Analysis and optimal design of a producer gas carburetor.
Air-fuel ratio characteristic exert a large influence on exhaust emission and fuel economy in Internal Combustion engine. With increasing demand for high fuel efficiency and low emission, the need to supply the engine cylinders with a well defined mixture under all circumstances has become more essential for better engine performance. Carburetors are in general defined as devices where a flow induced pressure drop forces a fuel flow into the air stream . An ideal carburetor would provide a mixture of appropriate air-fuel (A/F) ratio to the engine over its entire range of operation from no load to full load condition. Stoichiometric air-to-fuel ratio for producer gas is about 1.2 to 1.4 (on volume basis) based on the constituents of the gas. 
Producer Gas Carburetor
Mixing devices for gases used in gas engines generally referred to as carburetor, for mixing air and gaseous fuels are commonly attached to the intake manifold of an internal combustion engine. In gas carburetor the mixing of air and gaseous fuels needs to be in a proper ratio for a particular engine load and speed. Douglas  in his patented work has designed a carburetor which uses natural draw in order to preserve the energy generated by the system. The device permits start up with gasoline and a smooth transition from gasoline fuel start up to full operation on producer gas. The carburetor has a pair of concentric, stepped venturi inlets wherein the producer gas is introduced into an expanding venturi simultaneously and concentrically with the introduction of air through a reducing venturi, accelerating and commingling the fuel and air to provide a suitable mixture for introduction into the throat of a typical carburetor.
In designing the producer gas carburetor, simplicity and ruggedness have always been considered as a basic requirement to achieve easy adjustment and reproducible performance. The effective area reduction of gas and air entry holes is considered by taking a suitable coefficient of discharge. The air and fuel flow is through orifices into the mixing chamber of the carburetor which has baffle plates that enables proper mixing of air and fuel. The producer gas carburetor is being designed to have air and fuel flow at ambient conditions to be stoichiometry. The producer gas carburetor is as shown in the Figure 1 has orifices placed at air and gas inlets such that the A/F ratio at ambient flow condition should be stoichiometry for a engine suction pressure of a 25 kW engine. The amount fuel flow inside the carburetor is controlled by a butterfly valve which is located prior to the fuel inlet orifice. The pressure balancing electronic controller drives suitably the butterfly valve with the help of a motor that brings the valve for a null pressure differential across the manifolds for the fuel and air attached upstream to the main engine manifold and works in suction pressures. If the differential pressure at both the carburetor manifolds are maintained at zero, with the manifolds tuned for their effective flow areas to match the ideal mixture condition, then the mixture flow what one get at engine intake manifold will be stoichiometry. The Figure 1 shown is the geometric model of the producer gas carburetor designed and analyzed for optimal pressure drop with good mixing ability.
CFD Simulations are carried out on the producer gas carburetor geometric model as shown in Figure 1. The air and producer gas passes through inlet pipe of 50 mm diameter in the axial flow into mixing chamber through an orifice of 28.5mm and 26mm diameter respectively. Producer gas inlet pipe is provided with butterfly valve and geometric model for 0 to 90 degree in steps of 15 degree is built to carry out analysis at different valve openings. For different mass flow from 0.005kg/s to 0.05 kg/s insteps of 0.005kg/s at the carburetor outlet condition has been tried out. The Mesh geometry shown in Figure 2 is a 3- Dimensional model used to simulate the flow analysis in a carburetor with mesh grid density of around 270000 computational nodes is considered. The grid is a unstructured mesh with a maximum mesh grid size of 18mm. Inflated boundary meshes are created for the butterfly valve body and near inlet orifices with first layer mesh size of 1.8mm with an expansion factor of 1.2.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
Boundary and Initial Condition
The flow domain considered for simulation is the whole carburetor assembly with steady state flow. Here for simplicity of analysis a single gas (carburetor gas) entity is considered having air ideal gas and producer gas. The relative pressure of the carburetor domain is assumed to be 1 atmosphere with non buoyancy condition. For air inlet boundary condition mass and momentum, static pressure equivalent to domain reference pressure is set with flow condition being subsonic. The initial condition of flow through the air inlet with air ideal mass fraction as 1 is considered. The initial boundary condition for fuel inlet is same as the air inlet except for the flow of producer gas mass fraction being 1 at the inlet. The boundary condition for carburetor outlet is of different mass flow rate which is to be simulated is considered.
Geometry Optimization Of Differential Pressure Carburetor Model Using CFD Analysis
The producer gas carburetor that is being analyzed has been designed for a 25kWe engine capacity, to suit an existing testing facility. Different geometric configurations are tried with computational analysis so that the carburetor generates optimal mixture. The carburetors investigated earlier are presented with flow analysis captured by CFD analysis for three cases.
Initially (case1) the carburetor has been modeled with only one baffle plate at the centre of the mixing chamber as shown in Figure 3 . The carburetor inlets are designed to have orifices placed at air and gas inlets such that the A/F ratio at ambient flow condition should be stoichiometry for an engine suction of a 25 kW engine. The fuel flow to the carburetor is controlled by a control valve which is located in the upstream of the fuel orifice.
[FIGURE 3 OMITTED]
Figure 4 provides a contour plot with producer gas mass fraction for the geometry of case 1. The producer gas mass fraction at the exit is estimated to be varying in the range 0.3 to 0.6 than a desirable value of 0.5 uniformly. The mixing flow pattern at a plane perpendicular to the flow near the outlet of the carburetor with plot showing the contour for producer gas mass fraction, for a total mass flow rate of 0.025kg/s and with the condition of the control valve of gas inlet being fully opened (90[degrees]).
[FIGURE 4 OMITTED]
In this carburetor model (case 2) two baffle plates are being placed as shown in Figure 5, to get enhanced mixing as compared to the earlier case. The geometry considered is still simple enough that can be fabricated with much ease.
[FIGURE 5 OMITTED]
Figure 6 shows the flow pattern for case 2 carburetor in a 3 Dimensional view with a contour plot on a plane central cross section computed for an outlet mass flow rate of 0.025 kg/s. Figure 7 shows the mixing flow pattern at a plane perpendicular to the flow near the outlet of the carburetor with plot showing the contour for producer gas mass fraction, for a total mass flow rate of 0.025kg/s and with the condition of the control valve of gas inlet being fully opened (90[degrees]). The computations for the flow are seen to have reached to convergence at 148 iterations, with the residuals falling to less than [10.sup.-4]. In comparison to the previous configuration, there is a significant improvement in the mixing. However, as one can see, at the exit port still a fairly heterogeneous pattern of the mixture is visible indicating that this configuration also does not adequate homogeneity in the mixture at carburetor outlet and can not considered to be optimal in its performance.
[FIGURE 6 OMITTED]
The computations for the flow are seen to have reached to convergence at 134 iterations, with the residuals falling to less than [10.sup.-4]. It can be clearly seen that the present configuration does not achieve the mixing good enough to be considered as optimal.
In this carburetor model (case 3) two baffle plates are placed as in the earlier case and also a semi cylindrical plate at the entry of the outlet pipe of the carburetor is being placed as shown in Figure 8, to get enhanced mixing as compared to the earlier two cases. The geometry considered is still simple enough that can be fabricated with much ease. The same configurations have been optimized as the flow pattern obtained in these configurations shows a homogenous mixture generated at the outlet of the carburetor outlet. Detailed analysis of case 3 configuration is been discussed in the next section.
[FIGURE 8 OMITTED]
Figure 9 (a) and (b) represents the result of the most optimal set obtained that has three-dimensional and two-dimensional view of the carburetor. Figure 10 (b) shows the contour plot of the mass fractions within the mixture over a symmetric cross-sectional plane. One can see that as the flow passes across the orifices, baffle plates and the semi-cylindrical plates and before it gets out of the carburetor, the mixture is found to be getting more uniform and near the exit, the producer gas mass fraction is seen to be close to 0.5 uniformly, representing the most desirable condition for the carburetion. Figure 10 (c) shows the producer gas mass fraction lies around 0.5 of the mixture before it leaves the carburetor. The figure represents the flow pattern when the solver has converged with the residuals of all the variables have diminished to well within the acceptable limits (typically around [10.sup.-4]).
[FIGURE 9(a) OMITTED]
[FIGURE 9(b) OMITTED]
[FIGURE 10 OMITTED]
Experimental analysis of Differential pressure controller based producer gas carburetor
The differential pressure producer gas carburetor has been tested for transient state operation for 100 % load throw off condition. Experimental analysis carried out for around 3 minutes when the carburetor is fitted to an engine for transient state operation under 100% load throw off condition. Initially for 20 seconds the engine is run at no load condition and after 20 seconds the engine is electrically loaded for 20.56 kW at one instant and continued to be at that load for 70 seconds and then once again engine is brought back to no load condition. The Figure 11 shows the response curve of differential pressure sensing carburetor for 100% transient load throw off condition. The air to fuel ratio is around 1.2 at no load condition and at peak load it is 1.5. The differential pressure at transient 20.56 kW load condition is around null pressure.
[FIGURE 11 OMITTED]
The frequency of the engine is maintained at 50 Hz with a small deviation of +/- 2 Hz. This shows that the engine speed has been running at the rated speed of 1550 rpm except at load change off position wherein the engine has recovered fast at load throw off conditions. The control valve opening seems to be around 20 to 30 degree of opening. . It can be seen that the air to fuel ratio is around 1.2 at no load condition and at the peak load it is maintained at around 1.5. It was observed that the mass flow rate of fuel was 8 to 10 g/s and that of air to be in the range of 10 to 12 g/s at peak load conditions. The results confer that the differential pressure controller producer gas carburetor is able to maintain optimal air to fuel ratio even at 100% load throw off condition and can deliver stoichiometric ratio to engine inlet manifold at transient load operation.
The transient response studies were conducted at intermediate loads as well, for getting a clarity that the systems behave consistently and stably and at all loading conditions. Response data of the differential pressure carburetor were obtained at similar test conditions and procedures as explained for the full-load test, for transients of loads of 80%, 60% and 25% of the full-load. The figures12, 14 and 15 show the response curves for these sets respectively.
Figure.13 shows the response curve of differential pressure controller producer gas carburetor for a steady state operation for no-load condition. Experimental analysis when carburetor is fitted to an engine and a steady state analysis under no-load condition carried out for around 3 minutes shows that air--fuel mixture ratio between 1 and 1.2. Control valve opening found to be around 50 degree opening and the frequency to be around 50 Hz +/-0 .3 Hz. This shows that the engine speed was running at the rated speed of 1500 rpm.
[FIGURE 12 OMITTED]
[FIGURE 13 OMITTED]
[FIGURE 14 OMITTED]
[FIGURE 15 OMITTED]
Results and discussion
While comparing the data from the experiments and computations as seen from Figures. 16 and 17, it comes out that there is a larger deviation in the initial (15 to 25[degrees]) portion of the valve positioning and gets closer at the larger opening. This is attributed after an examination that certain mismatch in the contour edge of the valve disc used in the experiments due to the inaccuracies while machining it and during CFD simulations, the same has been assumed to be a clean elliptic edge making a the source of error. Further it could be interpreted for a butterfly valve that with the initial portions of the valves opening, where the blocking factor would be higher compared to its wide open conditions, the error contribution from comes down at larger openings of the valve, as also can be seen from the plots. While the trends of the curves are fairly well agreeing, considering the source of error in the experimental setup, the deviation in the prediction gets explained and it could be inferred that the results imply that the verification of CFD results in terms of its consistency on these sets are acceptable and the physical features of the operations has been captured. The table 1 below shows the result obtained from CFD analysis for different outlet flow rate conditions and for different control valve openings for inlet boundary conditions for both inlets to be 1 atmosphere. A characteristic curve of the air to fuel ratios obtained as an outcome of the study is shown in Figure.17.
The flow field captured by the CFD analysis, taking account of turbulence, in the 3-D domain of the carburetor is shown in the figure 15. The velocities on a plane just before the outlet of the carburetor is around traced to be about 5 m/s and at the centre of outlet on the plane to be around 11 m/s. Two vortices are traced to have formed just upstream of the outlet, due to the restriction by the semi-cylindrical plate introduced between the second baffle plate and the outlet. This arrangement has derived enhanced mixing of the gases and in meeting the desired criteria.
The plate 1 shows the flow pattern in a carburetor that is a photographed picture of a flow partical image velocimetry visualization technique at frequency of light chopping 2.5HZ. The Particle Image Velocimetry (PIV) is a non-intrusive, full field optical measuring technique, which is used to obtain velocity information about fluid motion, is one of the steps in devising experimental verifications in the analysis during the designing the new carburetor with differential pressure sensing, that enables studying of the flow patterns and to get an account of recirculation inside the carburetor. . The flow visualization study was carried out with table top test model made out of acrylic sheets. The experimental results compared with CFD results indicate that the flow patterns captured by the CFD analysis are close to the patterns captured from the experimental results as shown in figure 18and 19, implying that verification for the CFD analysis is complete with consistency among the two. 
[FIGURE 16 OMITTED]
[FIGURE 17 OMITTED]
[FIGURE 18 OMITTED]
[FIGURE 19 OMITTED]
The results obtained of differential pressure carburetor from the engine test run conditions shows that the controller has maintained stoichiometric mixture at different constant load conditions and at load throw off conditions. The non availability of an efficient, reliable, economical and commercial marketed producer gas carburetor was one of the major hurdles in establishing a 100% producer gas based power generation system. This problem has been successfully addressed. The producer gas carburetor developed has been tested leading to bringing out of an optimal design of the differential pressure Producer Gas carburetor used for prototype testing and for real-time testing.
 Klimstra J "Carburetors for Gaseous Fuels--on Air to Fuel ratio, Homogeneity and Flow restriction. SAE paper 892141
 Versteeg H K, and Malalasekara W An introduction to CFD-The finite volume method, 1995 (Longman Scientific and Technical)
 Anil T R, Tewari P G, and Rajan N K S " An approach for Designing of Producer Gas Carburetor for Application in Biomass based Power Generation Plants proceedings of the national conference of Natcon 2004 Bangalore"
 Anil T R, Ravi S D, Shashikanth M, P G Tewari and Rajan N K S. "CFD Analysis of a Mixture Flow in a Producer Gas Carburetor" International Conference On Computational Fluid Dynamics, Acoustics, Heat Transfer and Electromagnetics CFEMATCON-06, July 24-25, 2006, Andhra University, Visakhapatnam--530003, INDIA
 Ronald J Adrian, "Particle -Imaging Techniques for Experimental Fluid Mechanics" Annu. Rev. Fluid Mech. 1991. 23: 261-304.
 Douglas G Janisch patented work on producer gas carburetor; US Patent 5070851 December 10 1991,www.freepatentsonline.com
 Sridhar G, Shridhar H V, Dassapa S, Paul P J, Rajan N K S and Mukunda H S, "Development of Producer Gas Engine", Published in Institute of mechanical engineers, 2005.
T.R. Anil (1), P.G. Tewari (2) and N.K.S. Rajan (3)
(1) Professor, Dept. of Mechanical Engg. GIT Belgaum. E-mail: firstname.lastname@example.org
(2) Professor Energy Systems, Dept. of Mechanical Engg. BVBCET Hubli. E-mail: email@example.com
(3) Principal Research Scientist CGPL, Dept of A.E IISc. firstname.lastname@example.org
* Corresponding author E-mail: email@example.com
Table 1: Details A/F obtained from CFD analysis for 15 to 90[degrees] valve opening and for outlet mass flow rates ranging from 5 g/s to 50 g/s. Set Control Valve Air--Fuel Ratio Position 5 10 15 20 25 (in degree) (g/s) (g/s) (g/s) (g/s) (g/s) 1 15 4.10 4.03 4.05 4.00 4.00 2 30 2.33 2.30 2.29 2.33 2.33 3 45 1.63 1.64 1.63 1.63 1.63 4 60 1.34 1.33 1.33 1.29 1.32 5 75 1.26 1.26 1.26 1.26 1.27 6 90 1.31 1.25 1.29 1.28 1.28 Set Air--Fuel Ratio 30 35 40 45 50 (g/s) (g/s) (g/s) (g/s) (g/s) 1 4.54 4.00 4.05 4.23 4.43 2 2.34 2.30 2.31 2.31 2.68 3 1.67 1.63 1.63 1.63 1.67 4 1.28 1.33 1.28 1.32 1.31 5 1.27 1.26 1.27 1.25 1.26 6 1.26 1.27 1.27 1.27 1.27