Torque approximation, TCU torque integration

[AndreyB]

Two of my test mules are looking for about a dozen of torque-related CAN fields for happy operation. At the moment I am shopping around for any relevant wisdom.

[AndreyB]

Leiderman-Khlystov method of torque approximation by RPM came up which seems like a USSR wisdom?

A person not from forum has contributed two implementations but both seem to not produce realistic numbers for my 2.0T Volkswagen engine.

Lua for both methods is hopeffuly self-explanatory https://github.com/rusefi/rusefi/blob/master/unit_tests/tests/lua/test_lua_Leiderman_Khlystov.cpp

The problem is that both methods produce 2000 RPM torque above max torque from engine inputs. Is that a problem with my implementation or is it that Leiderman-Khlystov is not applicable for GDI forced induction?

[mck1117]

We already have cylinder airmass. Torque is very nearly linear to cylinder airmass, so you could just configure some value "torque per airmass" in units of Nm per gram.

That is certainly a good enough first order approximation for things like transmission control.

[mck1117]

the NA version of your engine makes what, 200Nm from 2 liters (~0.6g cylinder airmass max at 100% VE, so ~0.55 with "good" real VE), so that's 363 Nm/gram

[stefanst]

In first iteration: torque at crank = torque produced - torque losses
With
torque produced being roughly proportional to airmass per cycle
and losses being equal or greater to torque needed to sustain idle

So I would go with a first iteration of (airmass/cycle - airmass/cycle(idle)) * correction factor
The correction factor being determined by manufacturer stated max torque at max torque rpm

[AndreyB]

https://github.com/rusefi/opendbc/blob/rusefi_prod/vw_golf_mk4.dbc meets https://github.com/rusefi/rusefi_documentation/tree/master/OEM-Docs/VAG/2006-Passat-B6 produces a folder full of MLV logs of OEM engine idling or pulling

https://github.com/rusefi/can-log-tools/tree/main/output

image.png (78.88 KiB) Viewed 5283 times

[AndreyB]

torque produced being roughly proportional to airmass per cycle

torque produced being roughly proportional to fuel consumption you say? the log does agree.

image.png (57.12 KiB) Viewed 5095 times