Diesel engine simulator for starter motors

[istvanb]

Hello Community,

I need to build something like this:

http://youtu.be/zbXRRQmf0jI

The bottomline is that I'd like my prime mover (the big electrical motor I will control) to force the unit under test to follow a certain speed profile and by this simulating the stress it receives on a real combustion engine. The prime mover will be controlled by an Emerson made drive, which has two control modes: speed and torque. I believe I understand what is "speed control" and I believe thats what I need. (I have the speed profile recorded and this is what I want to apply on unit under test). But what does the "torque mode" mean exactly?

This is quite a project, I am looking forward to get your thoughts. Does somebody here has hands on experience with Emerson drives?

Thanks,
i-

 

[BrianPetersen]

In "speed" mode, it applies whatever torque is necessary to follow your "torque" profile. This is not what you want.

In "torque" mode, it applies the amount of torque that you specify, and the speed is whatever it wants to be.

An engine being cranked presents a certain amount of resistance (torque) and the speed is whatever it happens to be.

 

[istvanb]

Thanks Brian.

My assumption is that
1. if I record the speed of the flywheel on a real engine while the same starter is cranking it and then force the flywheel of my "simulator bench" to follow that speed profile, then the torque experienced by the UUT will be the same as what it would experience on the engine
2. similarly if I would record the torque applied by the engine and reproduce it on my simulator bench then the UUT speed will match to what I could record on the engine

Is this correct? Recording the speed instead of the torque is a LOT more easier.

 

[crshears]

Istvanb, why is the dot in your link between the 'u' and the 'b'? I for one won't click on it...

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[istvanb]

Thats is the link given by youtube once I click to the share button. (it shortens the link)

the original link to the video is:

https://www.youtube.com/watch?v=zbXRRQmf0jI&feature=youtu.be

Thanks!

 

[BrianPetersen]

Your assumption 1 is incorrect. There is a reason the speed profile is following the profile that it follows ... because the torque required to rotate the engine is varying (a lot) as each cylinder goes through compression strokes and as each valve spring has to get pushed open (and as the spring pushes back on the other side).

If you measure the speed profile and then emulate that speed profile on a bare flywheel sitting on bearings ... you get a flywheel going up and down in speed following that speed profile, without the underlying REASON (torque) that it is following that speed profile.

Sorry, but you'll have to measure both the speed and the torque ... Shortcuts won't work ...

 

[BrianPetersen]

Your eventual aim is to have the flywheel connected to an encoder which gives the flywheel position to a controller of some sort, and that controller then commands the servo in torque mode to absorb a torque that varies depending on the flywheel position ...

 

[crshears]

Let's begin at the beginning...

istvanb, your OP bears the title "Diesel engine simulator for starter motors." I'm taking this to mean the UUT is some sort of electric motor of unspecified type, intended to be used to crank a Diesel engine with its fuel rack set at zero, meaning no fuel being injected. If this is the case, you will be looking to supply power with the appropriate characteristics TO the UUT, since it is a motor, not a generator.

[...or do you wish to simulate a successful start of the Diesel engine so that after a short interval the speed of the UUT increases concomitant with negligible torque production?]

You write of your "prime mover (the big electrical motor I will control)" as a motor - yet the context seems to suggest that you wish to use this device as a dynamometer, viz., as a device that will apply a load to the UUT. If this is in fact the case, you desire to harness your big electrical "motor" as a generator, using the UUT as ITS "prime mover," correct?

And if THAT is the case, you are looking to take the output of this generator and either dissipate it in a resistive or other load bank, or regeneratively return this energy to your factory's mains, correct? [If memory serves, in the old days this was referred to as a "back loading" test.]

At the risk of being a snot, I submit that only once the terms of reference are understood can answering a question usefully becomes possible.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[crshears]

More...

Just read this thread:

http://www.eng-tips.com/viewthread.cfm?qid=173814

BrianPetersen, would it be doable for istvanb to employ a similar technique by programming the "regenerative" VFD to apply a pulse load to the UUT so as to mimic a Diesel engine being cranked?

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[istvanb]

@crshears: let me try clarifying the terms
UUT: unit under test in this case an automotive starter motor
Prime mover: a huge DC motor which I'd like to control in a way that it reproduces the load what the engine represents on the starter without having the engine itself.
Cranking: if you check the 23rd second of the video thats pretty much what I'd like to do. The blue line is the speed. It remains in the lower regions at the beginning moving the cylinders up and down. Finally the engine cranks and the speed ramps up. Related to this you see the current (green). There is an initial rush when it first tries moving the flywheel and then you see it on the lower region moving the cylinders. Once the engine cranks the UUT is deenergized and its current falls to zero amps.

The aim of the test bench is to "load" the UUT as it would be loaded on a real engine, while I dont have the engine in house.

Does this make sense? Let me know if not and I will try elaborate more!

@Brian: Previously I have worked with a $300k bench which requires only the speed as an input to create a simulated profile. (ok thats not true, it requires to know when do we start energizing the UUT, but thats irrelevant for now). It doesnt require torque. This is why I assume controlling the drive in speed mode is the way.

 

[crshears]

Hi istvanb,

One of the things that confused me was that what you are describing as a "prime mover" isn't actually one at all, since it isn't the thing doing the moving; it is in actual fact being moved by the unit under test, since it is substituting for the Diesel engine that isn't there.

The terminology I'm used to is that "moving the cylinders up and down" before fuel ignition takes place IS cranking...and taking that a step further, when you say, "finally the engine cranks and the speed ramps up," you mean to say, "finally the engine FIRES and the speed ramps up." Correct?

UUT I got right away...

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[istvanb]

@crshears: I am sorry mate, this just reflects that I am not a native speaker. Forget about "prime", lets call it simply a mover. I call it mover, because in my application it forces to move the UUT at a certain speed. And yes when I said "the engine cranks" I meant "the engine fires"... the language barriers

 

[crshears]

No worries about the language barriers; those exist so we can tear them down. Besides, engineering is the only universal language, right?

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[BrianPetersen]

Picture something looking like the flywheel of the engine except made of foam (weighs nothing) mounted on ball bearings (next to no friction). Picture that you have a motion profile (but no torque) for it. How much force is required to follow that motion profile? Next to nothing.

Even if you had the inertia right (by making the flywheel very heavy), it still won't simulate the sequential compression strokes of the engine.

You need the servo motor to operate in regenerative mode to absorb a torque-versus-displacement profile that corresponds to the normal starting of the engine.

 

[istvanb]

@Brian: yes its everything made out of foam then we have a bad scenario. But you wrote yourself that if I run the drive in speed mode then the right torque will be applied automatically. So - and I am only throwing numbers here - if the drive forces 100rpm then the right torque will be applied to keep that 100rpm no matter if it is 10Nm or 100Nm or 1000Nm for the sake of keep the UUT at 100rpm

please dont hesitate to educate me!!!!

 

[BrianPetersen]

NO, that's not what I said.

You have to run the drive in regenerative TORQUE mode. It has to apply a torque that follows the TORQUE profile as the flywheel rotates. (For example, at 0 degrees relative rotation angle apply 30 N.m back torque, at 10 degrees apply 55 N. back torque, at 20 degrees apply 10 N.m forward torque, or whatever the profile ends up being, these are only examples.)

The speed of the motor doing this will be whatever the starter motor drives it at. In TORQUE mode.

There will be an inertial component in addition to this, that is most easily matched by making your test rig have the same rotational inertia as the engine.

 

[Compositepro]

Istvanb, you want to apply a load to various starter motors that behaves like a diesel engine being started. First, this cannot be called a mover because it retards movement.

It is the starter motor that controls the speeds at any given given moment. The diesel engine that you wish to simulate simply applies a resistive torque that varies with time (engine speed and crank angle). You can measure this on a real engine while cranking at a spectrum of speeds.

Various servo motor technologies can be used to do what you ask. A torque sensor and speed sensor will be required, but you will be controlling the torque with your program, while the unit under test will determine the speed. The characteristics of the power supply to the uut will also affect the results.

 

[istvanb]

I agree that the term mover may not be correct. But otherwise I dont think my UUT would control the speed. If it would be controlled by the UUT then the speed would be something like 4000rpm. I think the key is that now matter the speed what my UUT wants rotate at, the "mover" (or "retarder" as you said) limits its speed to what we experience on a real application.

You right that the real engine applies a resistive torque. This torque forces the UUT to rotate at a lot smaller speed as it could rotate without that torque. I would like to reporduce the engine behaviour so I assume forcing the UUT speed to be the same I could measure on a live engine means that I should control the "mover" in speed mode.

Most of you recommended torque mode... this project could be above my skills right now...

 

[crshears]

Oh my...

"Control" and "determine" are becoming very loaded words in this thread!

My understanding is that most starting motors are DC series motors, and that DC series motors are defined as a constant-power motor, power in turn being defined as the product of rotational speed and torque, and that in a series motor, as load is applied, speed falls, with torque increasing so as to maintain a constant power output.

In a laboratory or factory test hall, the DC supply to the starting motor will likely be from a stabilized rectified AC source, meaning that the power available from the source remains more or less fixed for an indefinite period, whereas in real life the cranking will be from a battery, the output capability of which will rapidly decrease the longer cranking is continued.

The viscosity of the oil in the real-life engine will also have a profound effect on how rapidly the engine can be cranked, with a subsequent profound effect on how soon the engine will fire and become self-sustaining.

With these and a few other factors in mind, perhaps the best that can be said is that the speed at which a Diesel starting motor and its engine will rotate will be product of a resultant balance between the power being put in and the resistance being encountered...and as BrianPetersen stated, the amount of torque required to rotate any given engine through two complete revolutions [or just one, in the case of two-stroke cycle engines?] will vary continuously but quite predictably throughout the course of those revolutions, based on the work being either performed via the crankshaft as the compression stroke progresses or to the crankshaft by the piston as the trapped air re-expands.

Based on the foregoing, it could be necessary to program the dynamometer/generator regenerative circuitry to emulate more than one type of engine, especially if this test bed is to be used to develop starting motors for more than just one engine design...otherwise why go to the trouble for a one-off? In that case it would be better to just hire it done by somebody else and spare yourself all the headaches.

Just my thoughts.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Compositepro]

What controls the starter speed in real life? The diesel engine is a passive load with no speed control, and the starter creates motion and controls the speed by how its speed/torque curve interacts with the diesel's. This is what you are trying to simulate.