Presentation on theme: "Topic 4: Panel Discussions 1 April 5, 2014 EFC Panelist-inspired Open Discussion 1.Assessment of 4.1Dave Reuss 2.Assessment of 4.2 Dan Haworth 3.Assessment."— Presentation transcript:
Topic 4: Panel Discussions 1 April 5, 2014 EFC Panelist-inspired Open Discussion 1.Assessment of 4.1Dave Reuss 2.Assessment of 4.2 Dan Haworth 3.Assessment of 4.3 & 4.4 Tommaso Lucchini 4.Path for Spray G into an engine Scott Parrish 5.Diesel Needs Oivind Andersson EFC Priorities and Needs for ECN 3.x Action Item: Choose most important and universal problems inhibiting progress that would inspirer non-competitive unfunded effort?
Topic 4: Panel Discussions 2 April 5, 2014 ECN 3.X Topic 4.1 Measurements for Simulation benchmarking Experimental efforts require more than simulation benchmarking. Experiments provide observations beyond 5 senses. Contributions of simulation to experimental results Steady-state operation is not adequate for assessing LES effectiveness. Compute n cycles perturb cycle n+1 initial/boundary conditions compute next cycle(s) Compare with equivalent perturbation in the experiment Interpret Interpolate & extrapolate design Archive
Topic 4: Panel Discussions 3 April 5, 2014 ECN 3.X Topic 4.2 Simulation-to-simulation benchmarking (LES, RANS) - Haworth How do we know when (if) our physical models/numerical methods/mesh resolution/boundary condition specification/etc. are sufficient for the task at hand? – Engine simulations are expensive - especially multiple-cycle LES Should we require/suggest submission of results using reference models? – e.g., turbulence, ignition, flame propagation... – To what extent do results from different codes vary, even with nominally the same physical models? Should we require/suggest submission of results for one or more reference configurations? – Back to decaying homogeneous isotropic turbulence in a period box? Should we require/suggest submission of quantitative resolution metrics for LES results? – Several metrics have been proposed, e.g., xx% of TKE resolved Should we require/suggest submission of convergence metrics for LES? – e.g., resolved + residual (subfilter-scale) contributions to specified quantities independent of mesh size in appropriate limits
Topic 4: Panel Discussions 4 April 5, 2014 ECN 3.X Topics 4.3 Benchmarking: comparisons, analysis, and validation 4.4 CCV origins and abatement - Lucchini Simulations vs exp. data: How to validate a CFD code for engine simulations? 1)Identification of a common engine geometry which will be used for cold-flow validation and a single operating point. TCC is a good candidate. 2)How many cycles to simulate (RANS)? How many cycles to measure to properly get average quantities for RANS comparison? 3)Which are the proper boundary conditions to be set at boundaries: Total variable pressure? Mass flow? How do boundary conditions affect the predicted volumetric efficiency? How to evaluate this? 4)Effects of mesh structure (even more than grid size) to be absolutely evaluated. Is a coarse, flow-oriented grid much better than a refined non-oriented one?
Topic 4: Panel Discussions 5 April 5, 2014 Is the Spray G injector suitable for engine experiments? EFC – Pathway for Spray G into and engine YES! early injection, homogenous late injection, stratified late injection, HCCI, dilute combustion, etc. Central mount position required (no bend angle) What could be learned by putting Spray G into an engine? What problems should be pursued? MUCH TO BE LEARNED! Cylinder flow / spray interaction - stochastics Ignition and combustion Spray / wall interactions... Need to be careful to maintain focus
Topic 4: Panel Discussions 6 April 5, 2014 ECN 3.X Diesel Engine Needs for Spray B - Oivind Andersson Which important spray phenomena are not accounted for in spray vessels? Spray-swirl interaction Spray-spray interaction – Entrainment of hot products – Effects on flow between jets Spray-wall interaction – Processes downstream of impingement point Late cycle mixing (?) – How do sprays affect processes occurring after EOI? Multiple injections – Why does a post injection reduce soot, etc? – Combustion system effects!
Topic 4: Panel Discussions 8 April 5, 2014 4.1.1.Measurement and instrumentation standards 184.108.40.206.Required engine and system geometry 220.127.116.11.In-cylinder & extra-cylinder instrumentation needs. 18.104.22.168.1.EnsAve, crank-resolved volume-ave 22.214.171.124.2.Gas temperature measurements 126.96.36.199.3.Wall temperatures 188.8.131.52.Sensors: location, type, BC 4.1.2.Engine operation, test-to-test repeatability criteria. 184.108.40.206Metrics and criteria 220.127.116.11Run-time control 18.104.22.168Post-run analysis & charting 4.1.3.Motored flow characterization. 4.1.4.Motored spray characterization 4.1.5.Combustion characterization 22.214.171.124.Global analysis 126.96.36.199.Product gas Identification, 2-D/ 3-D 188.8.131.52.Resolved homogeneous-Flame Front 4.1.6.Error Analysis 184.108.40.206.Compression ratio, TDC, valve events, & pegging 220.127.116.11.Pressure measurements 18.104.22.168.Temperature measurements 22.214.171.124.PIV: u & dx dynamic range, u &du/dx accuracy 126.96.36.199.Spray measurements 4.1. Measurements for Simulation benchmarking
Topic 4: Panel Discussions 9 April 5, 2014 4.2.Simulation-to-simulation benchmarking (LES, RANS) 4.2.1.Engine simulation methodology 188.8.131.52 Complete test-bench coupled with 1D simulation 184.108.40.206 Type of boundary conditions 220.127.116.11 Consecutive vs. Perturbed engine cycles 4.2.2.Meshing strategy and resolution 18.104.22.168 Immersed boundary, auto mesh movement 22.214.171.124 Spatial resolution 126.96.36.199 Resolved-flow measures for sub-grid efficacy 4.2.3.Modeling issues 188.8.131.52. Numerical method (spatial and temporal) 184.108.40.206. Which models capture relevant physics & chemistry? 220.127.116.11. Robust CPU doable computations
Topic 4: Panel Discussions 10 April 5, 2014 4.3.Benchmarking: comparisons, analysis, and validation 4.3.1.Global engine metrics 18.104.22.168. In-cylinder Global & 0-D Metrics 22.214.171.124. Intake & Exhaust Systems 1-D quantities 4.3.2. In-cylinder flow characterization 126.96.36.199. Statistical Methods 188.8.131.52.1. phase-average and standard deviation 184.108.40.206.2. CCV vs. turbulence vs. noise 220.127.116.11.3. Required Number of sampled cycles. 18.104.22.168.4. Conditional sampling on flow events 22.214.171.124. POD: phase dependent, phase invariant. 126.96.36.199. Intra-cycle vs. inter-cycle comparison 188.8.131.52. Qualitative comparison of velocity fields, 2-D vs. 3-D 184.108.40.206. Sub-grid model efficacy from resolved-scale measures. 220.127.116.11. Simulation efficacy of scalar mixing. 4.3.3. Combustion modeling validation 18.104.22.168. Global Heat Release 22.214.171.124. Ignition and early flame development 126.96.36.199. Fully-developed turbulent flame
Topic 4: Panel Discussions 11 April 5, 2014 4.4.CCV Origins and abatement 4.4.1.Exploring CCV origin 188.8.131.52.Flow motion influence 184.108.40.206.1.Geometry of intake manifold and ducts 220.127.116.11.2.Flow around valve or exiting valve curtain 18.104.22.168.3.In-cylinder flow motion 22.214.171.124.4.Flow motion near spark plug at spark timing 126.96.36.199.5.Influence of the previous engine cycle 188.8.131.52.6.Influence of the engine scavenging process 184.108.40.206. Cylinder charging variation 220.127.116.11. Local mixture heterogeneity 18.104.22.168. Role of the spark-ignition period 22.214.171.124. Combustion process in CCVs 126.96.36.199.1.Early flame kernel propagation? 188.8.131.52.2.Occurrence of partial combustion & quenching 184.108.40.206.3.Heat transfer influence 4.4.2. Factors that increase CCVs 220.127.116.11. Engine geometry 18.104.22.168. Engine operating points 22.214.171.124. Mixture preparation 126.96.36.199. Flow motion pattern