MATLAB Answers

0

IC engine in Powertrain Blockset is able to simulate transient emissions?

Asked by Felix Pult on 24 Apr 2019
Latest activity Answered by Felix Pult on 9 May 2019
Hello together,
I want to simulate the transient emissions of a turbocharged IC engine. Therefore I found the ready made simulation model of an IC engine (https://de.mathworks.com/help/autoblks/ug/ci-engine-project-template.html?s_tid=srchtitle) which I combined with the IC controller (https://de.mathworks.com/help/autoblks/ref/cicontroller.html).
When I run a simulation with a load jump as load signal, the model simulates the fuel consumption, the emissions (except particulates) and the other important things of an IC engine, but I can not see any transient effects that should appear.
So my question is: Is this IC engine simulation model is able to simulate transient emissions or are only steady-state emissions possible? In the User's Guide of the Powertrain Blockset (https://de.mathworks.com/help/pdf_doc/autoblks/index.html) I think I read something to the transient emissions on the pages 3-12....
So thank you in advance for your help!

  0 Comments

Sign in to comment.

2 Answers

Answer by Mike Sasena on 8 May 2019
 Accepted Answer

Hello Alex,
The engine-out emissions of the CI Core Engine model referred to in your question responds dynamically to dynamic changes in the following control inputs to the engine: fuel mass, boost pressure, intake manifold gas temperature, intake oxygen percent, and fuel pressure to the extent that those inputs affect engine dynamic engine torque. The CI Core Engine model then uses the dynamic engine torque together with engine speed to look up quasi-steady engine-out emissions from lookup tables based on warmed up conditions. Our engine-out emissions can be calibrated to take into account air system and fuel system dynamics given a fixed engine calibration in the CI Controller, but not cylinder wall warmup dynamics.
If you want to focus on transient emissions for aftertreatment control and calibration, and account for cylinder wall warmup dynamics, you can make empirical dynamic emissions models using our Mapped CI Engine block, which allows a temperature input to affect emissions as the engine system warms up. You can fill 3-D lookup tables in the Mapped CI Engine block from measured engine-out emissions data using engine speed, fuel mass/inject, and coolant temperature/cylinder wall temperature as inputs. A coolant temperature/cylinder wall temperature model can be made using our Simscape family of products. Of course this empirical approach involves laboratory access and labor to calibrate the 3D lookup tables of the emissions.

  2 Comments

Hello Mike,
Thank you for your answer!
I'll ask you one more time, so I understand you correctly:
So in any case the model will read out the emissions out of a map. I can calibrate this map, but there is no "transient" calculation that simulates me my emissions?
Yes, the emissions maps themselves for CI Core Engine contain steady-state data, but the inputs to those maps are dynamic in response to driver demand (torque and speed), making the outputs of the maps dynamic in a gross sense.
For example important diesel emissions like engine-out NOx will move dynamically as engine torque moves dynamically, assuming instantaneous emissions equilibrium for each increment in torque. This sort of model is suitable for many types of aftertreatment control development activities since aftertreatment devices are well downstream of the engine.
In fine terms, however, emissions like NOx do NOT move directly in CI Core Engine in response to raw engine control inputs such as fuel injection timing, only secondarily as fuel injection timing effects torque. If the engine is calibrated then the movements of the raw engine control inputs are highly cross-correlated (by controller lookup tables) to the commanded torque, so the lack of direct connection between raw engine inputs and torque is often not of great consequence.
So if you want to control (and therefore model) engine-out emissions like NOx on a cycle-by-cycle basis, our model isn’t the right type of model for that. However:
  1. Emissions are not typically controlled on an engine cycle-by-cycle basis in production applications, because sensors and actuators are not available to control emissions on an engine cycle basis (excepting in-cylinder pressure sensors, which may see production at scale eventually).
  2. Cycle-by-cycle engine emission models to-date have been notoriously inaccurate compared to mapped emission models (current Powertrain Blockset approach) from measured data, and even cycle-by-cycle transient emission models require extensive calibration to real data.
It is important to define what you want to do with the emission outputs before judging their suitability. In Powertrain Blockset our purpose is to provide emission models for the purpose of drive-cycle vehicle-level emissions estimates and aftertreatment control design, and for that purpose a map-based approach is not only suitable but has been the industry norm for several decades due to the lack of suitability of alternative engine cylinder-based transient engine-out emission models.
Hope that context helps,
Mike

Sign in to comment.


Answer by Felix Pult on 9 May 2019

Ok, I think now I got it.
Because my goal is to investigate the behaviour of transient emissions regarding to different transient torque gradients I think your model is not the right tool.
But thank you anyway for your help!
Felix

  0 Comments

Sign in to comment.