Piston Bowl Swap - Emission Characterization study on a CAT3410 Diesel Engine using Converge CFD

I. INTRODUCTION

Compression ignition engine (Diesel engine) as the name suggests is an engine where fuel-injection takes place into the engine cylinder toward the end of compression stroke, just before the desired start of combustion. A closed cycle sector analysis of the cylinder is performed in this project as the main aim is to study the engine performance parameters & emission characterization under two piston bowl profiles - Omega & Open W type.

Closed cycle analysis is where there are no intake - exhaust ports, no moving geometry for simplification and the simulation is performed in the closed cylinder region. This is purely done for computational efficiency. The Case modelling is done such that the two piston bowls can be compared directly by keeping parameters like mass of fuel injected, compression ratio, spray and combustion modelling settings, mesh refinement and more constant in both cases. Engine parameters such as incylinder pressure, HRR, temperature field are studied alongside emissions such as hiroy soot, Nox & UHC. Combustion modelling is performed using SAGE - a detailed chemical kinetics solver in CONVERGE CFD software. The following report also explains theoretical modeling concepts applied in converge software for this case setup.

II. OBJECTIVE

 Setting up Spray modeling - Injector & Nozzle

2. Hydrodynamic Case setup - Combustion modeling, Grid refinement AMR parameters. 

3. Evaluate engine performance parameters for both piston bowl sector geometries of a CAT3410 engine. 

4. Emission characterization of the same - Soot, Nox & UHC & comparison.

III. MODELING THEORIES CONVERGE-CFD

III.1 SPRAY MODELING - Spray Breakup

Converge uses state of art models for spray processes involving liquid atomization, drop breakup, collision & coalescence, turbulent dispersion & drop evaporation. In terms of liquid sprays, Converge uses Lagrangian solver to model discrete particles and Eulerian solver to model the continuous fluid domain. Lagrangian specification of the flow field is a way of looking at fluid motion where the observer follows an individual fluid parcel as it moves through space and time. Plotting the position of an individual parcel through time gives the pathline of the parcel. The Eulerian specification of the flow field is a way of looking at fluid motion that focuses on specific locations in the space through which the fluid flows as time passes. The eulerian cell center is fixed while the lagrangian parcels moves. Heat, momentum & mass transfer occur between the discrete and continous phases via source terms in transport equations.

                                                             

Parcels are introduced into the domain at the injector, which presents a group of identical drops with same radius. velocity and temperature. Parcel is the basic unit converge solves instead of drop. Parcel undergoes several physical processes like, Pirmary breakup, secondary breakup, drop drag, collision & coalescence, turbulent dispersion & evaporation. 

                                                                           

III.2 Evaporation Model

The spray evaporation that takes place need to be considered and for that converge uses a certain standard pre-created models such as Frossling, Chiang or with boiling. In this project Frossling model is used and its not needed to provide particular equations of the model as Converge takes care of it. Converge asks for evaporation source to simulate multi-component vaporization. The source used in this project is "Source all base parcel species" which denotes multi-component liquid species evaporate into base species.   

As the evaporation source is specified, the maximum radius of ODE droplet heating needs to be specified as well, its basically a temperature discretization parameter which will control if drop temperature is uniform or radially varying. If drop radius exceeds max value then converge assumes a radially varying temperature or else Uniform temperature distrubution. 

III.2.1 Spray Penetration & Distribution

Converge writes the liquid and vapor penetration lengths at each output interval, it calculates vapor penetration length for each nozzle. For each cell that meets the critera that fuel vapor mass fraction exceeds user specified value or cell size does not exceed user specified bin size, converge calculates the distance from the center of nozzle to the center of the cell. 

The parcel distribution is done in two ways where the parcels are either distributed evenly throughout the cone or clustered near the center of the cone. Depending on requirement either distribution can be selected, for this project center of cone is selected.

                                                                                            

                                                                   

III.2.2 Collision/Coalescence & Drop Drag

If in a spray simulation parcels collide with parcels only in the same cell, it can lead to grid sensitivity, which is why converge has an adaptive collision mesh to reduce grid sensitivity. Once collision mesh is checked, the parcels can collide across grid cells, requiring no additional meshing or any other setup. In this project it is done using a standard NTC collision model & dynamic drop drag model with default TAB model constants to determine drop distortion. Dynamic drop drag calculates the drag coefficient to account for variation in drop shape as shown in spray breakup. In the pic below it is observable that parcels with collision mesh look more reasonable & natural. Other drop options are spherical drop drag & No drop drag. 

                                   

III.2.3 Injector & Nozzle

Once the spray breakup, & Parcel modeling is done, Injector and nozzles need to be specified, their locations, model, radius and such. One injector can have multiple nozzles and for this project there is 1 nozzle in one injector, This is done as per the engine manufacturer's requirement and sector simulation based on the amount of fuels and speed at which the fuel has to be sprayed inside the chamber. Thei injector has a profile rate_shape.in where its rate vs crank is determined is is shown below.

                                                                                

The fuel is specified for this injector as C7H16 with mass fraction 1 as only 1 fuel is being sprayed in this cylinder. There are two spray type, solid cone spray and hollow cone spray. Kelvin-Helmholtz & reyleigh taylor models are used for solid cone sprays. Nozzles are added next into the injector where nozzle parameters are determined such as diameter, injection radius and spray cone angle as well as co-ordinates of nozzle. 

III.3 COMBUSTION MODELING in CONVERGE

Combustion facilitates the energy transfer in the engine. Converge contains detailed chemistry solver and simplified combustion models. 

                                                         

III.3.1 SAGE detailed chemical kinetics solver (used in this simulation.) 

SAGE solver is the most predictive and accurate way to model combustion, it can accurately model ignition and laminar flame propogation. It uses local conditions to calculate reaction rates based on principles of chemical kinetics. Also determines kinetically limited phenomena such as engine knock and emissions. It requires a CHEMKIN-formatted input files to solve the reaction rates & ODE'S. The ODE solver used is called as CVODE solver. SAGE couples with transport solver via source terms in the species transport equations. 

A chemical reaction mechanism is a set of elementary reactions that describe an overall chemical reaction. The combustion of different fuels can be modeled by changing the mechanism (e.g., there are mechanisms for isooctane, gasoline, n-heptane, natural gas, etc.). SAGE calculates the reaction rates for each elementary reaction while the CFD solver solves the transport equations. Given an accurate mechanism, SAGE (in addition to AMR) can be used for modeling many combustion regimes (ignition, premixed, mixing- controlled). You can use SAGE to model either constant-volume or constant-pressure combustion.

In each time-step, the chemistry solver calculates the new species mass fractions immediately prior to solving transport equations. The change in species mass fractions is treated as source. 

                             

Omega is the relaxation source term, which takes the values. Detailed mathematical formulation can be found Converge 3.0 manual.

III.3.2 Accelerating SAGE 

As this is the most detailed solver, solving it in every cell can be expensive, time consuming. Converge has accelerating options which require increasing minimum cell temperature and min HC species mole fraction to reduce the number of cells in which combustion calculations are performed. Using start time and end time to limit sage to specific time interval & can limit SAGE operation to specific region. Other options include, Setting resolve only if temperature changes by specified value ~ no greater than 2 K. & using analytical jacobian to pass an analytically calculated jacobian to solver. SAGE cal also reduce time by using multizone modeling where it groups cells into bins (zones) for which appropriate options are provided. 

PISTON BOWL  - Design considerations 

Diesel combustion is known to be very lean with A/F ratios of 25:1 at peak torque, 30:1 at rated speed/maximum power conditions, and over 150:1 at idle for turbocharged engines. Yet, this extra air does not enter into the combustion process. It is rather heated during combustion and exhausted—causing diesel exhaust to be lean. Even though the average air-fuel ratio is lean, if proper care is not taken in the design process, regions of the combustion chamber can be fuel rich and lead to excessive smoke emissions. A key objective in designing the combustion bowl then is to ensure that mixing of fuel and air is adequate to mitigate the impact of fuel rich regions and allow the engine to meet its performance and emissions targets. Turbulence in the air motion within the combustion bowl is found to be beneficial to the mixing process and can be used to achieve this goal. Swirl induced by the intake port can be enhanced or squish can be generated by the piston as it approaches the cylinder head to create more turbulence during the compression stroke through proper design of the bowl in the piston crown.

Two piston bowl profiles are studied in this project, OMEGA & OPEN W. 

IV. Sector extraction & applying bowl profiles in CONVERGE CFD

Sector geometry is an extraction of the full 360 deg cylinder into desired angle. The sector's can be pulled based on spray to angle ratio, for example if the cylinder has 8 nozzles then 360/8 gives 45 deg. Or if 6 nozzles then 360/6 = 60 deg. This sector will have a single nozzle and will be 1/6th part of the cylinder making this simulation extremely computationally efficient coupled with close cycle analysis. 

There is an option in converge called - Make engine sector surface from where we can pull create sector geometries & provide bowl profile. The cylinder shown below is the cylinderical surface geometry which is used in the sector tool to extract & create geometry. 

                                

The Box below represents the "Make engine sector surface" tool box in converge cfd, After providing the following cylinder parameters such as bore, stroke, con rod length, we provide the above mentioned surface.dat file & then import the bowl profile. The red diagram below represents the bowl profile shape as per the data. The below box represents OPEN W type then Omega piston bowl profile. After pressing extract the geometry is created. Both sectors are 60 degree angled.

                                                                                                     Open W bowl profile - sector extraction

                                

                                                                                                     Omega bowl profile - sector extraction

                             

Extracted surface within the surface.dat geometry:

This is how the extraction of sector surface looks. This image is only for demonstration of the geometry. The longer portion is the desired geometry with its bottom most brown color boundary being the bowl profile. Open W piston is shown here. 

                                                                        

V. SECTOR GEOMETRY 

The geometry created & labeled as per the boundary flagging. After making sector surface, converge automatically flags them as its an inbuilt feature. There is only one region created in this geometry called cylinder which comprises of all the boundaries shown below. 

Top view - Cylinder Head.

                               

VI. FULL HYDRODYNAMIC CASE-SETUP

The Case-setup is tabulated with the flow of settings done in CONVERGE STUDIO. Detailed boundary conditions, regions & initializations, spray & combustion modeling conditions are laid out in the tables. Standardized models & algorithms used have been displayed as well. 

                             

                             

VI.1 SPRAY MODELING 

                             

                                                                                                        Injector & Nozzle Setup                            

                             

Spray rate graph 

This graph is obtained from tools - Spray rate preview, where we can see the graph of rate shape, peneteration velocity, injection pressure etc wrt crank angle provided in rate shape.

                                                

 

VI.2 COMBUSTION MODELING SETUP

                               

VI.3 FUEL & PARCEL SIMULATION

The fuel used in this simulation is DIESEL2 / C7H16, the composition of both is same DIESEL2 in converge is more refined for simulation. Parcel simulation specifies fuel being used, it can be setup from predefined liquids in parcel simulation option in converge. Once the chosen fuel is selected converge incorporates all its properties as seen below, to utilize it in chemical reactions and combustion simulation.

                            

VII. GRID CONTROL SETUP

VII.1 AMR - SGS CONCEPT 

Adaptive Mesh Refinement is a method of adapting & enhancing the accuracy of a solution within certain sensitive or turbulent regions of simulation, dynamically and during the time the solution is being calculated. When solutions are calculated numerically, they are often limited to pre-determined quantified grids as in the Cartesian plane which constitute the computational grid, or 'mesh'

In this Project, Subgrid scale AMR is used in the Temperature & Velocity variables of the Simulation . Converge calculates the second order of the variable subject to AMR embedding and while calculating if the difference is above the Sub-grid criterion, it will perform a refinement of the cells with embed level as provided.  The base grid is 0.004 m 

The formula used is : 

                                                      

                                                       

                                                                                                             Velocity AMR - SGS

                                                     

                                                                                                           Temperature AMR - SGS

VII.2 Fixed Embedding 

Fixed embedding follows same refinement formula except its not adaptive to solution progressing and is there for the mentioned period. 

1. Nozzle embedding 

It is important to have refinement at the nozzle just slightly before spray parcels peneterate into the cylinder, this automated generated mesh refines the variables through the parcels - fluid that enter and as theyre mixing with the air and atomizing. This improves the simulation accuracy while also maintaining a high quality grid control measure which is computationally efficient. 

                                                  

2. Piston & Cylinder head embedding

Piston & cylinder head boundary will have embed scale of 1 refinement according to the formula. The piston refinement is done to capture the fuel hitting on it & how it spreads accross its boundary and to capture the automizaton done, fluid solid interaction, capturing of swirl generated at the boundary and other wall conditions. Similarly cylinder head value is important to capture as well hence another fixed embedding refinement. 

                                            

In this image the spray AMR can be seen before the spray released and after the spray injection. The top and bottom embedding is cylinder head and piston embedding on clip plane in PARAVIEW.

                                                                    

Clip mesh Animation of geometry (Fixed embedding + AMR) - PARAVIEW.

                                            

Velocity AMR during fuel injection -

                                            

 

VII.3 TOTAL CELLS

This is the total meshing graph where we can observe how the mesh first decreased then increased. This is acting as per fixed embedding & AMR.

Omega piston geometry produces slightly more highest cell count than open W piston geometry but toward the end there is more no. of cells in open W as compared to omega.

                                  

VII.4 TIMING MAP 

Timing map is like a pathway map which lays out all simulation timing details into something like a process chart which displays what procedure is taking place at what degree such as injector function, it displays the type of Grid control parameters are set on it. It gives a handy tool to verify everything in one place & see if some refinement or function is placed as desired or not.

                                

VIII SOLUTION & POST PROCESSING

After complete modeling of the cases, the simulation is ran using CYGWIN command line terminal on a 4 core machine. Once its completed the data is converted to post process in PARAVIEW. The data plotted shows comparisons between both piston geometries.

1. In-cylinder pressure

A very high compression pressure is required in case of C.I engines which is when the fuel enters and combustion takes place. Open W piston produces 11.5373 Mpa peak pressure as compared to 11.2583 Mpa produced in omega piston geometry. They're extremely close with a difference of only 2.41 %. There is also slight delay noted in omega pistons pressure rise & drop.

 

                                

2. PV Plot

The C.I engine follows a diesel cycle. As opposed to theoretical cycles, nothing is constant here, pressure rise exists in the initial combustion phase. Since only compression & combustion stroke are simulated intake and exhaust plots are not available. However for sector simulation this is sufficient to evaluate alot of design considerations. This pv diagrams were invented to improve the efficiency of engines originally. Theyre utilized to calculate the work done as well which is usually given by 

`W = intP*dV` for a complete cycle. 

 

                                       

It can be also calculated by calculating the area under the curve. Fortunately, Converge provides an engine performance calculator built in which can evaluate these terms. 

2.1 Engine performance calculator for both Geometries

                                                                                                           Omega Piston

                                 

                                                                                                           Open W Piston

                                              

   

Omega Piston Work done (Gross) = 3409.28 Nm Power = Work/Time Where, Time is the combustion duration - `270.171°` RPM - 1600  RPS – 1600/60 = 26.66 RPS  DPS – 26.66*360 = 9597.6 Combustion duration in seconds = `270.171/9597.6 = 0.02815 s` Power = Work/Time = 3409.28/0.02815 = 121.11 KW. Torque (T) = `(60P)/(2piN) = (60*121111)/(2*pi*1600) = 722.8 Nm`

Open W piston Work done (Gross) = 3028.53 Nm Power = Work/Time Where, Time is the combustion duration - `270.156°` RPM - 1600  RPS – 1600/60 = 26.66 RPS  DPS – 26.66*360 = 9597.6 Combustion duration in seconds = `270.156/9597.6 = 0.02815 s` Power = Work/Time = 3028.53/0.02815 = 107.585 KW. Torque (T) = `(60P)/(2piN) = (60*107585)/(2*pi*1600) = 642.1 Nm`

 

                                                                           

3. Heat release rate & Combustion Efficiency 

Heat release & Integrated heat release is plotted against crank angle. It gives the Amount of heat i.e total amount of energy released from combustion process.  From the total heat release in the combustion chamber, the combustion efficiency of the engine can be determined.

HR RATE

The heat energy released rate is very high in the omega piston as compared to open W. It gives an idea that the temperature generation in the omega piston is higher. The lower heat release rate in open W indicates some unburnt fuel or delay in burning due to geometrical parameters - Design of the piston shape as that is the only parameter varying in both case. The drops observed in the openW is noise in the data and should be ignored.

                                 

INTEGRATED HR 

OpenW burns a bit slow and a smooth burning curve lower than Omega piston is observed. The omega piston engine burns faster and at a higher release which is advantageous. At the end of the plot the openW engine ends at a lower release aswell. From the engine performance parameters above & Integrated HR data we can determine that efficiency of the openW piston is lower as compared to Omega. Further accounting the emissions produced in each of these cases can validate the overall better engine. 

                                 

4. Mean Temperature

As observed from the Heat release and engine performance characteristics, the omega piston produces higher temperature and provides more burning & in case of openW the temperature isnt as high during combustion which is probably due to some incomplete burning at the combustion phase however it does completely burns later due to still higher temperature. This temperature data from incylinder region can be utilized in liner design, engine thermal analysis and more. 

                                 

Temperature animation

                                                 

6. EMISSIONS

 The major cause of emissions are non-stiochiometric combustion, dissociation of nitrogen, and impurities in the fuel and air. Even though we assumed a stiochiometric combustion for this project, it produced emissions because perfect combustion does not occur. The emissions of concern are unburnt hydrocarbons (HC), oxides of carbon (COx), oxides of nitrogen (NOx), SOx & Soot. The plots of emissions produced in this simulation is shown below. The emissions are produced in Cylinder which is then pushed out from exhaust port. we're looking at the emissions produced in the Cylinder region. 

6.1 HIROY SOOT

Soot is a mass of impure carbon particles resulting from incomplete combustion of hydrocarbons. The observation is as soon as the combustion process starts Soot started being accumulated in small quantity then peaks when the pressure is highest then dissolves and averages onto a particular value. These are generated in the fuel rich zones within cylinder during combustion & come out as exhaust smoke. Soot formation takes place in diesel combustion environment at temperatures between about 1000 and 2800 K, at pressures 50 to 100 atm, and with sufficient air overall to burn fully all the fuel. The time available for the formation of solid soot particles from a fraction of fuel is in order of milliseconds. In graph below, we have soot formation graph as well as soot oxidation graph. Large fraction of soot oxidizes within the cylinder before the exhaust process even begins. 

The soot formed in OpenW piston is significantly higher than in Omega which is as predicted as there is unburnt fuel at the beginning of combustion in openW and hence is fuel rich zone at a high temperature. Approximately 25 - 29 % difference only because of difference in piston bowl design. At the end of the combustion as well openW leaves with a significantly higher Soot which will pass through exhaust.

                                        

                                                    

                                                     

6.2 NOx 

With the higher temperature in Omega piston and higher heat release, the NOx is significantly higher here about nearly 82 % more than open W. Diesel fuels contain more nitrogen than gasolines and hence they produce higher nox as compared. This is Zeldovich Nox. NOx production is low in OpenW however soot is higher, different tradeoffs compared with engine performance will determine a better design. The critical time period is when burned gas temperatures are at a maximum i.e between start of combustion and shortly after the occurence of peak cylinder pressure. Mixture which burns early in the combustion process is especially important since it is compressed to a higher temperature, increasing the NOx formation rate, as combustion proceeds and cylinder pressure increases. And during expansion due to mixing of high and cooler burned gas freezes the NOx chemistery. This effect only occurs in Diesel fuel. 

                                    

                                                       

                                                       

6.3 Unburnt HydroCarbons

Hydrocarbons are consequence of incomplete combustion of the hydrocarbon fuel. Hydrocarbon emission indicates combustion inefficiency. As observed from all the data so far, it shows why OpenW has more unburnt HC production, however it also disolves later with a high degree of delay and averages to still slightly higher HC than present in Omega piston geometry. 

                                       

6.4 CO

CO emissions from IC engines are controlled primarily by fuel/air equivalence ratio. Higher peak is obtained by omega piston however it declines lower right after combustion, but there is a significant delay in OpenW piston engine as the fuel remains unburnt for a while its delayed. there is 2-3 % difference at peak CO but very longer degree of disolution in openW and slightly higher quantity left overall. These emissiosn data can be utilized to make improvements and changes in the designs and combustion operations of the IC engines. 

                                      

7. Gaseous Fuel plot 

The Gaseous fuel injected and diffusion basically its combustion burning is dispalyed in the plot below in Kgs that is in mass form, There is a steep increase in Omega piston design's gaseous fuel intake that is due to its design as shown in the contour below, there is significant peneteration length in this design which allows it to vapourize in the high temperature whereas alot of fuel impingement occurs in openW piston due to which high vapour isnt formed, yet as observed the fuel combusts in the end due to high temperature. Another advantage of omega type design is the swirl motion it creates as observed from previous animation of mesh refinement - velocity in Grid control, The circular swirl helps again in vapourization of fuel. This causes high burning in Omega piston design and less in openW, optimum ratio needs to be obtained while accepting appropriate tradeoff between emission & performance. No engine is perfect and there is always room for improvement. Modification of injector angle in OpenW piston is recommended to observe more improvement and can be done as part of future scope. 

                                       

                                        

IX CONCLUSION

1. Omega piston engine has an edge over OpenW piston engine in this case, there are tradeoffs between both of them but comparatively better one is Omega due to more benefit analyzed from it. 

2. Engine performance parameters are significantly better in Omega piston engine, this factor is crucial as they are exactly same except for piston bowl profile, 12 % more power is obtained in Omega among other factors.

3. Similarly Emissions in Omega design are either low or nearly equivalent to openW in nearly all forms except for NOx, so if primary goal is reducing NOx then OpenW has an edge here. But then again OpenW design faces issues like unburnt fuel during combustion & higher impingement, delays in combustion as very less vapourization of fuel as compared to omega design. Whooping difference of nearly 25-30% is found in Soot, 82 % in NOx & nearly 40 % in unburnt hydrocarbon. This difference is calculated with the value taken from peaks strictly and their difference is taken with high value determining the overall %. Thus if NOx can be traded off or slight modification allowed then Omega piston is a better choice. 

4. Further modification is recommended in case of openW piston, injector angle should be varied which would yeild in better vapourization of fuel spray. Similarly same can be done in Omega to see if more better performance is obtained or faster swirl is obtained. 

5. Spark - ignition and diesel engines are major source of urban air pollution & it is extremely important to keep developing methods, refining engines, creating strategies to perform complete combustion in order to reduce pollution which is ever increasing. 

It is critical to perform the spray & combustion modeling accurately in converge as that would route the entire simulation correctly, otherwise the data obtained is not useful.

 7. Grid control parameters like embedding & AMR are excellent in Converge & it should be utilized effectively as mesh determines the results accuracy & are critical to reduce over all simulation time for such extremely detailed cases. 

 8. The engine data obtained from this simulation can be taken out & analyzed in data visualizer programs automating plot post processing & calculation. One of the data visualizer & engine performance tool built by me can be checked out from here - Data Analysis of IC - Engine simulation using PYTHON 

 9. The emissions data obtained can be used to understand if it meets the criteria against new emission norms such as BS VI & redesigning, refinement can be carried out.

 

 keywords - CFD, COMBUSTION, IC-ENGINE-CFD, CONVERGE-CFD, PARAVIEW, SIMULATION, CAE

 

--------------------------------------------------------------------------------------------------- THE END -------------------------------------------------------------------------------------------------------