Final Report Summary - ULTRAVISC (Sensor-based ultrasonic viscosity control for the extrusion of recycled plastics)
Executive summary:
There is a need is to develop a cost-effective, potentially retrofit system to address the current deficiencies in reprocessing waste plastics with respect to plastic flow and the issues of contamination and variability in the feedstock. Motivation for progressing beyond the state-of-the-art is to encourage the large number of European Union (EU) small and medium-sized enterprises (SMEs) to continue to use and increase their ability to re-use waste plastics. The development of the ULTRAVISC technology has enabled waste plastic processors to continuously monitor and activate changes that will reduce the production of scrap / waste whilst giving the potential opportunity to produce thinner walled products through the application of ultrasound via a closed-loop feedback control system.
The ultrasonic modulation system comprises a newly developed sonotrode and ultrasonic chamber. The system is designed to operate at 20 kHz and able to deliver a range of ultrasonic energy levels based on the input power. A number of process trials using a variety of dies configurations were carried out using virgin and recycled feedstock and demonstrated the potential of ultrasonics to reduce both melt viscosity and increase material throughput.
The new closed-loop ULTRAVISC control system demonstrated that it is able to monitor process conditions and apply ultrasonics modulation as and when is required to maintain a constant viscosity. Furthermore, the system is also able to control the ultrasonics, heaters or a combination of the two in order to target a specific throughput or die pressure. A variety of recycled feedstock with different melt flow index (MFI) has been processed and it has been shown that the system is able to cope with the both sequential changes to MFI and material blends.
As in the case of the ultrasonic modulation system, the ultrasonic filter was newly developed. It is based on a reversed candle filter whereby the polymer melt runs along the length of the single-piece sonotrode and then through a candle filter. A variety of process trials were carried out to demonstrate the positive effect of ultrasonics on the unclogging of filters. The trials showed that the ultrasonic was unable to unclog the filter but did cause damage to it on several occasions. On a more positive note, the ultrasonic filter system has shown it is able to operate at pressures over 350 bar.
The project has successfully met its principle objectives; a prototype ultrasonic system with closed-loop control has been developed and successfully shown that it is able to monitor and control the viscosity of recycled feedstock in real-time. Furthermore, trials have also shown that a greater throughput is possible using ultrasonics despite the power requirements being less than a standard extruder system.
Project context and objectives:
The ULTRAVISC project which was funded under the European Commission (EC)'s Seventh Framework Programme (FP7), was carried out over a period of 24 months beginning on 1 October 2009 and ending on 30 September 2011.
ULTRAVISC addresses the EU Directive 1999/31/EC by introducing a novel processing technology that enhances the extrusion of batch-variable plastic waste and provides the means for the EU to improve recycle rates and extend recycling targets.
Plastic recycling and reducing the volume of plastic waste going to landfill is an emotive subject, highly publicised by our increasing dependence on oil, fears over future supply-demand and concern for our environment. Social responsibility and political debate have introduced increasingly stringent EU mandates aimed at encouraging the plastics industry to invest in processes and technologies that minimise the disposal of plastic waste at landfill. To enable recycled plastics to compete with virgin polymers technological developments that provide the mechanisms with which recycling and reprocessing of plastic waste becomes more economically viable and a commercially attractive alternative is needed. The main issue for extruders of recycled plastics concerns the variability of the recyclate feed material which currently requires a novel solution. We have seen an opportunity to provide a viable solution to meet this currently unmet market need.
The main technical objectives that have been achieved through process innovation are:
- Development of a variable intensity ultrasound device for the normalisation of polymer MFI to within a target MFI range of ± 5 % at the die, such that processing remains consistent and / or controlled under a specific set of extrusion conditions, regardless of variations in the bulk MFI of the recycled plastic.
- Operation of ultrasonic probes in extrusion machines with a range of process throughputs, operating temperature of up to 200 degrees of Celsius, and pressures of up to 20 MPa (200 bar), using an ultrasound frequency of 20 kHz.
The technical development work was divided into a number of work packages (WPs) each contributing to the overall objective of the project:
- WP1: Material characterisation and ultrasound
- WP2: Soft sensor
- WP3: Viscosity control
- WP4: Low / high power sonic filter
- WP5: ULTRAVISC system integration.
There is a need is to develop a cost-effective, potentially retrofit system to address the current deficiencies in reprocessing waste plastics with respect to plastic flow and the issues of contamination and variability in the feedstock. Motivation for progressing beyond the state of the art is to encourage the large number of EU SMEs to continue to use and increase their ability to re-use waste plastics. The development of the ULTRAVISC technology has enabled waste plastic processors to continuously monitor and activate changes that will reduce the production of scrap / waste whilst giving the potential opportunity to produce thinner walled products through the application of ultrasound via a closed-loop feedback control system.
The original objective was to implement a Soft-sensor to accurately monitor and control the melt viscosity of the waste plastic melt from knowledge of the processing parameters and conditions (i.e. melt temperature, screw speed, back pressure, motor torque). Control is achieved using ultrasonics to modulate melt viscosity and filter the material. However, process trials using the Soft-sensor highlighted that a different model is required for each extrusion configuration; this is considered to be somewhat impractical and expensive. Therefore, it was agreed that an alternative control strategy be used. The alternative strategy is based on proportional-integral-derivative (PID) control whereby viscosity is kept constant by controlling the ratio of pressure to throughput of the polymer melt in the die.
The ultrasonic modulation system comprises a newly developed two-part sonotrode and associated ultrasonic chamber. The system is designed to operate at 20 kHz and able to deliver a range of ultrasonic energy levels based on the input power. A number of process trials using a variety of dies configurations were carried out using virgin and recycled feedstock and demonstrated the potential of ultrasonics to reduce both melt viscosity and increase throughput.
The new closed-loop ULTRAVISC control system demonstrated that it is able to monitor process conditions and apply ultrasonics modulation as and when is required to maintain a constant viscosity. Furthermore, the system is also able to control the ultrasonics, heaters or a combination of the two in order to target a specific throughput or die pressure. A variety of recycled feedstock with different MFI has been processed and it has been shown that the system is able to cope with the both sequential changes to MFI and material blends.
As in the case of the ultrasonic modulation system, the ultrasonic filter was newly developed. It is based on a reversed candle filter whereby the polymer melt runs along the length of the single-piece sonotrode and then through a candle filter. A variety of process trials were carried out to demonstrate the positive effect of ultrasonics on the unclogging of filters. The trials showed that the ultrasonic was unable to unclog the filter but did cause damage to it on several occasions. On a more positive note, the ultrasonic filter system has shown it is able to operate at pressures over 350 bar.
In summary, the project has successfully met its principle objectives; a prototype ultrasonic system with closed-loop control has been developed and successfully shown that it is able to monitor and control the viscosity of recycled feedstock in real-time.
The project consortium originally comprised 8 participants (5 SMEs and 3 research and technological development (RTD)) with the co-ordinator being Cherry Pipes. During the course of the project one of SMEs Plastitehase (Estonia) had to withdraw due to financial difficulties. Project results:
WP1
The first WP of the project focused on the development of software / hardware architecture, materials characterisation and ultrasonics knowledge enhancement.
The hardware architecture includes a data acquisition (DAQ) system that has the ability to measure various extrusion process parameters including barrel pressure, die pressure-drop, screw speed, and temperature along the extrusion line. Software was developed to investigate the closed-loop control system and the proposed soft-sensor technique which describes the relationships between the machine parameters and melt viscosity at the die.
The viscosity model was successfully tested with virgin low-density polyethylene (LDPE) using a Collins 25 mm single-screw extruder. However, polymer feeding could not be fully controlled when working with high-density polyethylene (HDPE) and so a larger extruder (38 mm) was used for all subsequent trials.
In parallel with the work on viscosity modelling, a closed-loop control system was developed for the integration of the soft-sensor and the control system. The output of the soft-sensor is used as feedback viscosity for the purpose of control. Based on the error between the set value for viscosity and the model output, the controller has been designed to control the screw speed and the temperature in order to keep the viscosity constant across a range of polymer types and compositions. Since the viscosity is proportional to pressure drop and the reciprocal of screw speed the target of the closed-loop control system is to maintain the ratio of the two constant. This was successfully been achieved using PID control and a pc as the user interface.
Recycled HDPE routinely used by Cherry Pipes underwent process trials and subsequent characterisation in order to help build a consistent process methodology in preparation for ultrasonic trials. This work included the design and manufacture of a slit die that was used by each of the RTD participants to ensure consistency.
A finite element analysis (FEA) model was developed to help understand how ultrasound energy is transmitted from the sonotrode through the polymer melt. The effect of probe geometry, depth within the melt, and amplitude on the propagation of ultrasonic energy was been investigated. One of the main modelling results indicated that a single sonotrode is sufficient to transmit to the ultrasonic energy through the chamber.
The first stage of the ultrasonic filter development involved the evaluation of several designs which included a vibrating sonotrode in the form of a mesh located within the extrusion line. Discussions within the consortium and with third parties led to this design being abandoned as it could not be achieved within the budget and timeframe of the project. Consequently, a new design was developed in which a standalone probe is used with a candle filter.
It is important that the filter is effective in-use and so work was carried out to estimate the cleaning interval of the filter. A 100 Mesh filter was considered appropriate and it was estimated that such a filter would require a cleaning interval somewhere between 1 and 10 tonnes of recycled material. Replacement of the filter was also considered at the design stage to minimise any downtime.
WP2
The implementation of soft sensor is based on a prediction-feedback observer mechanism. A 'prediction' model generates an open-loop estimate of the melt viscosity based on the process inputs; this estimate is used as an input to a second, 'feedback' model to predict the pressure of the system. These two models were trained off-line from the process data and rheological properties of material by using genetic algorithm. The output of pressure model is compared to the actual measured melt pressure and the error used to correct the viscosity estimate in real time.
A DAQ system was built to enable the extruder screw speed and temperature settings to be controlled from the PC, as well as collection of pressure and temperature data.
As mentioned in WP1, a slit die was used in the process trials using the soft sensor.
The extruder throughput was shown to have an approximately linear relationship with screw speed. Model identification trials were carried out whereby the extruder was run at a range of screw speeds and temperature settings with different feed materials to generate data for modelling. The pressure and viscosity models calculated from genetic algorithm which fit best with the experimental data were calculated.
It was agreed by the project partners to use two virgin HDPE materials with different MFIs (Marlex HHM-TR 144, MFI: 0.18 and Sabic B6246LS, MFI: 0.5) plus a number of blends in order to 'train' the model.
The models required by the soft sensor are specific to the machine / die set-up which is a disadvantage; the end-user would need to 'train' the model for each die configuration which could be both costly and time consuming.
The viscosity prediction model is material specific but as this model is adjusted by the pressure feedback model it is not essential that this is accurate, however a new model should be investigated if the type of polymer is changed (e.g. from HDPE to PP). The prediction model can be developed doing an offline test.
Due to the requirement for the soft sensor to be trained for each extruder configuration it was agreed that an alternative control strategy be implemented; a control system that can be more easily adapted to each industrial application.
As the viscosity is proportional to the ratio pressure to throughput and its measurement may not be readily available for different die geometries, the target of viscosity control can be achieved by keeping both throughput and pressure of the polymer melt in the die constant.
In order to do this, a strategy was selected to implement two controllers separately in computer software: the first controller is used to control the ultrasonic energy or temperature to maintain constant die pressure, the second controller controls screw speed for the constant throughput of the polymer melt. Both controllers are designed with specific algorithms to run simultaneously during the extrusion process. For the purpose of keeping the die pressure constant, controller 1 uses ultrasonic only, temperature only or the combination of ultrasonic and temperature as its control variables.
WP3
A new design of sonotrode has been developed which differs considerably from the more standard cylindrical design that is only partially submerged into the polymer melt. Background knowledge and technical discussions led to a novel sonotrode design that incorporates grooves and extends through the barrel chamber. This design increases the surface area of the sonotrode that is in contact with the polymer melt which should result in a more uniform reduction in viscosity.
Prior to the manufacture of the ultrasonic modulation hardware, further understanding of the pressures in the ultrasound chamber was required to optimise the gap between sonotrode and chamber. An ultrasonic pressure test chamber with a 'dummy' sonotrode was, therefore, designed and, trials ran. Extrusion trials showed that die pressure was effectively unaffected by the size of gap. Consequently, the smaller gap was selected for the 'real' sonotrode / chamber set-up.
A new two-part grooved sonotrode fabricated from titanium was commissioned. The sonotrode is clamped top and bottom and has maximum operational amplitude of approximately 40 µm at its midpoint, and a nominal operating frequency of 20 kHz. The system is such that the 4 kW generator provides sufficient power to oscillate the sonotrode at the pre-set amplitude. Development trials work focused on running the system between 50 and 90 % of the maximum amplitude.
Initial ultrasonic trials were carried out using virgin HDPE (Marlex TR 144, MFI: 0.18) to investigate the general effect of ultrasound on-time and amplitude on the polymer melt.
%The polymer melt exiting the die when the system is running at 50 % amplitude continuously for a period ('on time') of 30 seconds is shown. The ultrasonic effect on the polymer melt is clearly evident. The splitting of the polymer melt was shown to be a result of the applying the ultrasonics for too long; no splitting occurred when the 'on-time' was 10 or 20 seconds. It was also observed that there is a delay between when the ultrasonics is applied and when the effect is seen.
The sonotrode is designed to oscillate within a fairly narrow frequency range: 20 kHz ± 500 Hz and has nodal points (zero displacement) at the clamping points (winglets). Pulsed and continuous ultrasound has been applied and it was shown by all RTDs that extended continuous application resulted in a significant rise in the melt temperature. The best mode of operation was shown to be continuous pulsing (or pulse width modulation).
The chamber is designed such that the sonotrode can be mounted either vertically or horizontally. Due to the considerable weight of the ultrasonic unit, a support stand is advised. The cost of the prototype ultrasonic hardware including the sonotrode, booster, convertor and generator is ~20,000. A commercial system would be larger in size and, consequently, the cost would be higher. However, if the use of ultrasonics is able to halve Cherry Pipes waste it could save them in excess of EUR 80 000 per annum.
A tubular die was commissioned to evaluate the effect of the ultrasonic modulation on a representative tubular profile. In order to test the robustness of the ultrasonics a number of different types of recycled feedstock were processed both sequentially and pre-homogenised. The results demonstrate that ultrasonics can increase throughput as well as reduce system energy consumption. Ultrasonic modulation has been demonstrated to have real commercial potential for recyclers such as Cherry Pipes.
Notes:
(i) Material historically used by Cherry Pipes to produce their drainage pipes
(ii) Produced in house by Cherry Pipes using recycled HDPE milk bottles
(iii) Black rechip sourced locally by Cherry Pipes
WP4
The final design of the ultrasonic filter is based on a 'reversed' candle filter (polymer melt flows from the inside of the filter outwards). The design of the probe takes into account various parameters such as ultrasonic efficiency, mechanical resistance, optimisation of the filtration surface, connection with the surrounding elements and cost. After finalising the probe design, the next step was to design the housing taking into account operational requirements such as positioning and fixation of the pressure and temperature transducers and system integration.
The dimensioning of the most critical components of the unit has been performed using the finite element method. The selected parts have been studied under an internal pressure of 200 bars which represents the ultimate specification in industrial configuration.
Optimisation has been carried out on the elements that showed some weakness under this loading. The flow propagation has also been optimised using computational fluid dynamic (CFD) analyses with objective to maximise the filtration effectiveness and consequently the cleaning intervals.
Numerous trials were performed with the aim to demonstrate the positive influence of the sonic effect on the clogging of the filters, but none have been totally successful: either no clogging evidence has been observed; or failure of the filters (see Figure 9) detected as a direct consequence of the sonic effect. The most representative test was carried out with artificially polluted HDPE. Well known stainless steel (SS) particles had been incorporated at a concentration of 10 wt% within a virgin polyethylene (SABIC 6254) prior to the trial. The pressure evolution within the unit during the tests and the determination of the residual content of SS particles through burning analyses of extruded samples confirmed that sonic assistance causes direct failure of the filters. Nevertheless, a novel ultrasonically-assisted filter system was successfully developed and tested at pressures up to 350 bar.
WP5
The ultrasonic modulation and new control system were integrated into an extrusion line for trials. Due to the varying results of the ultrasonical filter it was evaluated as a standalone system and not integrated during the project.
The integrated system includes a number of signals such as pressure transducer, ultrasonics and throughput each of which was comprehensively tested to ensure functionality. The ultrasonic power applied to polymer melt is regulated by supplying a DC control voltage to the Herrmann Ultrasonic generator. Two main ultrasonic control modes were investigated in the integrated control system; continuous mode and pulse mode. Process trials showed that a saw-tooth pulse signal provided the best mode of operation for the ultrasonic system.
Power consumption for the whole control system including motor drive, ultrasonic generator and all heater bands was also measured via a watt meter connected to the main power supply.
In order to integrate the ultrasonic modulation unit with the control system, open-loop trials were successfully carried out to study the effect of ultrasonic energy on the polymer melt.
The trials showed that there is a significant reduction of die pressure at lower throughput (< 5kg / h) when the ultrasonics is activated. However, the extrudate becomes visibly damaged if the throughput is such that excessive ultrasonic energy is applied to the polymer melt.
The first generation user control interface for closed-loop control is shown. In this interface, all main parameters of extrusion process are monitored in real time including set and current values of die pressure, set and current values of throughput, screw speed, barrel pressure, power consumption, temperature and ultrasonic control signal.
A number of trials using the ultrasonic / control system were performed with recycled HDPE materials. The first set of trials focused on a single material (MFI: 0.6) and showed that it takes only approximately two minutes for the die pressure to reach a pre-set figure when both ultrasonic and temperature controllers are activated. This is far quicker than simple temperature control plus the energy consumption is lower.
Trials were also performed with a blend of two recycled HDPE of different MFI. Again, the control system was able to activate the ultrasonic in combination with temperature controllers to reach a target pressure quicker than temperature control alone and more efficiently (lower energy consumption).
The machine installation for ULTRAVISC system integration is shown. This new technology allows us to keep melt viscosity constant regardless of variations in MFIs due to differences in the source of the waste materials used. Through our experiments, die pressure can be kept constant even when increasing of throughput to 33 %. This also helps to increase the products of the extrusion process.
In summary, the project have achieved all its technical objectives and provided the SMEs with a platform for up-scaling and exploiting the ULTRAVISC technology.
Contact details:
Dr Paul Beaney
12 Derryhirk Road
Dungannon
Co. Tyrone BT71 6NH
Northern Ireland
Tel: +44-028-38853900
Fax: +44-028-38853901
Email: Paul@cherrypipes.com
List of websites: http://www.ultravisc.org
There is a need is to develop a cost-effective, potentially retrofit system to address the current deficiencies in reprocessing waste plastics with respect to plastic flow and the issues of contamination and variability in the feedstock. Motivation for progressing beyond the state-of-the-art is to encourage the large number of European Union (EU) small and medium-sized enterprises (SMEs) to continue to use and increase their ability to re-use waste plastics. The development of the ULTRAVISC technology has enabled waste plastic processors to continuously monitor and activate changes that will reduce the production of scrap / waste whilst giving the potential opportunity to produce thinner walled products through the application of ultrasound via a closed-loop feedback control system.
The ultrasonic modulation system comprises a newly developed sonotrode and ultrasonic chamber. The system is designed to operate at 20 kHz and able to deliver a range of ultrasonic energy levels based on the input power. A number of process trials using a variety of dies configurations were carried out using virgin and recycled feedstock and demonstrated the potential of ultrasonics to reduce both melt viscosity and increase material throughput.
The new closed-loop ULTRAVISC control system demonstrated that it is able to monitor process conditions and apply ultrasonics modulation as and when is required to maintain a constant viscosity. Furthermore, the system is also able to control the ultrasonics, heaters or a combination of the two in order to target a specific throughput or die pressure. A variety of recycled feedstock with different melt flow index (MFI) has been processed and it has been shown that the system is able to cope with the both sequential changes to MFI and material blends.
As in the case of the ultrasonic modulation system, the ultrasonic filter was newly developed. It is based on a reversed candle filter whereby the polymer melt runs along the length of the single-piece sonotrode and then through a candle filter. A variety of process trials were carried out to demonstrate the positive effect of ultrasonics on the unclogging of filters. The trials showed that the ultrasonic was unable to unclog the filter but did cause damage to it on several occasions. On a more positive note, the ultrasonic filter system has shown it is able to operate at pressures over 350 bar.
The project has successfully met its principle objectives; a prototype ultrasonic system with closed-loop control has been developed and successfully shown that it is able to monitor and control the viscosity of recycled feedstock in real-time. Furthermore, trials have also shown that a greater throughput is possible using ultrasonics despite the power requirements being less than a standard extruder system.
Project context and objectives:
The ULTRAVISC project which was funded under the European Commission (EC)'s Seventh Framework Programme (FP7), was carried out over a period of 24 months beginning on 1 October 2009 and ending on 30 September 2011.
ULTRAVISC addresses the EU Directive 1999/31/EC by introducing a novel processing technology that enhances the extrusion of batch-variable plastic waste and provides the means for the EU to improve recycle rates and extend recycling targets.
Plastic recycling and reducing the volume of plastic waste going to landfill is an emotive subject, highly publicised by our increasing dependence on oil, fears over future supply-demand and concern for our environment. Social responsibility and political debate have introduced increasingly stringent EU mandates aimed at encouraging the plastics industry to invest in processes and technologies that minimise the disposal of plastic waste at landfill. To enable recycled plastics to compete with virgin polymers technological developments that provide the mechanisms with which recycling and reprocessing of plastic waste becomes more economically viable and a commercially attractive alternative is needed. The main issue for extruders of recycled plastics concerns the variability of the recyclate feed material which currently requires a novel solution. We have seen an opportunity to provide a viable solution to meet this currently unmet market need.
The main technical objectives that have been achieved through process innovation are:
- Development of a variable intensity ultrasound device for the normalisation of polymer MFI to within a target MFI range of ± 5 % at the die, such that processing remains consistent and / or controlled under a specific set of extrusion conditions, regardless of variations in the bulk MFI of the recycled plastic.
- Operation of ultrasonic probes in extrusion machines with a range of process throughputs, operating temperature of up to 200 degrees of Celsius, and pressures of up to 20 MPa (200 bar), using an ultrasound frequency of 20 kHz.
The technical development work was divided into a number of work packages (WPs) each contributing to the overall objective of the project:
- WP1: Material characterisation and ultrasound
- WP2: Soft sensor
- WP3: Viscosity control
- WP4: Low / high power sonic filter
- WP5: ULTRAVISC system integration.
There is a need is to develop a cost-effective, potentially retrofit system to address the current deficiencies in reprocessing waste plastics with respect to plastic flow and the issues of contamination and variability in the feedstock. Motivation for progressing beyond the state of the art is to encourage the large number of EU SMEs to continue to use and increase their ability to re-use waste plastics. The development of the ULTRAVISC technology has enabled waste plastic processors to continuously monitor and activate changes that will reduce the production of scrap / waste whilst giving the potential opportunity to produce thinner walled products through the application of ultrasound via a closed-loop feedback control system.
The original objective was to implement a Soft-sensor to accurately monitor and control the melt viscosity of the waste plastic melt from knowledge of the processing parameters and conditions (i.e. melt temperature, screw speed, back pressure, motor torque). Control is achieved using ultrasonics to modulate melt viscosity and filter the material. However, process trials using the Soft-sensor highlighted that a different model is required for each extrusion configuration; this is considered to be somewhat impractical and expensive. Therefore, it was agreed that an alternative control strategy be used. The alternative strategy is based on proportional-integral-derivative (PID) control whereby viscosity is kept constant by controlling the ratio of pressure to throughput of the polymer melt in the die.
The ultrasonic modulation system comprises a newly developed two-part sonotrode and associated ultrasonic chamber. The system is designed to operate at 20 kHz and able to deliver a range of ultrasonic energy levels based on the input power. A number of process trials using a variety of dies configurations were carried out using virgin and recycled feedstock and demonstrated the potential of ultrasonics to reduce both melt viscosity and increase throughput.
The new closed-loop ULTRAVISC control system demonstrated that it is able to monitor process conditions and apply ultrasonics modulation as and when is required to maintain a constant viscosity. Furthermore, the system is also able to control the ultrasonics, heaters or a combination of the two in order to target a specific throughput or die pressure. A variety of recycled feedstock with different MFI has been processed and it has been shown that the system is able to cope with the both sequential changes to MFI and material blends.
As in the case of the ultrasonic modulation system, the ultrasonic filter was newly developed. It is based on a reversed candle filter whereby the polymer melt runs along the length of the single-piece sonotrode and then through a candle filter. A variety of process trials were carried out to demonstrate the positive effect of ultrasonics on the unclogging of filters. The trials showed that the ultrasonic was unable to unclog the filter but did cause damage to it on several occasions. On a more positive note, the ultrasonic filter system has shown it is able to operate at pressures over 350 bar.
In summary, the project has successfully met its principle objectives; a prototype ultrasonic system with closed-loop control has been developed and successfully shown that it is able to monitor and control the viscosity of recycled feedstock in real-time.
The project consortium originally comprised 8 participants (5 SMEs and 3 research and technological development (RTD)) with the co-ordinator being Cherry Pipes. During the course of the project one of SMEs Plastitehase (Estonia) had to withdraw due to financial difficulties. Project results:
WP1
The first WP of the project focused on the development of software / hardware architecture, materials characterisation and ultrasonics knowledge enhancement.
The hardware architecture includes a data acquisition (DAQ) system that has the ability to measure various extrusion process parameters including barrel pressure, die pressure-drop, screw speed, and temperature along the extrusion line. Software was developed to investigate the closed-loop control system and the proposed soft-sensor technique which describes the relationships between the machine parameters and melt viscosity at the die.
The viscosity model was successfully tested with virgin low-density polyethylene (LDPE) using a Collins 25 mm single-screw extruder. However, polymer feeding could not be fully controlled when working with high-density polyethylene (HDPE) and so a larger extruder (38 mm) was used for all subsequent trials.
In parallel with the work on viscosity modelling, a closed-loop control system was developed for the integration of the soft-sensor and the control system. The output of the soft-sensor is used as feedback viscosity for the purpose of control. Based on the error between the set value for viscosity and the model output, the controller has been designed to control the screw speed and the temperature in order to keep the viscosity constant across a range of polymer types and compositions. Since the viscosity is proportional to pressure drop and the reciprocal of screw speed the target of the closed-loop control system is to maintain the ratio of the two constant. This was successfully been achieved using PID control and a pc as the user interface.
Recycled HDPE routinely used by Cherry Pipes underwent process trials and subsequent characterisation in order to help build a consistent process methodology in preparation for ultrasonic trials. This work included the design and manufacture of a slit die that was used by each of the RTD participants to ensure consistency.
A finite element analysis (FEA) model was developed to help understand how ultrasound energy is transmitted from the sonotrode through the polymer melt. The effect of probe geometry, depth within the melt, and amplitude on the propagation of ultrasonic energy was been investigated. One of the main modelling results indicated that a single sonotrode is sufficient to transmit to the ultrasonic energy through the chamber.
The first stage of the ultrasonic filter development involved the evaluation of several designs which included a vibrating sonotrode in the form of a mesh located within the extrusion line. Discussions within the consortium and with third parties led to this design being abandoned as it could not be achieved within the budget and timeframe of the project. Consequently, a new design was developed in which a standalone probe is used with a candle filter.
It is important that the filter is effective in-use and so work was carried out to estimate the cleaning interval of the filter. A 100 Mesh filter was considered appropriate and it was estimated that such a filter would require a cleaning interval somewhere between 1 and 10 tonnes of recycled material. Replacement of the filter was also considered at the design stage to minimise any downtime.
WP2
The implementation of soft sensor is based on a prediction-feedback observer mechanism. A 'prediction' model generates an open-loop estimate of the melt viscosity based on the process inputs; this estimate is used as an input to a second, 'feedback' model to predict the pressure of the system. These two models were trained off-line from the process data and rheological properties of material by using genetic algorithm. The output of pressure model is compared to the actual measured melt pressure and the error used to correct the viscosity estimate in real time.
A DAQ system was built to enable the extruder screw speed and temperature settings to be controlled from the PC, as well as collection of pressure and temperature data.
As mentioned in WP1, a slit die was used in the process trials using the soft sensor.
The extruder throughput was shown to have an approximately linear relationship with screw speed. Model identification trials were carried out whereby the extruder was run at a range of screw speeds and temperature settings with different feed materials to generate data for modelling. The pressure and viscosity models calculated from genetic algorithm which fit best with the experimental data were calculated.
It was agreed by the project partners to use two virgin HDPE materials with different MFIs (Marlex HHM-TR 144, MFI: 0.18 and Sabic B6246LS, MFI: 0.5) plus a number of blends in order to 'train' the model.
The models required by the soft sensor are specific to the machine / die set-up which is a disadvantage; the end-user would need to 'train' the model for each die configuration which could be both costly and time consuming.
The viscosity prediction model is material specific but as this model is adjusted by the pressure feedback model it is not essential that this is accurate, however a new model should be investigated if the type of polymer is changed (e.g. from HDPE to PP). The prediction model can be developed doing an offline test.
Due to the requirement for the soft sensor to be trained for each extruder configuration it was agreed that an alternative control strategy be implemented; a control system that can be more easily adapted to each industrial application.
As the viscosity is proportional to the ratio pressure to throughput and its measurement may not be readily available for different die geometries, the target of viscosity control can be achieved by keeping both throughput and pressure of the polymer melt in the die constant.
In order to do this, a strategy was selected to implement two controllers separately in computer software: the first controller is used to control the ultrasonic energy or temperature to maintain constant die pressure, the second controller controls screw speed for the constant throughput of the polymer melt. Both controllers are designed with specific algorithms to run simultaneously during the extrusion process. For the purpose of keeping the die pressure constant, controller 1 uses ultrasonic only, temperature only or the combination of ultrasonic and temperature as its control variables.
WP3
A new design of sonotrode has been developed which differs considerably from the more standard cylindrical design that is only partially submerged into the polymer melt. Background knowledge and technical discussions led to a novel sonotrode design that incorporates grooves and extends through the barrel chamber. This design increases the surface area of the sonotrode that is in contact with the polymer melt which should result in a more uniform reduction in viscosity.
Prior to the manufacture of the ultrasonic modulation hardware, further understanding of the pressures in the ultrasound chamber was required to optimise the gap between sonotrode and chamber. An ultrasonic pressure test chamber with a 'dummy' sonotrode was, therefore, designed and, trials ran. Extrusion trials showed that die pressure was effectively unaffected by the size of gap. Consequently, the smaller gap was selected for the 'real' sonotrode / chamber set-up.
A new two-part grooved sonotrode fabricated from titanium was commissioned. The sonotrode is clamped top and bottom and has maximum operational amplitude of approximately 40 µm at its midpoint, and a nominal operating frequency of 20 kHz. The system is such that the 4 kW generator provides sufficient power to oscillate the sonotrode at the pre-set amplitude. Development trials work focused on running the system between 50 and 90 % of the maximum amplitude.
Initial ultrasonic trials were carried out using virgin HDPE (Marlex TR 144, MFI: 0.18) to investigate the general effect of ultrasound on-time and amplitude on the polymer melt.
%The polymer melt exiting the die when the system is running at 50 % amplitude continuously for a period ('on time') of 30 seconds is shown. The ultrasonic effect on the polymer melt is clearly evident. The splitting of the polymer melt was shown to be a result of the applying the ultrasonics for too long; no splitting occurred when the 'on-time' was 10 or 20 seconds. It was also observed that there is a delay between when the ultrasonics is applied and when the effect is seen.
The sonotrode is designed to oscillate within a fairly narrow frequency range: 20 kHz ± 500 Hz and has nodal points (zero displacement) at the clamping points (winglets). Pulsed and continuous ultrasound has been applied and it was shown by all RTDs that extended continuous application resulted in a significant rise in the melt temperature. The best mode of operation was shown to be continuous pulsing (or pulse width modulation).
The chamber is designed such that the sonotrode can be mounted either vertically or horizontally. Due to the considerable weight of the ultrasonic unit, a support stand is advised. The cost of the prototype ultrasonic hardware including the sonotrode, booster, convertor and generator is ~20,000. A commercial system would be larger in size and, consequently, the cost would be higher. However, if the use of ultrasonics is able to halve Cherry Pipes waste it could save them in excess of EUR 80 000 per annum.
A tubular die was commissioned to evaluate the effect of the ultrasonic modulation on a representative tubular profile. In order to test the robustness of the ultrasonics a number of different types of recycled feedstock were processed both sequentially and pre-homogenised. The results demonstrate that ultrasonics can increase throughput as well as reduce system energy consumption. Ultrasonic modulation has been demonstrated to have real commercial potential for recyclers such as Cherry Pipes.
Notes:
(i) Material historically used by Cherry Pipes to produce their drainage pipes
(ii) Produced in house by Cherry Pipes using recycled HDPE milk bottles
(iii) Black rechip sourced locally by Cherry Pipes
WP4
The final design of the ultrasonic filter is based on a 'reversed' candle filter (polymer melt flows from the inside of the filter outwards). The design of the probe takes into account various parameters such as ultrasonic efficiency, mechanical resistance, optimisation of the filtration surface, connection with the surrounding elements and cost. After finalising the probe design, the next step was to design the housing taking into account operational requirements such as positioning and fixation of the pressure and temperature transducers and system integration.
The dimensioning of the most critical components of the unit has been performed using the finite element method. The selected parts have been studied under an internal pressure of 200 bars which represents the ultimate specification in industrial configuration.
Optimisation has been carried out on the elements that showed some weakness under this loading. The flow propagation has also been optimised using computational fluid dynamic (CFD) analyses with objective to maximise the filtration effectiveness and consequently the cleaning intervals.
Numerous trials were performed with the aim to demonstrate the positive influence of the sonic effect on the clogging of the filters, but none have been totally successful: either no clogging evidence has been observed; or failure of the filters (see Figure 9) detected as a direct consequence of the sonic effect. The most representative test was carried out with artificially polluted HDPE. Well known stainless steel (SS) particles had been incorporated at a concentration of 10 wt% within a virgin polyethylene (SABIC 6254) prior to the trial. The pressure evolution within the unit during the tests and the determination of the residual content of SS particles through burning analyses of extruded samples confirmed that sonic assistance causes direct failure of the filters. Nevertheless, a novel ultrasonically-assisted filter system was successfully developed and tested at pressures up to 350 bar.
WP5
The ultrasonic modulation and new control system were integrated into an extrusion line for trials. Due to the varying results of the ultrasonical filter it was evaluated as a standalone system and not integrated during the project.
The integrated system includes a number of signals such as pressure transducer, ultrasonics and throughput each of which was comprehensively tested to ensure functionality. The ultrasonic power applied to polymer melt is regulated by supplying a DC control voltage to the Herrmann Ultrasonic generator. Two main ultrasonic control modes were investigated in the integrated control system; continuous mode and pulse mode. Process trials showed that a saw-tooth pulse signal provided the best mode of operation for the ultrasonic system.
Power consumption for the whole control system including motor drive, ultrasonic generator and all heater bands was also measured via a watt meter connected to the main power supply.
In order to integrate the ultrasonic modulation unit with the control system, open-loop trials were successfully carried out to study the effect of ultrasonic energy on the polymer melt.
The trials showed that there is a significant reduction of die pressure at lower throughput (< 5kg / h) when the ultrasonics is activated. However, the extrudate becomes visibly damaged if the throughput is such that excessive ultrasonic energy is applied to the polymer melt.
The first generation user control interface for closed-loop control is shown. In this interface, all main parameters of extrusion process are monitored in real time including set and current values of die pressure, set and current values of throughput, screw speed, barrel pressure, power consumption, temperature and ultrasonic control signal.
A number of trials using the ultrasonic / control system were performed with recycled HDPE materials. The first set of trials focused on a single material (MFI: 0.6) and showed that it takes only approximately two minutes for the die pressure to reach a pre-set figure when both ultrasonic and temperature controllers are activated. This is far quicker than simple temperature control plus the energy consumption is lower.
Trials were also performed with a blend of two recycled HDPE of different MFI. Again, the control system was able to activate the ultrasonic in combination with temperature controllers to reach a target pressure quicker than temperature control alone and more efficiently (lower energy consumption).
The machine installation for ULTRAVISC system integration is shown. This new technology allows us to keep melt viscosity constant regardless of variations in MFIs due to differences in the source of the waste materials used. Through our experiments, die pressure can be kept constant even when increasing of throughput to 33 %. This also helps to increase the products of the extrusion process.
In summary, the project have achieved all its technical objectives and provided the SMEs with a platform for up-scaling and exploiting the ULTRAVISC technology.
Contact details:
Dr Paul Beaney
12 Derryhirk Road
Dungannon
Co. Tyrone BT71 6NH
Northern Ireland
Tel: +44-028-38853900
Fax: +44-028-38853901
Email: Paul@cherrypipes.com
List of websites: http://www.ultravisc.org
