Optimization Technique using GTP

1.Explore Example- Diesel_4cyl_DIpulse_injratemap and write a detailed report

Let's explore the example of the software:

The model represents a 4-cylinders CI engine, direct injection, turbocharged with Westgate valve control, the total displacement of 2L, cooled EGR, DOC, and DPF (chemistry wasn't considered for the after-treatment systems).

INTAKE SYSTEM

The intake system is composed of inlet pipes and an airbox, upstream the compressor, and an intercooler with the other part of the intake system, till the cylinders, downstream the compressor. The airbox is modeled with interconnected flow splits, as shown below:

CYLINDERS

The main tab of the cylinders is shown below:

A "finite element model" is used to predict cylinder, head and piston temperature, called: "TWallSoln-Cyl1-5"; a DIPulse combustion object is used and an in-cylinder heat transfer is modeled using WoschiGT:

The object used for the injector allows defining different profiles, as shown below into the "profile tab", where injected mass, injection timing and the pressure profile are defined:

The quantity of fuel injected every cycle is controlled by the "controllerDInj-1", receiving from the engine the value of rpm, BMEP, and the inlet air mass flow, to reach the BMEP target.

ENGINE

Inside the engine, the cylinder geometry is defined:

as visible from stroke and bore data, it is a perfect square engine, so it's plausible to think that it will run not fast (low rpm).

EXHAUST SYSTEM

The exhaust system is composed of pipe rounds that are directly linked with the turbine. But before arriving at this turbomachinery, there is the EGR valve, that allows the recirculation of exhaust gases. It's controlled by a valve actuated by EGR controller, that works receiving the following signals: EGR mass flow rate and air mass flow rate (from the intake system). An articulate EGR cooler is used:

The turbine is a no-twins scroll turbine, but it has a Variable geometry controller that choose how the turbine must work and which maps use:

apparently seems that no different maps are used for rack positions, but clicking on the map object is possible to visualize that there is a turbine map for every rack array defined:

After that the exhaust gases pass through the after-treatment systems:

only geometrically and not chemically modeled.

In the end, there is a muffler, modeled with two different main elements: flow splits and flow splits absorbing, which represents a volume with general absorbing material:

1. Run all designs for a case and find the optimum values

Parameters-SOI, Rail Pressure

Response- NO and bsfc minimum

Comment on your observations

For case 1, shown below, the optimum values have been found:

The parameters that can be changed to obtain the lowest value of NO and BFSC, are P_Rail and SOI_PRE.

The P_RAIL represents the pressure of the injected fuel.

The SOI_PRE represents the start of the first injection of the fuel. In fact, there are two different quantities of fuel injected sequentially.

Let's visualize the optimizer set-up:

an integrated designer optimizer is chosen, multi-objectives (Pareto), optimizing every case independently (but only one case is run).

The factors tab is shown below, choosing the lower and upper limit of the factors (SOI_PRE and P_Rail):

And the responses tab, showing the targets: minimizing the NOx generation and BSFC:

After completing all the 150 simulations, the results are discussed:

RESULTS

In GT-Post the results are discussed. At first 2d Pareto is analyzed, comparing the values of NOx with BSFC:

This curve of points represents the best simulations obtained. It's clearly visible that having a low NOx generation causes a very high BSFC, causing the formation of other pollutants; in the same way, having a very low BSFC causes a very high formation of NOx. So, it's not possible to define the optimum situation having low BSFC and low NOX, but it's necessary to do a trade-off between these two quantities.

Another important aspect that needs to be analyzed is the sensitivity comparison for the following parameters:

Responses vs Factor

Start of injection vs Pressure rail

Start of injection vs Pressure rail

It's clearly visible that the most important parameter that influences the NOx formation and BSFC is the P_Rail.