Periodic Reporting for period 1 - PET (Printing Electro-Tomography)
Periodo di rendicontazione: 2022-04-01 al 2023-09-30
3D printing is a revolutionary technique able to create parts and components that are not possible with conventional techniques such as machining and that are not suitable for large scale molding due to the time and investment required. 3D printing can be performed in many ways, however one of the most common processes is forming a part by stacking 2D layers to create the desired geometry. In the Fused Filament Fabrication (FFF) method, which was the focus of this work, this typically consists of a polymer material being extruded melting the layers of the to be produced together.
Typically this FFF method faces a trade-off in terms of the speed at which is being printed and the precision at which the deposition process occurs, due to the open loop control of the extrusion process. Typically only filament feed rate and hot-end/nozzle temperature are being regulated, with user calibration being the determining factor of the rate of over- or under-extrusion. The movement of the material depositing print-head is also often open loop, and filament consistency, environmental conditions and component wear state are often unknowns affecting the extrusion consistency. Therefore the ability to have a manner of real time monitoring of the process would greatly improve the consistency and quality of FFF 3D printed parts and components.
Our work introduces conductively doped filaments to provide a real-time way to perform direct measurement of the fusion and extrusion process of deposited layers. This is achieved by measuring the electrical impedance between the printbed and extruding nozzle, deriving a tomographic description of the print as a resistive network resulting from the printed layers. The addition of a predictive model for the expected print impedance then allows for identification of part construction issues, which can then be adapted to result in real time defect characterisation with the options of correction or print abortion in the case of failed prints.
Typically this FFF method faces a trade-off in terms of the speed at which is being printed and the precision at which the deposition process occurs, due to the open loop control of the extrusion process. Typically only filament feed rate and hot-end/nozzle temperature are being regulated, with user calibration being the determining factor of the rate of over- or under-extrusion. The movement of the material depositing print-head is also often open loop, and filament consistency, environmental conditions and component wear state are often unknowns affecting the extrusion consistency. Therefore the ability to have a manner of real time monitoring of the process would greatly improve the consistency and quality of FFF 3D printed parts and components.
Our work introduces conductively doped filaments to provide a real-time way to perform direct measurement of the fusion and extrusion process of deposited layers. This is achieved by measuring the electrical impedance between the printbed and extruding nozzle, deriving a tomographic description of the print as a resistive network resulting from the printed layers. The addition of a predictive model for the expected print impedance then allows for identification of part construction issues, which can then be adapted to result in real time defect characterisation with the options of correction or print abortion in the case of failed prints.
The initial stage of the project involved the creation of a measurement platform capable of performing printing electro-tomography impedance measurements. Starting with the evaluation of bed electrode contacts coatings and patterns for optimal contact with the conductive filament and the evaluation of 2-, 3- and 4-wire measurement modalities allowing for removal of the bed to print and nozzle to print contact impedance. To speed up the development a Ender S1 Pro 3D printing platform with support from the open source community was selected, and modified with customized Marlin firmware to adapt a electrode equipped print bed. The printing platform was subsequently equipped with rotary encoders on each of its print axis and extrusion system, to allow for motion tracking of the print head. Data acquisition was handled with the use of TiePie HS5 digital oscilloscopes capable of synchronous encoder and impedance measurements.
To evaluate the method measurements of the commonly used 'Benchy' and 'XYZ test cube' were performed, resulting in the finding that a 2-wire measurement between four corner and a nozzle electrode resulted in sufficient measurement resolution. To allow for additional validation of the geometry of the printed parts post printing, a Einscan SP 3D scanner was acquired to aid next to the initial standard of visual inspection.
Further measurements were performed with a developed single 0.4mm walled hollow 20x20x20mm test cube. This thin wall structure minimized thermal mass and conduction effects resulting in dominant convection cooling to the print bed temperature. These were printed in a solid form, a perforated form where the perforations were smaller than the extrusion nozzle (resulting in no breakage of contact) and in a form with larger perforations (resulting in contact breakage). Measurements with varying electrode patterns were evaluated to determined the influence of geometrical effects in the test structure as well as measurements with varying temperatures to evaluate the effect of temperature on material resistance and the effects of print warping. Measurements with a 3- and 4-wire modality were also performed to evaluate the contributions of bed contact and nozzle contact resistances.
The collected data was visualized with the use of a developed tool allowing for inspection of independent print layers (especially the first layer), print paths and print resistance as a function of location in the printed structure in a three dimensional rotatable figure. Development of a predictive print model was also achieved for the thin walled test geometries allowing for the detection of print defects as deviations from this predictive model. Future work will expand upon this model to allow for real time prediction of print impedance for complex geometries.
Throughout the process an industrial partners actively gave input on the design choices. The work has been commercialized through sale the patent to our industrial partner with an anti-shelving clause, requiring to actively develop the technology and to bring it to the market within a fixed period. Further work will also be performed in close collaboration with our industrial partner to aid commercialization of the methodology.
To evaluate the method measurements of the commonly used 'Benchy' and 'XYZ test cube' were performed, resulting in the finding that a 2-wire measurement between four corner and a nozzle electrode resulted in sufficient measurement resolution. To allow for additional validation of the geometry of the printed parts post printing, a Einscan SP 3D scanner was acquired to aid next to the initial standard of visual inspection.
Further measurements were performed with a developed single 0.4mm walled hollow 20x20x20mm test cube. This thin wall structure minimized thermal mass and conduction effects resulting in dominant convection cooling to the print bed temperature. These were printed in a solid form, a perforated form where the perforations were smaller than the extrusion nozzle (resulting in no breakage of contact) and in a form with larger perforations (resulting in contact breakage). Measurements with varying electrode patterns were evaluated to determined the influence of geometrical effects in the test structure as well as measurements with varying temperatures to evaluate the effect of temperature on material resistance and the effects of print warping. Measurements with a 3- and 4-wire modality were also performed to evaluate the contributions of bed contact and nozzle contact resistances.
The collected data was visualized with the use of a developed tool allowing for inspection of independent print layers (especially the first layer), print paths and print resistance as a function of location in the printed structure in a three dimensional rotatable figure. Development of a predictive print model was also achieved for the thin walled test geometries allowing for the detection of print defects as deviations from this predictive model. Future work will expand upon this model to allow for real time prediction of print impedance for complex geometries.
Throughout the process an industrial partners actively gave input on the design choices. The work has been commercialized through sale the patent to our industrial partner with an anti-shelving clause, requiring to actively develop the technology and to bring it to the market within a fixed period. Further work will also be performed in close collaboration with our industrial partner to aid commercialization of the methodology.
The achievement of a 3D representation of the print fusion in a real time manner through the performance of DC resistance measurement is a great step forward for the quality assessment. Current real time optical and thermal observations can still result in long post processing times where this method results in quick directly observable resistance jumps in the case of voids and print-line defects. It also offers the benefit of providing a measurement modality for internal defects that might not be directly observable as they are being printed, forming due to thermal stress in the part in subsequent layers. Thermal stress events in previous layers such as print warping, layer splitting and loss of print bed adhesion are also easily observed through the impedance tomography method.
In addition the ability of observing these internal defects, the method can also be applied to determine critical first layer adhesion, which if insufficient leads to the majority of print failures. The monitoring of the first layer leveling and corresponding adhesion can be achieved by measuring the variation in individual print-lines by placement of parallel electrodes, monitoring of the heating and extrusion of the base layer of the print. Currently both a 3D model for simple geometries and a 2D model for complex first layers have been developed, along with a characterisation of the thermal resistivity of the materials and the effect of geometry upon the cooling rate of samples.
The achievement of a direct real time monitoring modality for Fused Filament Fabrication based 3D printing has a great impact on the industrial applicability of this fabrication method. Current polymer printing processes rely on expensive post-printing part validation through X-ray or CT-scan inspection for critical parts. The implementation of our methodology by our industrial partner on their high-end industrial machines would have a large impact on improving the affordability of the production of high quality polymer parts in small batches, leveling the playing field for small start-ups and businesses.
In addition the ability of observing these internal defects, the method can also be applied to determine critical first layer adhesion, which if insufficient leads to the majority of print failures. The monitoring of the first layer leveling and corresponding adhesion can be achieved by measuring the variation in individual print-lines by placement of parallel electrodes, monitoring of the heating and extrusion of the base layer of the print. Currently both a 3D model for simple geometries and a 2D model for complex first layers have been developed, along with a characterisation of the thermal resistivity of the materials and the effect of geometry upon the cooling rate of samples.
The achievement of a direct real time monitoring modality for Fused Filament Fabrication based 3D printing has a great impact on the industrial applicability of this fabrication method. Current polymer printing processes rely on expensive post-printing part validation through X-ray or CT-scan inspection for critical parts. The implementation of our methodology by our industrial partner on their high-end industrial machines would have a large impact on improving the affordability of the production of high quality polymer parts in small batches, leveling the playing field for small start-ups and businesses.