness) 800 Tumble 270, 225, 180, 135, 90, -90, -180, -270 600
800 Swirl 270, 225, 180, 135, 90, -90, -180, -270 600
1500 Tumble 270, 225, 180, 135, 90 400
1500 Swirl 270, 180, 90 400
Tomographic PIV (8mm light sheet thickness) 800 Tumble 270, 225, 180, 135, 90 600
1500 Tumble 270, 90 400


Description of the experiment

Particle image velocimetry (PIV) was used to capture the flow field within the central tumble plane. The in-cylinder flow was measured by conventional planar particle image velocimetry (PIV), stereo-PIV, high-speed PIV, and tomographic PIV. Silicone oil droplets (∼ 1μm diameter) were used for seeding the intake flow. Further details on the setup are provided in [12].

Conventional PIV and stereo-PIV

Conventional 10Hz PIV was applied for the statistically independent sampled data to achieve highest possible quality and resolution. A dual cavity Nd:YAG laser (Gemini New wave) was used for illumination and a double frame CCD camera (PCO SensiCam, 1376×1040 pixels, 12 bit) to detect the scattered light of the seeding particles. Measurements were performed with a resolution of 2.2mm and 0.6mm based on a 32×32 pixel interrogation window resolving the global in-cylinder flow (FOV = 65×65 mm2) and the small scales (FOV = 20×15 mm2). The conventional setup provides the two in-plane velocity components (2D2C). With the addition of another CCD camera in a stereo setup all three velocity components were captured in the plane (2D3C).

Tomographic PIV

The ability to capture the 3D structure of the in-cylinder flow field was accomplished from Tomographic PIV. A probe volume of 4mm and 8mm thickness centered around the central tumble plane was illuminated with a dual-cavity Nd:YAG laser (PIV-400, Spectra Physics) with an average pulse energy of 375mJ. Four CCD cameras (ImagerIntense, LaVision, 1376×1040 pixels, 12 bit) with identical lenses (50mm, Nikon) in Scheimpflug arrangement were setup circularly around the optical access, such that each camera projection provides independent line-of-sight information of the illuminated volume. The alignment of the cameras achieved a visualized area of 47×35 mm2 located centrally between the intake valves with a vector spacing of 0.4mm.

High-speed PIV

High-speed 2D2C PIV measurements were recorded at 4.8kHz sampling rate (i.e. CAD resolution at 800 RPM) to determine the temporal evolution of the in-cylinder flow over the entire engine cycle for consecutive cycles. A dual-cavity diode pumped Nd:YVO4 slab laser (Edgewave) was used together with a high-speed camera (LaVision, HSS6). Due to the wide dynamic range of the velocities within a cycle (from 1 to 50m/s) time separation between both laser pulses needed to be adjusted depending on the imaged CAD. To capture the entire cycle, pulse separation was adjusted variable during the measurements, optimized on the observed CAD.

Uncertainties

The uncertainty in the experimental data depends on the uncertainty of the PIV setup. For the statistical quantities the number of samples needs to be considered in the uncertainty estimation. Additionally the reproducibility of the engine operation has an impact.

Uncertainty in PIV measurements

PIV data quality depends on parameters such as tracer particles, camera settings, and choice of PIV algorithm. Much work has been done to identify and optimize experimental parameters that minimize measurement uncertainties associated with digital PIV [34]. Optimized seeding densities, time delay between the double-images, and focused particles provided correlation peak values in the order of 0.4 to 0.9. Uncertainty assessments in the literature regarding the aforementioned particle image quality yields a 2–3% measurement uncertainty. The PIV processing adds additional errors, which were estimated to be within 4%. The average accumulated uncertainty in the PIV measurements was within 5% [4], while the presented TPIV measurements have uncertainties up to 9% [5].

Uncertainty of mean velocity estimation

Statistical analysis of the in-cylinder flow velocities requires a sufficient sample size to accurately represent the ensemble-mean values. Due to limited computational resources the number of simulated cycles is limited and varying cycle numbers have been reported to sufficiently reproduce the ensemble-mean and standard deviation to a certain extent as obtained in the experiment, i.e. [2118]. The convergence of flow velocity statistics based on finite sample sizes are shown in Figure 10 by the standard deviation of the sample-means for the x- and y-velocity components at 4 specific locations. It reveals that the convergence is spatially dependent. For a sample size (n) of 50 the uncertainties of values are up to 30% of the maximum as obtained for n = 2 and sample sizes as large as 2700 samples are needed to provide minimum uncertainty of the mean values. At 270° bTDC uncertainties are highest for the y-velocity component in the stagnant flow region where the annular flow from the intake impinges on the reversing flow above the piston. The location of this stagnant flow region varies in size, shape, and location among the different sample-means, resulting in the high relative uncertainty for this region. During compression, at 90° bTDC, the largest uncertainty exists near the tumble vortex center, indicating variances in size, shape, and location of the tumble center among the different sample-means.


AC2-10 convergence.png
Figure 10: Convergence of the standard deviation of the sample-means at specific locations within the velocity field for 270° and 90° bTDC. This figure is taken from [4].

Repeatability of engine operation

The use of complementary measurement techniques allows to identify errors and to improve the reliability of the data set. The comparison of all velocity measurements as obtained from the different PIV techniques emphasizes the reliability of the velocity data (Figure 11). Statistics are based on the maximum number of phased-locked flow images available for each PIV technique. Good agreement is shown between each PIV measurement for ensemble-average velocity profiles, while slight differences are shown for the standard deviation. Profiles of standard deviation show overall good agreement, but specific locations show differences among the PIV data. Standard deviation profiles for the HS-PIV and high-resolution PIV data can deviate from the other profiles. It is likely that a larger sample size for the HS-PIV data (78 cycles) is needed to better resolve the standard deviation profile. The high-resolution PIV data reveal a higher standard deviation in some locations, which could result from the spatial filtering of PIV techniques.


AC2-10 veloprofile.png
Figure 11:Velocity profiles of the phase-average and standard deviation extracted along y=-30mm for 270° and 90° bTDC from all PIV techniques. Statistics are based on the maximum number of samples available for each PIV technique. This figure is taken from [4].

Data files

Data is available on request. Please contact Dr. Benjamin Böhm (boehm@rsm.tu-darmstadt.de).




Contributed by: Carl Philip Ding, Rene Honza, Elias Baum, Benjamin Böhm, Andreas Dreizler — Fachgebiet Reaktive Strömungen und Messtechnik (RSM),Technische Universität Darmstadt, Germany


Contributed by: Brian Peterson — School of Engineering, University of Edinburgh, Scotland UK


Contributed by: Chao He, Wibke Leudesdorff, Guido Kuenne, Amsini Sadiki, Johannes Janicka — Fachgebiet Energie und Kraftwerkstechnik (EKT), Technische Universität Darmstadt, Germany


Contributed by: Peter Janas, Andreas Kempf — Institut für Verbrennung und Gasdynamik (IVG), Lehrstuhl für Fluiddynamik, Universität Duisburg-Essen, Germany


Contributed by: Stefan Buhl, Christian Hasse — Fachgebiet Simulation reaktiver Thermo-Fluid Systeme (STFS), Technische Universität Darmstadt, Germany; former: Professur Numerische Thermofluiddynamik (NTFD), Technische Universität Bergakademie Freiberg, Germany

Front Page

Description

Test Data

CFD Simulations

Evaluation

Best Practice Advice


© copyright ERCOFTAC 2018