<< Go to homepage

Introduction

Team from my university - PUT Chemistry - 3D Print section -started to work on RecycleBot. We already 3D printed every necessary part, but we found out some problem I gonna describe below:

Auger

We bought 16 mm auger instead of 5/8 " which is pretty silly mistake, but it may be common while building RecycleBot outside of US, because for example in Poland it is pretty hard to find inch augers. I have found an universal solution to this problem, which for now seems pretty easy (I hope it gonna be like it seems), but I have created an FreeCAD model using Assembly4 workbench (I am thinking to use spreadsheets instead, as they are available in vanilla FreeCAD), which let us use an variable to describe auger diameter, and fitting (I have used same backlash, as original authors, but in my project it is possible to whatever one). Moreover for now it just change the bore diameter, and flange remains the same, which might be changed if sb wants to use even bigger auger.

It is possible to find the FreeCAD file in my cloud, but there is upgrade which reduces wobbling of the hopper due to axial forces from auger.

RPM values

Auger (again…)

Second issue we found is diffrent gearbox on our NEMA23. In orginal work it is 15:1, and we have 18:1. In the code I have found and augerDesiredRPM variable which is originally set to 35 RPM, and it is a little bit confusing for me, because it seems like it is stepper motor speed, and the variable name is just set as augerDesired. I think like that, because augerDesiredRPM value is multiplied by 200, which is equal to NEMA23’s steps per revolution.

It means:

$$15:1 → 35 RPM$$ $$18:1 → x RPM$$

$$x = \frac{35 ⋅ 18}{15} \rightarrow x=42$$

So in my humble opinion we would need to change augerDesiredRPM variable to 42 to reach the same results as in original project. It could be nice to write simple script in Python to make it more universal.

Puller DC motors

That is another problem, as we have 70 RPM DC motors, and in orginal project there are 50 RPM DC motors, but I gonna update the page in few days about that problem.

FEM analysis

For one of mine term project I need to finish the FEM analysis. I would like to check the differences between our Auger assembly - our one is longer, which can lead to better extrusion parameters.

I’d like to modify:

It appeared to be pretty hard project, I have spent on it really too much time (about 100hrs, literally 7 days 10-12 hrs each), mostly fighting with FreeCAD mesh (later on I’ve found out that it was bug from 2017 connected with 2 diffrent materials calculations).

I have finished to use 2 programmes(later I gonna use OpenFOAM, rn I havent got enough time :( )):

I have to use Fusion360 to “model” filament in the extruder, due to CalculiX bug while running thermomechanical simulation with 2 diffrent materials.

Whole project, with many grammatical mistakes, ugly and improper TechDraw, and many others not is available here (still better than rest of this post, eventhough text is almost the same, as I am not able to live without Emacs, so I just copy-pasted content from there to Overleaf).

I promise to repair both post and report after exam time, I am posting it as it is, so my teammates can view my results (and laugh at my grammar as well, my Emacs reversed engineered grammarly is not working). Moreover I know, I should use list to import so many .csv, but working on it few hours before deadline didnt help on thinking at all.

Other files connected with the project are available here

Methodology

As I have said I needed to use 2 programmes. Firstly temperature inside the barrel was calculated using FreeCAD and CalculiX. Later on Fusion360 Plastic mold was used to simulate the PLA flow. The inject point was placed on the place of extrusion point in RecycleBot’s extruder (2mm diameter). In this case the actual flow become inversed to obtain results.

For first step FreeCAD and CalculiX were used, due to its simulation over time abilities: 6 version were used to receive the temperature after 30s of heating, and choose nearest to 220 degrees, as it is PLA extruding temperature in RecycleBot.

In Fusion360 the results were completly useless - cap outside temperature was 15 degrees, and in the middle barrel had 1500 degrees.

Heater coil calculations

Electrical calculation

To obtain heat source power, Ohm law and electrical power formula were used:

Ohm law: $$R = \frac{U}{I}$$

where: R - resistance [Ohms]
U - voltage [V]
I - current [A] \

Electrical power formula:

$$P = {U}\cdot{I}$$

where: P - power [W]

Both of these combined results in heater power formula which is:

$$P =\frac{U^2}{R} \cdot \eta$$

where:

$$\eta$$ - efficiency

Heater efficiency due to Bassily and Clover raport is fluctuating between 80-90 %.

Heater coil used in PUT Chemistry version had resistance equal to 4.77 Ohm/m

Geometrical calculations

In next step wire lenght was calculated using the formula:

$$WL = \frac{L}{P} \cdot \pi \cdot D$$

where: L - wire Lenght [mm] P - wire Pitch [mm] D - outside diameter [mm]

In PUT Chemistry machine wire resistance is 5 Ohm/m, pitch is same as in original RecycleBot, so 1 mm, and wall thickness of the pipe is 2mm, so outside diameter is equal to 20 mm.

FEM Fusion360

In Fusion the Plastic flow was simulated to observe the plastic flow through the barrel and around the auger. The tool was used in wrong way, but still gives information about flow mechanis. Nevetrheless OpenFOAM analysis would be better, and should be carry on in the future. Fusion appeared to be useless due to lack of time properity.

FEM FreeCAD

Unfortunatly simulation of whole extruder with the screw was not possible due to CalculiX bug connected with simulating 2 diffrent materials. The error obtained was:

Nevertheless thermomechanical analysis of the extruder barrel was done in FreeCAD, 30 second with 5 second step.

There were 2 boundary condition:

On the end max temperature of 551 K was obtained, with 312 K minimal.

FEM mesh with boundary parameters can be seen on photo.

FEM plotting

Two scripts were used to obtain the temperature over time charts, and polynominal regression of the time required to set the temperature on desired setting. Unfortunately the Paraview bug interrupted the workflow, which made any changes of the charts impossible. According to this situation these scripts were written in Python.

First of them was used to plot relationship between temperate and horizontal position. The PlotOverLine function in Paraview was used, nextly the results were exported to .csv file. Probing line can be seen on photo, the position was chosen to be 9 mm away from central barrel axis, due to stable temperature requirement.

Temperature and horizontal positon shows big thermal gradient between cap and the heated zone. Temperature amplitude increases over time, it is necesarry to cool down the extrusion product, though model shows temperature on the cap is below PLA melting temperature, while temperature in heating zone exceed nominal working temperature, admittedly this phenomenon going to be reduced over the time, due to augmented heat flow between heater zone and cap, comparing to convectional heat exchange, whose effect is shown on 0 to 25 mm on the chart, where the curves over the time begin to be more parabolic than 5 second curve, which is almost flat.

Next required parameter was the time needed to reach desired temperature. For this task NumPy package was used, with its polyfit method which allowed to obtain polynomial regression analysis:

$$-4.76256e-06x^6+ 0.0079313x^5+ -5.28158x^4+ 1757.894661x^3+ -292432.891258x^2+ 19451166.732051x$$

<< Go to homepage