Can a Model Accurately Simulate Pressure Variations in Tillage Tools?

[serbring]

Hi all,

I have just measured the pressure on a point of a tillage tool and the resulting signals is reported in the following link.

In literature, tillage tool are modeled with a constant pressure, but i reality the pressure changes over time and I would like to model the pressure variation over time. How could I do it? Could you suggest me anything to study which will permit me to accomplish that?

Thanks

 

[Nidum]

Structural mechanics and Finite Element analysis .

 

[serbring]

Structural mechanics and Finite Element analysis .

Not really, I measured the pressure because to design the frame of tillage tools, a FEM analysis is carried out. This requires the assumption of a pressure distribution over the plough and for fatigue analysis, its variation over time/distance should be assumed/hypotized. That is the reason, I measured the pressure. That model, can be something like arma (?) where it is based on stochastic parameters. The next step of the activity will be trying to verify, what affects the parameters, probably, soil conditions, moisture content, and so on.

 

[berkeman]

reality the pressure changes over time

Why is the pressure spiking so much and with such a large variation? Are you hitting rocks in the soil or something? Can you post a picture of your test setup and the ground you are tilling? How fast is the tow vehicle moving?

https://upload.wikimedia.org/wikipedia/commons/thumb/a/ab/Ploughtilling_the_field.jpg/275px-Ploughtilling_the_field.jpg

 

[serbring]

We mounted sensor pressure on all the plough components on untilled soil. We made this measurement in two different years, with different setup (i.e. sensor protection) and on different soils, different speed (around 2, 4 and 6km/h) and the signals are always so spiky.
I found a paper back to the '70s and similar magnitude and spiky behaviour was found. I believe this is due to the granular behaviour of soil.

 

[berkeman]

The baseline of 0 pressure for the graph seems incorrect to me. When the tiller is resting on the soil or tilling the soil, the baseline must be positive because of the weight of the arms and rotors. If the baseline is really 500 or more, then a delta of 20-50 may be reasonable. Is this an AC-only measurement? What are the approximate weight and surface areas involved?

 

[serbring]

It's not a tiller but a plow (or plough). Why the baseline shouldn't be zero? No soil over the plough surface, no pressure. The measurements were DC, we used tactile sensors. The plough body is about 40kg.

 

[Nidum]

Structural mechanics and Finite Element analysis

I had in mind trying to model the action of the plough as it was dragged through the ground .

 

[serbring]

I had in mind trying to actually model the action of the plough as it was dragged through the ground .

Others tried this approach, but the results I saw in the paper are totally different from the one I measured.

 

[berkeman]

Did you try any tests to get a better idea of what the pressure would look like with very uniform soil? Like try it in fine sand, or in very high quality soil? What did you see as differences between damp soil and dry soil?

 

[serbring]

That is the point... on monday, I am going to perform test on tilled soil with power harrow, that should be very fine. I would love to make this measurement on fine sand, not really easy for us, because we do not have a soil bin, so we can only carry out the test on real field conditions.

 

[berkeman]

What software do you have for doing the final modeling? I would think you could model the pressures based on your measurements (and add in random fluctuations of about the amplitudes you are seeing in your experiments). That would get you closer to being able to do the fatigue analysis you want to do...

 

[serbring]

Actually, I do not want to model the plough but only the pressure signals. This will serve as an input to the FEM analysis. Maybe we will use Ansys, but it is not sure, FEM modeling is not my field, a coworker will do it for me.

 

[Nidum]

(1) I would like to properly understand the mechanisms generating those pressure signals and factors affecting the accuracy of those signals .

It is correct that these pressure signals are effectively being used to measure the total force acting on the plough blade ?

(2) The plough blade is being pulled through soil with approximately uniform physical properties but with a random arrangement of stones , roots and voids mixed in .

The depth of engagement of the plough in the ground varies continuously but not by a very large amount .

The plough is moving at constant horizontal velocity .

For any particular plough , tractor and ground quality combination this should give a pulling force v time plot which has a basic force level undulating relatively slowly above and below a mean value but with additional short duration spikes and troughs .

(3) At any instant the pulling force equals the soil reaction force on the blade^(a) .

The soil reaction force though is made up of many small force contributions acting on an equal number of small areas on the blade . These small force contributions will vary in magnitude to some extent according to location on blade^(b) and also to some extent they will track the variations of total force acting on the complete blade .

There is however likely to be a large variation in magnitude of these force contributions due to local variations in the soil mixture .

(4) Determining the total force acting on the plough blade at any instant just by measuring the force contribution acting on a small sampling area is likely to give results which are very noisy and with possible large magnitude errors .

(5) There is likely to be a high level of vibration and acoustic noise generated by the movement of the plough blade through the soil . If picked up then this could cause many false spikes and troughs to appear in the force results .(a) : Neglecting pro tem any dynamic response effects .
(b) : Including effects of local variations in blade geometry .

PS: I do not see FEA modelling of the plough / soil interaction as being particularly difficult or likely to produce grossly inaccurate results .

 

[Tom.G]

Actually, I do not want to model the plough but only the pressure signals. This will serve as an input to the FEM analysis.

Digitize the measurements you have taken and use them as a table-lookup in the simulation?

 

[serbring]

(1) I would like to properly understand the mechanisms generating those pressure signals and factors affecting the accuracy of those signals .

It is correct that these pressure signals are effectively being used to measure the total force acting on the plough blade ?

We measured the total draft force acting on the plough. The signal is very different as it is always above zero (no surprise). I suppose the draft force is mostly caused by the cutting process rather than the moving of the ploughshare in the soil.

(2) The plough blade is being pulled through soil with approximately uniform physical properties but with a random arrangement of stones , roots and voids mixed in .
The depth of engagement of the plough in the ground varies continuously but not by a very large amount .

The plough is moving at constant horizontal velocity .

Test were carried out at full throttle and at constant ploughing depth. Therefore, the ploughing speed may vary a little as soil shear strength is not constant.

For any particular plough , tractor and ground quality combination this should give a pulling force v time plot which has a basic force level undulating relatively slowly above and below a mean value but with additional short duration spikes and troughs .

(3) At any instant the pulling force equals the soil reaction force on the blade^(a) .

The soil reaction force though is made up of many small force contributions acting on an equal number of small areas on the blade . These small force contributions will vary in magnitude to some extent according to location on blade^(b) and also to some extent they will track the variations of total force acting on the complete blade .

There is however likely to be a large variation in magnitude of these force contributions due to local variations in the soil mixture .

(4) Determining the total force acting on the plough blade at any instant just by measuring the force contribution acting on a small sampling area is likely to give results which are very noisy and with possible large magnitude errors .

Our idea is not to evaluate the total force from the measured pressure, but to model the pressure which can be used to validate FEM/DEM model,'

(5) There is likely to be a high level of vibration and acoustic noise generated by the movement of the plough blade through the soil . If picked up then this could cause many false spikes and troughs to appear in the force results .(a) : Neglecting pro tem any dynamic response effects .
(b) : Including effects of local variations in blade geometry .

I have not thought about it, your are on right. It would be interesting to evaluate this factor.

PS: I do not see FEA modelling of the plough / soil interaction as being particularly difficult or likely to produce grossly inaccurate results .

you're on right, but FEM/DEM/CFD do not really reproduce the process even if they can reproduce the results by means parameters tuning.

 

[serbring]

Digitize the measurements you have taken and use them as a table-lookup in the simulation?

That is possible, but a model would permit to simulate the plough in different soil conditions.