Optimization of BSFC and Nox in a 4-cylinder diesel engine using GT-Power
Objective :
To optimize parameters such as NO & BSFC for a 4-cylinder diesel engine using optimization tool in GT-Power
wer.
Description :
Optimization is the process of maximizing or minimizing a model's output(s) (responses) by systematically changing one or more input variables (factors). An optimizer is an automated process that iteratively sets the input variables, runs the model, evaluates the model output(s), and changes the inputs again to achieve some optimization objective, or until some stopping criteria are met. An optimizer uses a search method or algorithm to intelligently determine the next set of factor values in as few model evaluations as possible, since each execution of the model may be computationally expensive and time consuming.
An optimizer seeks to minimize or maximize an objective function, which is constructed from one or more model outputs. The model outputs are called responses.
The space over which the factors are varied is called the design space, design domain, search space, or search domain, and is defined by the user by setting lower and upper bounds on each factor. In GT-SUITE, because an optimization factor must be defined with a parameter, the search space is sometimes also called the parameter space.
Diesel 4-Cylinder DI Pulse Model :
The example 'Diesel_4cyl_DIPulse_InjRateMap' demonstrates several modeling concepts important for automotive diesel engines, including variable geometry turbine, boost control, intercooler, EGR circuit with cooler and EGR rate control, injection limiting for smoke control, and exhaust after treatment device modeling (geometry only, no chemistry).
Engine :
The engine is a 4-cylinder, four-stroke, direct injection, 2L diesel engine. The engine runs in "speed" mode, where the engine speed is entered by the user and the engine torque is calculated. Combustion is modeled by the DIPulse method, a predictive combustion model using the detailed injection profiles. In-cylinder heat transfer is modeled using the Woschni model. A FE model is used to predict the cylinder, head and piston temperatures. This simulation will run until it reaches steady-state convergence for six different engine speeds at full load, three part load operating points, and an additional idle point (total of 10 cases).
Intake System :
The intake system consists of the air piping upstream of the compressor inlet, and contains a simple model of an air box. The air box is modeled as a group of interconnected flowsplits that represent the total volume of the air box. At the inlet and outlet of the air box, "bellmouth" orifice connections are used to model smooth transitions.
The Intercooler :
The intercooler is found downstream of the compressor outlet. This intercooler is an air-to-air intercooler that is modeled using a "semi-predictive" method based on heat exchanger effectiveness. Note that if heat exchanger performance data is available, it is generally preferred to model the heat exchanger using 'HxMaster' and 'HxSlave' parts (which will use the available data to create non-dimensional Nusselt correlations defining heat exchanger performance). This approach will generally extrapolate better than an assumed effectiveness to conditions beyond those present in the heat exchanger data. However, if heat exchanger data is not available (or if the simulation will be run at conditions where HX effectiveness is well known), the semi-predictive effectiveness method is appropriate.
The intercooler is modeled using a pipe part. The Number of Identical Pipes attribute is used to model a bundle of 25 identical pipes of 7.0 mm diameter each (giving a large surface area for heat transfer). The pressure drop across the intercooler is calibrated using the Friction Multiplier in the in the pipe. Flowsplits are required to model the inlet and outlet of the intercooler since pipes cannot connect to pipes with the Number of Identical Pipes attribute greater than one. The desired intercooler outlet temperature is calculated by a simple control system and is imposed by actuating the wall temperature of the intercooler pipe. The Heat Transfer Multiplier in the pipe is increased to a large enough value so that the outlet temperature matches the wall temperature. The desired outlet temperature is calculated in a 'IntercoolerEff' part from the effectiveness table of the intercooler.
Design Optimizer :
Parameters used are Fue injection pressure and Start of Fuel Injection.
Optimization parameters are BSFC and NO as asked in the objective. We are setting up the case to obtain the minimum values of the both, hence minimization has been selected.ss
Post-Processing :
Case 1 : Speed - 4000RPM
NO vs BSFC contour plot :
BSFC contour plot :
NO contour plot :
Case 2 : Speed - 4000RPM.
NO vs BSFC contour plot :
BSFC contour plot :
NO contour plot :
From the above cases, we can conclude that as the BSFC decreases Nox increases, as more amount of air is used for complet combustion of A/F mixture. As the Nox decreases, BSFC increases due to incomplete combustion of the mixture during rich A/F mixture. Hence optimum operating plot s need to be used during the development process.
Significance of Injection Pressure :
More the injection pressure, quicker is the atomization and vaporization of fuel, hence easier mixing of fuel with air, making the A/F ratio more homogenous. Homogenous mixture for SI engine burns completely giving maximum performance. As seen in the above results during higher rpms injection pressure is more for proper air fuel mixing, but as speed decreases, pressure also decreases due to less amount of air intake, hence creating fuel rich zones to compensate for the loss of air and ease proper combustion.
Significance of Start of fuel injection :
As the speed increases, the process of combustion is much faster and there is less time for fuel mixing, hence injetion is advanced for complete combustion, giving less BSFC, and more Nox. Further Nox can be controlled using control strategies such as EGR or SCR.