Explore FRM model, conversion of detailed model into FRM model using FRM converter and build the FRM model using FRM builder tool in GT-POWER

Objectives: To explore FRM (Fast Running model) model tutorial and to build the FRM model using the FRM builder approach for the following specifications.

1. Bore- 102mm, Stroke- 115mm, CR- 17
2. CI engine, Nos. of cylinders- 6
3. Twin scroll turbine
4. GT Controller

Given: FRM converter tutorial no 9 which includes 8 steps given. Create FRM using the FRM builder approach for 6 cylinder CI engine having a Twin scroll turbine and GT controller.

Theory:

FRM (Fast Running Model)
FRM stands for the Fast Running Model. Whenever simulation speed is of the high priority we require the FRM model it can be done on a high fidelity GT-POWER model by simplifying it using the standard conversion process mentioned in GT-POWER or we can directly build FRM model with the required configuration. The high fidelity model is often too slow in running the simulation in which it solves transient cycles where cycle time is very high so we cannot afford a detailed model simulation approach.
FRM models are fully physical, dynamic and these models are designed specifically to run fast. Fully physical means we can capture physical data from the simulation e.g. in-cylinder pressure of engine where we get Pressure vs Crank angle diagram. FRM is also an equation-based model where output values are governed by input values or based on input values output will change.

There are two different methods by means of which we can achieve fast run in FRMs: 1. increasing the simulation time step size. 2. Decreasing nos. of calculation per time step.
The above methods can be achieved by combining various flow volumes together which will reduce nos. of subvolumes and also allow for larger time step size by increasing effective subvolume length which is responsible for larger time step size. Besides simplifying and combining flow volumes, there are additional flow solver options that can reduce nos. of calculation per time step and, thus further decrease the run time without changing the time step size.

Tutorial No: 9:

In this tutorial, the process of creating a Fast Running Engine model from a detailed, high fidelity engine model described. Initially, it starts with SI 4 cylinder turbocharged GDI engine. The turbine wastegate is being dynamically controlled in order to target BMEP. A full-loa sweep is being simulated, from 5000 RPM to 1000 RPM. Run this model and generate the results that will be used to calibration and comparison of the FRM.

Before going to the conversion process we need to make sure what are the different parts limiting the time step. It indicates lower the value, the more likely the part will limit the time step. This will help to determine multiple parts that will limit the timestamp. These values we will get in GT-POST by going to this path Select RLT Contour Map -> Pipes/FlowSplits -> Flow Control -> "Minimum Courant Time Step (Explicit Circuits Only)".
Next, we need to take care of is to check nos. of subvolumes for each pipe modeled. The more subvolumes are used, the execution time will be slower. The Number of Subvolumes can be selected under Pipes/FlowSplits > Flow Control > Number of Subvolumes.

We are starting with a high-fidelity detailed model in GT-ISE, FRM tagging needs to be done.

FRM tagging has been done as follows:

1. Boost pipes
2. Exhaust Manifold
3. Exhaust pipes
4. Intake Manifold
5. Intake pipe

Launch FRM converter tool to convert detailed models and monitor Key RLT results throughout the conversion process.

In order to validate FRM results, we can add key RLTs to monitor or track the results during the progress in the "Accuracy Checking" dialog box with specified tolerances.

STEP-1: Exhaust Manifold

This is the very first subsystem of the model we are going to simplify. For detailed, high-fidelity models always an exhaust manifold where the velocities will be higher hence it will turn restricts the time step.

Step Name: Exhaust Manifold
Tagged Subsystem: Exhaust Manifold 1
Simplify For Accuracy- Here we are interested in the accuracy of the model. It will combine exhaust port/runners together and keep separate from exhaust manifold volumes. If we required a real-time capable model choose "Simplify for Speed" where port, runners, and manifold volumes will be combined together at once.

Run the calibration and we get calibrated values then run the complete case setup by considering all cases. We can see how simplification at exhaust manifold has been done.

After doing the calibration check and verifying the results are satisfactory, we can proceed further to the next step.

STEP-2: Exhaust Pipe

For model simplification, we will take the exhaust pipes downstream the turbine.

Step Name: Exhaust Pipe
Tagged Subsystem: Exhaust Pipe 1
Simplify For Accuracy

Exhaust pipes have been converted into the CAT part.

Run the calibration and check for calibration results to assure satisfactory results.

STEP-3: Intake Manifold

Step Name: Intake Manifold
Tagged Subsystem: Intake Manifold 1
Simplify For Accuracy

Some modifications need to be done on characteristics length and Expansion diameter attributes value to combined volume.

Run the calibration and check for calibration results to assure satisfactory results.

STEP-4: Compressor Outlet Pipes

Step Name: Compressor Outlet Pipes
Tagged Subsystem: Boost Pipe 1
Simplify For Accuracy

Reconnect the Control links, sensing Upstream Temperature from "BP1_Pipe1-1" Pipe and Upstream Mass Flow Rate from the "6” OrificeConn upstream of the "BP1_CAC1-1" Flowsplit, and sending the actuator output to both "BP1_CAC1-1" Flowsplit (actuating Wall Temperature) and to the HeatExchangerConn (actuating Imposed Fluid Temperature).

Run the calibration and check for calibration results to assure satisfactory results.

STEP-5: Intake Pipes

Step Name: Intake Pipes
Tagged Subsystem: Intake Pipes 1
Simplify For Accuracy

Since every subsystem has now been simplified, we will assure that there are no remaining pipes with short discretization lengths. Go to the Data tab in the GT-ISE toolbar and click "Delete Unused Objects".

Run the calibration and check for calibration results to assure satisfactory results. Here, we have simplified all the subsystems hence, at this time we can see simulation run time data. We can observe plot below where simulation run time has plotted for both original and FRM models where we can see a drastic drop in a real-time factors.

STEP-6: Additional Changes

Step Name: Additional Changes
Tagged Subsystem: Exhaust Manifold 1
Simplify For Speed

In this step we are interested in the speed of simulation run hence all exhaust runner volumes are now merged into single flow volume.

Run the complete case setup and verify the results we got.

Here, we can how drastically FRT has drop to close to 5 in a later case for all operating range.

STEP-7: Cylinder Slaving

In this step of simplification, we are utilizing the cylinder slaving capability of the EngCylinder component.

Step Name: Cylinder Slaving

In this case, one master cylinder has been chosen for calculation and rest of the cylinders will follow this master cylinder.

This Cylinder slaving further helps in reducing the factor of real-time close to almost 2 for all operating conditions with high accuracy as we can see above.

STEP-8: Real-time solver

As a name suggest it real-time solver will require an RT license to run the real-time solver. However, certain templates are not allowed with real-time solver, and so they must be removed first.
Since we don't have a GT-SuiteRT license we are not able to run this.

In this way, we have reduces the FRT of the original detailed model and bring down to the Fast running model by simplifying all the subsystems. Along with fast simulation run time, we have achieved a great level of accuracy in the results which are feasible.

Build FRM model Using FRM builder tool:

This is another approach to build a complete Fast Running Model with the help of a series of steps only.

STEP-1: Create New engine model.
STEP-2: Choose Engine configuration

1. SI- Spark Ignition
2. CI- Compression Ignition

STEP-3: Provide Cylinder geometry

1. Bore
2. Stroke
3. Compression Ratio

STEP-4: Choose a turbo type.

1. Twin Scroll
2. VGT

STEP-5: Choose an engine type.

1. 4 cylinder
2. 6 cylinder

STEP-6: Choose a control logic type

1. GT-SUITE Controller
2. Coupled to Simulink

We have to choose the engine configuration according to the given specifications.

• Bore- 102mm, Stroke- 115mm, CR- 17
• CI engine, Nos. of cylinders- 6
• Twin scroll turbine
• GT Controller

Results:

Here, we have run a total 25 cases by considering various load targets, BMEP targets, Boost pressure targets at different engine speeds.

We have plotted some of the important engine parameters by considering various load conditions at different engine speeds to obtain the desired BMEP. In order to do that we have fixed boost pressure or compressor outlet pressure depending on load requirement.

Brake power and torque requirement have been adjusted according to load conditions we want to operate.
We can see that BSFC and Air-Fuel ratio is very high for 20% load condition at all engine speeds. In order to limit in-cylinder temperature which succumbs NOx formation, the engine operates at a rich mixture at higher engine speed i.e. 4000 RPM.

Average Run time taken to complete simulation for all cases is presented below:

We are interested in Run time of the simulation hence we made gt model using FRM Builder approach. We can see above, FRM plays a significant role in reducing simulation time without loss in model fidelity.