Week-7 : Converting a detailed engine model to a FRM model
AIM
* To explore tutorial 9 which in detail explains how to convert an existing model into a fast-running model (FRM).
* To build an FRM using the FRM builder approach with
bore = 102 mm
Stroke = 115 mm
Compression ratio (CR) = 17
CI engine, Twin Scroll Turbine, GT controller
INTRODUCTION
FRM is an abbreviated form of Fast running model, as far as GT-suite models are concerned throughout we have run many models whether it is pertaining to the engine, after treatment or engine + after treatment, turbochargers, etc, all these are detailed models. When the simulation is run for different cases in any of the detailed models it takes a considerable amount of time, being a detailed model there will be a lot of components, for example, if we take an intake manifold in a detailed engine model it is being modeled with the number of circular pipes, orifices, and flow split, so when the simulation is run the flow equations (Navier stokes equation) needs to be solved in each of these components and hence it becomes time-consuming. To compensate for the time consumption Fast running models are developed, in a broad way we can put it like FRM sums up different components into one component with simplified geometry and attributes. If we take the same example of an intake manifold, instead of circular pipes, orifices, and flow split in FRM it will be a single entity or a geometry called an intake manifold.
When FRM is used instead of a detailed model, of course, we have an edge over the simulation time, and CPU time but at the same time the accuracy of the results will be a bit compromised also the fidelity of the model will be low. The purpose of building an FRM is that when the software in the loop, hardware in the loop, etc testing is being done there will be a real-time ECU or any software plugged into the models, if the detailed model is used it will be difficult to match an up with the real-time hardware or software so FRM comes handy.
Exploring tutorial number 9:
* The tutorial explains in detail how the high fidelity (detailed model) can be converted into an FRM using the FRM converter tool. FRM converter is a handy tool provided by GT-suite which keeps track of the entire conversion process.
* First we need to select a detailed model so as to convert it into FRM, this tutorial number 9 considers the example case SI_4Cyl_GDI_Turbo_WOT, a snip of the project map of this detailed model is shown below.
* We can either convert this detailed model into FRM based on accuracy or speed, in this tutorial the simplification is done based on accuracy so as to compare the engine performance parameters from the detailed model as well as FRM and to determine the accuracy.
* Detailed model which is open in GT-suite is first done FRM tagging, this option can be seen under the tools tab. It tags all the subsystems in the detailed model. We can start by simplifying the exhaust system, since high-velocity flow is more at the exhaust this affects the courant number and hence simplification is started from the exhaust manifold.
* Before we apply simplifications to the exhaust manifold, the maximum time step limit of 1 crank angle degree will be removed. This is a default limit that is imposed when an Eng Cylinder component is in the model. If the courant condition allows for a larger time step to increase in size. When the exhaust system is simplified it looks as follows,
* Similar to the exhaust system exhausted from the turbine, a snip of the turbine exhaust system from a detailed model is shown below,
* This system is simplified in the same way as that of the exhaust system of the engine. When simplified we get the exhaust from the turbine as,
This completes the second step in the conversion of the detailed model to FRM and hence we have the complete project map of the second step with simplified exhaust and turbine exhaust.
* Next is a simplification of the intake manifold, the steps followed are the same as followed for the exhaust system, but minor modifications need to be made to the characteristic length and expansion diameter attributes value for the combined flow.
* The characteristic length is defined as the distance that the flow entering the boundary will travel before crossing another boundary or impacting a surface. Boundary #1 represents the boundary between the IM1_pipe1-1. The fluid entering the flow split in this direction can travel all the way across the plenum before impacting a surface, so the characteristic length is equal to the length of the plenum.
* The other boundaries are adjacent to the runners, where backflow from the cylinder would travel across the diameter of the plenum before impacting the plenum wall, which is why they are given a characteristic length of 70 mm.
* The expansion diameter defines the maximum diameter that the flow may expand to after entering a boundary. For boundary#1 it is simply the diameter of the plenum because the fluid expands from IM1_pipe1-1 into the plenum in a straight direction. The expansion diameters for the other boundaries are equal to the plenum diameter multiplied by 1.4. This is because as the fluid backflows into the plenum, it encounters a wider cross-section than the plenum diameter to expand into. A snip of the modification window is given below,
This completes the step3 of the conversion, a snip of the same is given below.
* Next is a simplification of compressor outlet pipes, select boost pipes 1 from the tagged subsystem list and click on simplify by accuracy. Since the intercooler core has been replaced with a flow split and HeatExchangerConn part. The BP1_CACC1-2 HeatExchangerConn part will be used to assure the intercooler outlet temperature is maintained.
* When the compressor outlet pipes are simplified this completed the step4 and hence we have
* Intake pipes select intake pipe 1 from the tagged subsystem list. Click on simplify for accuracy. Collapse the FRM converter to provide space to work in the model map. This completes the simplifying process. We have some things like the below after step5
* We started with simplifying the exhaust system on an accuracy basis now we can further simplify the exhaust system on a time step basis, by doing this we get a much-simplified exhaust system as shown below,
* Final step or simplifying is cylinder slaving. By right-clicking on any of the EngCylinder slaving options. For full cylinder slaving to function, each valve must be specified as either an intake or exhaust valve.
So above tutorial gives us an idea of how the existing detailed model can be converted.
* To build an FRM:
The second part of this challenge is to build an FRM with the following parameters,
bore = 102 mm
Stroke = 115 mm
Compression ratio (CR) = 17
CI engine, Twin Scroll Turbine, GT controller.
The main purpose of building an FRM is to check how much the simulation time, as well as CPU time, varies, so let us first take a detailed CI model which has a turbo and a controller.
Let us consider the Diesel_6Cyl_TC example case, this model has a twin turbocharger, a control unit, and 6 cylinders. This example resembles much of the FRM we are building.
Simulation is run at engine speed = 1500 rpm, below is the snip that shows the other parameters. Even this detailed 6-cylinder engine model is set with the above-mentioned bore, stroke, and compression ratio values.
Following is the time taken for the simulation to run,
Now let us build an FRM and then run the simulation,
Under the tools option, the FRM builder option is available, when we click this option we get the below window where CI engine is selected.
* When we click next we get attributes that define bore, stroke, and compression ratio. Below is the snip of the window that shows the attributes defined,
* When we click next we get a window where the type of turbo is set,
* By clicking next we define the number of cylinders,
* In the last window the controller type is set,
This gives us an engine model as shown below,
Results obtained as follows,
* If we compare the time taken to complete the simulation in the detailed model as well as FRM it is seen that the detailed model took less time and FRM took more time, but the more time taken by the FRM is because of a number of periods for which the simulation is run. FRM is run for 128 periods and the time taken for the same is 10.24 sec. But the detailed model run for 75 periods and the time taken for the same is 6 seconds if the detailed model would have run for 128 periods the time taken would have been more than 10 seconds.
* Comparing the CPU time even after running 128 periods of simulation FRM took only 37.79 min but the detailed model that run for 75 periods took 1:53:56, which is almost 2 hrs of CPU time.
* From the above result, it is observed that for Sil, Hil, and Mil type of simulations using FRMs will be a good option.