Final Report Summary - OPTIJECT (A novel NIR-based instrument for in-line monitoring during injection moulding)
Executive summary:
Plastic products are present in all areas of our daily life from toys, kitchen utensils, bottles and electric appliances to car parts and approximately 30 % of them are processed by injection moulding, making it the major plastic converting technique. A proper control of the process as well as material parameters such as temperature, moisture, colour and speed of injection is required to obtain good parts repeatedly. However this can be difficult and currently most adjustments are based on quality control carried out after injecting the parts, leading to economic losses due to the amount of scrap generated. A technology that would allow moulders to monitor and control the injection process inline would bring substantial benefits due to improved quality control practices, increased productivity associated with reduced scrap and improved cost competitiveness.
The OPTIJECT project answers to these different need as it has successfully developed an online system for monitoring the injection moulding process with the view of improving the quality of resulting parts. The OPTIJECT system is based on the development of a tailored optical probe capable of performing both reflectance and transmission measurement in near infrared (NIR) and visible (VIS) ranges whilst generally enduring the harsh processing conditions of injection moulding. In addition, the OPTIJECT system has been calibrated for online monitoring of a number of parameters both in terms of the properties of the polymer stream, a key novelty, and of the process parameters during the part injection.
The OPTIJECT probes are located within the machine nozzle area and can be suited to different injection machines since only the mechanical interface where the probes are installed needs to be adjusted.
An initial survey of the industry to define the parameters on which the project should focus, i.e. those that industrials believe is the greatest causes for defects. In particular, in terms of NIR, OPTIJECT was shown to allow access to the material moisture and degradation levels in real time, i.e. possibly even before the part has left the mould, which is a real breakthrough. In terms of Vis, the system allows online determination of the colour, i.e. Commission internationale de l'éclairage (CIE) LAB value, and via a coupling with the determination of the concentration of the colour additive via NIR, it possible to optimise the amount of expensive masterbatch to provide a given colour to the final part with no need to visual colour matching. Although probes exist for the monitoring of some of the process variables (e.g. temperature, speed and pressure) but often these values are taken at the level of the machine as opposed to OPTIJECT which will allow measuring those variables via NIR within the polymer stream to reach a more reliable picture of the effect of the process on the quality of the material. Indeed, an additional feature incorporated in the OPTIJECT software is the artificial intelligence aimed at reducing the amount of time it takes to set-up, turn around and diagnose in process problems based on the information from the OPTIJECT system and previously programmed responses based on expert knowledge. This decision support system is based on the processing in real time of the NIR and VIS data and allows flagging any parameter which is out of specification or even advising the operator on the changes needed after some training (which could not be fully performed during the OPTIJECT project due to lack of time).
Further abilities of the system, after related calibration, include monitoring of infrared (IR) active additives such as chemical blowing agent, food contact contaminants, copolymer ratio, etc.
After the development period and hardware integration with the software, the OPTIJECT prototype was used during two demonstration sessions to the industry especially through its installation of injection moulding machines of different sizes.
The project also included other activities targeted at maximising the potential for the OPTIJECT system. This included the management intellectual properties, a preliminary market study and business plan for exploiting the new system but also number of dissemination activities throughout the project. The public website of the project (see http://www.optiject.eu) informs the public on the technology, as well as the latest news and the progress of the project.
Project context and objectives:
Plastic products are present in all areas of our daily life from toys, kitchen utensils, bottles and electric appliances to car parts and approximately 30 % of them are processed by injection moulding making it the major plastic converting technique. The proper control of the process as well as material parameters such as temperature, moisture, colour and speed of injection is required to obtain good parts repeatedly. However this can be difficult and currently most adjustments are based on quality control carried out after injecting the parts, generating economic losses due to the amount of scrap generated. A technology that would allow moulders to monitor and control the injection process inline would bring substantial benefits due to improved quality control practices, increased productivity associated with reduced scrap and improved cost competitiveness.
The injection moulding process represents a complex series of events that ultimately affect the quality of the final part so a thorough understanding of each process step is needed:
1. heat up polymer until it melts
2. inject molten polymer into an empty cavity to give it a final form
3. cool the polymer down so it stays the same shape as the mould.
The entire injection moulding process can be summarised in a flowchart. Therefore, the operator has a limited magnitude of adjustment of process parameters to control the series on hidden steps along the process that may affect the parts quality checked by offline a posteriori control and for which OPTIJECT may now give an answer.
The most common practices used at injection moulders who mass produce commodity goods (e.g. drinks bottles, vacuum cleaners and automobiles) include the following practices with an aim to ensure part quality:
1. drying of material using the resin manufacturers specifications
2. visual inspection of colour and distortion (during setting and production)
3. part stability and shrinkage is measured during machine setting on small batches of parts and sporadically throughout production as part of an International Organisation for Standardisation (ISO) quality control regime
4. residual stresses and physical strength are evaluated during the machine setting and are performed by either impact testing or tensile testing.
Producers of high tech goods and pharmaceutical products use the above quality control steps as well as:
1. humidity and moisture measurements after the aforementioned drying step
2. melt flow index or shear viscosity measurements of material to be processed (monitor change in flow properties)
3. differential scanning calorimetry for identification of glass transition (freezing), melting temperature and the oxidative induction time (Identification of correct temperature and maximum processing time)
4. spectrometric determination of colour, translucency and the parts response to polarised light (sample consistency)
5. accurate three-dimensional (3D) measurements using automated equipment to ensure micrometer tolerances (3D scanning equipment to monitor wrapage and shrinkage).
These steps implemented by moulders are only able to monitor the properties of either the raw materials or of the produced parts and do not provide direct information about the materials properties or moulding conditions.
Only a limited number and types of on- and in-line control systems are currently available. The use of sensors to monitor machine variables such as the barrel temperature, melt front velocity, injection pressure and mould temperature rely on indirect measurements coming from parts of the injection moulder itself and not the actual polymer. An example of this remote monitoring is the processing temperature; the processing temperature of polyethylene terephthalate (PET) is approximately 280 °C so you set the barrel temperature to 280 °C but is it the temperature of the polymer in the barrel? The answer is no! The actual melt temperature is much more complex and is influenced by multiple factors including the injection speed, diameter of the barrel and nozzle, backpressure, additives and polymer viscosity. The actual temperature has been found to change by as much as 25 °C from the barrel walls to its centre therefore the temperature measurement is a measure of the metals temperature that surrounds the polymer; this is also dependent on the accuracy of your thermocouple.
Sensors used in injection moulding have to be tough and hard wearing so are often based on simple thermo-couples or piezoelectric devices which therefore can only supply information that is directly measured or calculated using previously established relationships e.g. apparent viscosity.
These indirect measurements are used to minimise the appearance of moulding defects such as dieseling or flash by providing information about the overriding factors causing the problems such as excessive injection pressure and speed. These sensors are however unable to tell if the material is burning or has an excessive moisture content.
The injection moulder is able to monitor machine conditions but cannot give direct process or part information and because of this does not provide 'real' information about the goods being produced and cannot therefore act against scrap.
In such context, OPTIJECT investigated the online use of NIR spectroscopy. Spectroscopy in general uses radiation from the electromagnetic spectrum as a way to provide energy to a sample. The difference between the radiations in the electromagnetic spectrum for example X-rays and radio waves, is the frequency and the energy of the associated photons. The energy of the radiation used can be found using Planks relationship 'E = h?', where E is the energy of the photon (light source in this instance), h is Planks constant and ? is the frequency of the electromagnetic wave. Basically, different types of radiation have different energies but they are all methods of providing energy.
NIR spectroscopy uses non-visible light to excite the chemical bonds in organic and inorganic materials. A light source with an emissive range between 700 and 5000 nm is used to irradiate a sample and cause excitation of some of the bonds present in the sample which absorb a small part of the incident light. The light that returns to the detector is now missing characteristic wavelengths of light which are known to relate precisely to molecular vibrations and rotations. The identification of specific bonds present in the sample can be used to calculate what molecule or combinations of molecules are in the sample. NIR spectroscopy is particularly good at detecting molecules with a large difference between the atomic weight of the atoms in the bond e.g. hydroxide (OH) and NH.
In addition, OPTIJECT explored the possibility on combining NIR with VIS spectroscopy. VIS spectrometry uses light with a wavelength of 380 to 780 nm to induce electronic excitations in the sample being irradiated. The technique is called VIS spectrometry because it utilises a light source with a wavelength detectable by the human eye. In the same way that NIR spectroscopy works, a sample is irradiated and if a particular wavelength of light matches the energy needed to excite a ground state electron (commonly but could be greater) to a higher energy state this wavelength range will be absent from the returning light to the detector. These packets of light (quanta) or energy correspond to chemical groups with spare orbital's or unpaired electrons in the sample and can be used to identify the molecule. Additionally, the ability of the human eye to detect light in this range can be used to calculate the colour of objects. As an example, an object that appears blue to the eye has had the majority of wavelengths above 480 nm absorbed by the sample and thus appears blue.
CIE have developed a system with which the results from VIS spectrometry can by projected onto a colour space that is easily interpreted by the non-spectroscopist. The most common colour space, the LAB space can be interpreted by:
1. L* is scaled between 0 and 100 and can be interpreted as either the white / brightness of the sample or the transparency of the sample dependant upon the observation mode
2. A* represents the colours green and red who's colour intensity is defined by the magnitude of the A* value e.g. of A* = -3 would represent green and A* = 3 would represent red
3. B* represents blue and yellow colours with blue values being negative and yellow positive.
The CIE LAB space is source independent and provides the same value regardless of the source energy distribution (light temperature).
Overall, the objective of this project centred on the development of an online system for monitoring the injection moulding process with the view of improving the quality of resulting parts. A tailored optical probe combination allows performing both reflectance and transmission measurement in NIR and VIS ranges. In addition, the OPTIJECT system has been calibrated for online monitoring of a number of material parameters both in terms of the properties of the polymer stream and of the process parameters during the part injection.
Firstly, the participating small and medium sized enterprises (SMEs), as well as a representative sample of European plastic converters were consulted and results used to determine their needs and specifications and define the industrial specifications for the OPTIJECT prototype. This bottom-up approach ensured that the prototype which would be tested and validated in a real industrial environment, would meet not only the technological requirements of industry, but socioeconomic requirements.
Secondly, the lab scale tests using NIR and VIS spectroscopy started from the beginning of the project and focussed on monitoring various materials parameters of greatest importance for the SMEPs such as colour, humidity and degradation during a large number of off line tests (e.g. by the study of injected material and pellets). Subsequently, when suitable probes were available to be installed in the injection machine, similar tests were performed online. As a result, calibration curves have been obtained for these different parameters, along with the spectral ranges of interest fuelling into the definition of the features that the prototype needed to meet.
Then, a subsequent goal was to deliver the design of the OPTIJECT prototype including a system processing unit (with VIS and NIR spectrometers), optical fibre and connectors, optical probe and mechanical interface with the injection moulding machine. It has been built and used in thorough functional tests during which various iterations of improvements have been necessary.
In addition, various artificial intelligence techniques have been evaluated for their suitability for the treatment of the chemometric data resulting from the online monitoring of the injection process using the OPTIJECT hardware but also later for the accounting of additional qualitative criteria set by the operator. They have been programmed in specific OPTIJECT software.
The pre-competitive OPTIJECT prototype consisting both of hardware, as part of work package three (WP3) and software (WP4), assembled in WP5 have been delivered and validated at an industrial scale at different participating SMEs in WP6 and demonstrated to the industry in WP7.
A further goal was to carry out demonstration activities proving the viability of the OPTIJECT system, outline its potential economic and environmental advantages. In terms of industrial validation trials, they were done in injection machines of different features and their specific requirements especially in terms of using the visible system for online colour detection.
The overriding goal of this project was to ensure that the pre-competitive OPTIJECT prototype would fulfil the threshold requirements of the industry to ensure its further development post-project into a fully industrial system that would be taken to market, where its beneficial impact on the quality of injected parts could be felt at European level. This was also empowered by the realisation of dissemination activities throughout the project and the layout of an exploitation strategy by the consortium.
Project results:
The prototype system developed during the OPTIJECT project transcends the boundary of the current state-of-the-art in monitoring the injection moulding process, as it is capable of simultaneously measuring real-time parameters that are critical for the quality of the final moulded parts and the concentrations of the different compounds. Should it reach the market, it will overcome many of the limitations of the existing practices which currently do not allow in-line measurements of all the parameters targeted in the project, especially regarding the materials.
The main scientific and technological (S&T) elements of foreground consist in:
1. calibration curves for using the OPTIJECT system. This foreground consists in the set of specific methodologies for the calibration of various plastic parameters, as well as the resulting NIR calibrations in given conditions. This 'library' of conditions results from over 1 000 experiments performed in UNEXE.
2. OPTIJECT probe design: a robust probe withstanding the high temperature and pressure in the injection moulder that can simultaneously work in VIS and NIR range with a version for transmittance and a version for reflection with the ability to work also as a receiver for transmittance mode.
3. OPTIJECT hardware and software: the hardware consist in a system processing unit (with VIS and NIR spectrometers), the above described probes and mechanical interface with the injection moulding machine. Software for processing the VIS and NIR data was also provided with intelligent features to guide the user in the improvement of part quality.
More details are given below regarding the S&T features of each element of foreground.
Parameters that were calibrated
In response to the needs identified in the stakeholder study performed in the past project in terms of factors driving defects during the injection moulding process, the consortium selected several parameters to calibrate the system. The system has been shown to extent the state-of-the-art through the identification of visual information as well as spectral information that was previously achieved in the literature as described in the previous paragraph.
The tests performed have focussed on materials parameters such as moisture, colour concentration and short term polymer degradation indicators, but also on the determination of the machine variables such as pressure, temperature and flow speed (with temperature showing a particular affinity for measurement via NIR), leading to the delivery of the specifications of the system to use and for the conditions to calibrate it.
Industrial processers can benefit from immediate quality improvement potential thanks to moisture detection. Elevated moisture levels cause premature degradation and discolouration of materials throughout processing. Most polymers are hydroscopic; plastics processors therefore dry the material before use according to manufacturer's recommendation. This drying process is seldom followed and may lead to 'wet' material being processed causing a knock on effect to processing and part quality. In addition, moulders face difficulties to identify the cause of defects as they often lack of a moisture meter in their facilities. The use of NIR to detect moisture is well documented for off-line tests but direct identification during processing at elevated temperatures has not been performed outside of the OPTIJECT project. The strong absorption of water in the IR region makes it readily discernible from other spectral features and is somewhat process independent as a consequence. The predicted moisture concentration can be that of the polymer in the hopper or of the moulded part and has a parts-per-million accuracy.
Colour monitoring thanks to OPTIJECT is believed to have the potential to become an online colour quality control and allow parts manufacturers to terminate with a non reliable reliance upon visual inspection.
While being a low cost component, the VIS part of the OPTIJECT system has been shown to be one of the most versatile measuring components due to the ability to deal with large fluctuations in moulding variables. CIE LAB values, which can be calculated after the VIS spectra, are an absolute representation of the tri-stimulus values our eyes use to identify colour projected into a colour space providing a representation of perceived colour. The software developed throughout the project detects whether the part meets the predetermined CIE LAB value. This capability of the system is of international importance as signified by acceptance of our findings to the International Polymer Processing Symposium (PPS 28).
The colorimetric analysis of molten PET and polypropylene (PP) has been performed during injection moulding respectively for red and blue colour in transmission and reflection (further details on the conditions of the tests can be found on the project website, article 'Real-time-vis-spectroscopic-monitoring-of-colour-during-injection-moulding'). Results taken during the injection moulding processing for variable amounts of masterbatch allowed calculating CIE LAB values. Online data were compared to offline data showing good agreement and therefore the potential to provide an absolute colour value for produced parts. The greatest difference between the on- and offline CIE LAB values can be found in the yellow characteristic in the B value. The difference in translucence between solid and molten parts is believed to be the source of this discrepancy, but such difference could be easily taken into account by the OPTIJECT software in case the colour specifications are based on CIELAB values measured offline.
In addition, the NIR system was used to monitor the real colour concentration in the polymer online in the melt. The online PET colour calibration has been generated using 20 repeat spectra collected at 21 different concentrations of masterbatch. Overall, the conversion of VIS spectroscopic data to the CIE LAB colour space has allowed the monitoring of moulded part colour fluctuation shot by shot and the combination with NIR data regarding the colour concentration holds the possible benefit of optimising the colour concentration to reach a target colour.
In addition to the previously presented parameters which were widely studied, the proof of feasibility of monitoring contaminants as well as of chemical blowing agents were made. The system was used during the processing of recycled PET previously used for food packaging which hast strict European Union (EU) guidelines and trace amounts of oxo-degradation additives and contaminants from food stuffs, e.g. limonene, could be detected. Based on a preliminary calibration, the opportunities for the OPTIJECT system are vast when considering the range of moulded products that are produced using injection moulding and via other processes such as extrusion.
Hardware
It should be noted that the proposed system was also prepared and tested in transmission using a second receiving probe.
Software
The OPTIJECT software is based on an artificial intelligence (AI) feature to give process set up recommendations to the moulder. At the moment, it can mainly flag single parameters which are out of operator's expectations such as colour, colour concentration, humidity, but further training is needed to provide solutions in terms of modifications of the process parameters that would lead to parts with better quality.
Thus, OPTIJECT technology can bring the following key advancements to the injection moulding process:
1. monitoring the material directly, which is completely novel in the industry
2. enabling recommendations of readjustment of the injection moulding process to improve part quality
3. most effective positioning of the NIR sensors in the nozzle of the injection moulding machine as opposed to the monitoring of single cavities - monitoring of the flow as opposed to the finished component
4. ability to rapidly measure a broad spectrum of critical parameters with good selectivity and via a single system - in terms of moisture the accuracy is as good as offline devices such as Karl Ficher (ie. ppm range)
5. ability to monitor raw materials batch-to-batch variations causing unexpected defects at the moulder
6. flexibility to shift between NIR and visible, in transmission or reflection depending of the mode that may be best suited for each set of parameter and material monitored, without handling probes.
Potential impact:
The results of the project consist of a precompetitive prototype monitoring system (system hardware, software and interface) with calibration curves for a range of parameters and a methodology highlighting how to implement and calibrate the system in new injection operations (technical service that will belong to the post-project exploitation of the system). As a conclusion of the information gathered in WP1, the general strengths, weaknesses, opportunities and threats of the system were outlined and it was shown that the OPTIJECT system, as a tool for controlling the injection process using NIR and visible spectrometers could meet a real need in the plastic converting industries to assist them in supplying ends users with parts of better quality and reducing scrap. It will also involve the supply and value chain whereby machine and sensors manufacturers will play a key role. As previously mentioned, the SMEs have jointly applied for a demonstration activity to lead the OPTIJECT system to the market.
In terms of economic aspects, one of the key economic requisite from the SMEs regarding the OPTIJECT system was that it should be low cost to open a wider market by making it commercially attractive to prospective clients. This has been one of the key challenges met during the up-scaling of the prototype and need further improvements to optimise the cost of the system as the cheapest version did not allow working online in the injection machine and only a version of greater price worked in those conditions.
There are numerous socioeconomic impacts that will be derived from the results of this OPTIJECT research project. Central to the expected socioeconomic impacts, is the boosting of the competitiveness of companies operating along the injection moulding chain- from raw materials and sensors suppliers to parts moulders - by improving the quality of European injection moulded plastic products and decreasing their scrap. The OPTIJECT system will be introduced in a very large market. Indeed there are well over 13 000 SME injection moulders in EU27 and an average production of 200 million of plastic components per company with an average annual growth of about 5 %. The total revenues are estimated at approximately EUR 280 billion. Directly and indirectly, 1.5 million jobs can be accredited to the European injection moulding industry. Making this industry more competitive will have impacts for the European economy, growth and jobs.
Furthermore, by virtue of their technological features, the OPTIJECT systems could improve working conditions by doing away with time consuming and chemical-based sampling techniques.
There are also environmental benefits associated with the OPTIJECT systems. Quality monitoring of materials will prevent batches unfit for injection from entering the processing chain. This will prevent the waste of energy, raw materials and resources. Furthermore, OPTIJECT will assist European producers to adhere to the various European Commission (EC) guidelines that exist in the plastic industry, such as, for example, for processed items intended for food packaging and other even more stringent guidelines when it comes to medical applications. Moreover, by reducing the energy costs involved in plastics processing, as well as raw materials consumption and waste (due to lower scrap generation), the results of this project will also contribute to assisting the EU in meeting its carbon dioxide (CO2) cutting objectives under the Kyoto protocol.
In addition, further opportunities were disclosed during the OPTIJECT project. In particular, while the system focussed on the validation of the injection moulding process, the part quality depends not only on the injection process itself, but also on the raw material quality and consistency. Therefore, the use of the similar system at the manufacturer of pellets would allow a global management of the quality parameters with a similar system throughout the supply and value chain. The development is reduced and less challenging and will be undertaken here opening up potential in an additional 30 % market share of the plastic converting techniques.
Main dissemination and exploitation activities:
The expected economic and technical benefits of the OPTIJECT technology have served as basis for the dissemination and exploitation activities. As previously mentioned, OPTIJECT is expected to have a significant impact on the quality management in the plastic industry, whereby injection moulding is a very important conversion technique.
The project included a series of other activities targeted at maximising the potential for the OPTIJECT system.
A great importance was given to the management of the intellectual property rights generated and in agreement to the dissemination of non-confidential information throughout the project. A patent review was carried out and concluded that OPTIJECT does not infringe upon existing protected intellectual property (IP). Although the project developed significant know-how exploitable by the SMEs such as specific designs for the probes, the calibration curves for their injection processes and material parameters of interest, their novelty did not allow filling a patent.
The plastic market in which OPTIJECT will be introduced is a highly promising one, following an upward trend in terms of volume produced but also of quality aspirations and scrap reduction. In addition, to date none of the currently available solutions offer the technological innovations or performance of OPTIJECT. The foreground generated by the OPTIJECT project is broad giving rise to the potential exploitation opportunities for each member of the consortium. A preliminary business plan as well as post project development work have been laid out benefitting from the synergistic role of the partners covering the whole supply and value chain. Based on the results of the project, the SMEs have access to a system that they can exploit on the market place. They jointly applied for a demonstration activity project in order to bridge the gap to commercialisation. Findings from the industry questionnaire-based survey and the industry consultations have revealed a willingness to invest in such a system in view of tangible technical and envisaged economic benefits. These figures predict a fast return on investment for OPTIJECT system.
Non-confidential information on the results and principles of the OPTIJECT technology has been disseminated throughout the project. The public website of the project (see http://www.optiject.eu online) informs the public and interested parties on the basics of the technology, latest news and the progress of the project insofar as it can be disseminated without threatening the proper protection of the developed intellectual property. Various press releases were circulated in professional plastic technology magazines and websites, raising the awareness of OPTIJECT both in industry and in the public domain and generating positive feedback. The project partners also took part in number of trade fairs. In total over 50 dissemination efforts were realised with contribution from the whole consortium.
Project website: http://www.optiject.eu
Contact details for further information:
Dr Ing. Elodie Bugnicourt
IRIS Innovació i Recerca Industrial i Sostenible, Spain
ebugnicourt@iris.cat
Plastic products are present in all areas of our daily life from toys, kitchen utensils, bottles and electric appliances to car parts and approximately 30 % of them are processed by injection moulding, making it the major plastic converting technique. A proper control of the process as well as material parameters such as temperature, moisture, colour and speed of injection is required to obtain good parts repeatedly. However this can be difficult and currently most adjustments are based on quality control carried out after injecting the parts, leading to economic losses due to the amount of scrap generated. A technology that would allow moulders to monitor and control the injection process inline would bring substantial benefits due to improved quality control practices, increased productivity associated with reduced scrap and improved cost competitiveness.
The OPTIJECT project answers to these different need as it has successfully developed an online system for monitoring the injection moulding process with the view of improving the quality of resulting parts. The OPTIJECT system is based on the development of a tailored optical probe capable of performing both reflectance and transmission measurement in near infrared (NIR) and visible (VIS) ranges whilst generally enduring the harsh processing conditions of injection moulding. In addition, the OPTIJECT system has been calibrated for online monitoring of a number of parameters both in terms of the properties of the polymer stream, a key novelty, and of the process parameters during the part injection.
The OPTIJECT probes are located within the machine nozzle area and can be suited to different injection machines since only the mechanical interface where the probes are installed needs to be adjusted.
An initial survey of the industry to define the parameters on which the project should focus, i.e. those that industrials believe is the greatest causes for defects. In particular, in terms of NIR, OPTIJECT was shown to allow access to the material moisture and degradation levels in real time, i.e. possibly even before the part has left the mould, which is a real breakthrough. In terms of Vis, the system allows online determination of the colour, i.e. Commission internationale de l'éclairage (CIE) LAB value, and via a coupling with the determination of the concentration of the colour additive via NIR, it possible to optimise the amount of expensive masterbatch to provide a given colour to the final part with no need to visual colour matching. Although probes exist for the monitoring of some of the process variables (e.g. temperature, speed and pressure) but often these values are taken at the level of the machine as opposed to OPTIJECT which will allow measuring those variables via NIR within the polymer stream to reach a more reliable picture of the effect of the process on the quality of the material. Indeed, an additional feature incorporated in the OPTIJECT software is the artificial intelligence aimed at reducing the amount of time it takes to set-up, turn around and diagnose in process problems based on the information from the OPTIJECT system and previously programmed responses based on expert knowledge. This decision support system is based on the processing in real time of the NIR and VIS data and allows flagging any parameter which is out of specification or even advising the operator on the changes needed after some training (which could not be fully performed during the OPTIJECT project due to lack of time).
Further abilities of the system, after related calibration, include monitoring of infrared (IR) active additives such as chemical blowing agent, food contact contaminants, copolymer ratio, etc.
After the development period and hardware integration with the software, the OPTIJECT prototype was used during two demonstration sessions to the industry especially through its installation of injection moulding machines of different sizes.
The project also included other activities targeted at maximising the potential for the OPTIJECT system. This included the management intellectual properties, a preliminary market study and business plan for exploiting the new system but also number of dissemination activities throughout the project. The public website of the project (see http://www.optiject.eu) informs the public on the technology, as well as the latest news and the progress of the project.
Project context and objectives:
Plastic products are present in all areas of our daily life from toys, kitchen utensils, bottles and electric appliances to car parts and approximately 30 % of them are processed by injection moulding making it the major plastic converting technique. The proper control of the process as well as material parameters such as temperature, moisture, colour and speed of injection is required to obtain good parts repeatedly. However this can be difficult and currently most adjustments are based on quality control carried out after injecting the parts, generating economic losses due to the amount of scrap generated. A technology that would allow moulders to monitor and control the injection process inline would bring substantial benefits due to improved quality control practices, increased productivity associated with reduced scrap and improved cost competitiveness.
The injection moulding process represents a complex series of events that ultimately affect the quality of the final part so a thorough understanding of each process step is needed:
1. heat up polymer until it melts
2. inject molten polymer into an empty cavity to give it a final form
3. cool the polymer down so it stays the same shape as the mould.
The entire injection moulding process can be summarised in a flowchart. Therefore, the operator has a limited magnitude of adjustment of process parameters to control the series on hidden steps along the process that may affect the parts quality checked by offline a posteriori control and for which OPTIJECT may now give an answer.
The most common practices used at injection moulders who mass produce commodity goods (e.g. drinks bottles, vacuum cleaners and automobiles) include the following practices with an aim to ensure part quality:
1. drying of material using the resin manufacturers specifications
2. visual inspection of colour and distortion (during setting and production)
3. part stability and shrinkage is measured during machine setting on small batches of parts and sporadically throughout production as part of an International Organisation for Standardisation (ISO) quality control regime
4. residual stresses and physical strength are evaluated during the machine setting and are performed by either impact testing or tensile testing.
Producers of high tech goods and pharmaceutical products use the above quality control steps as well as:
1. humidity and moisture measurements after the aforementioned drying step
2. melt flow index or shear viscosity measurements of material to be processed (monitor change in flow properties)
3. differential scanning calorimetry for identification of glass transition (freezing), melting temperature and the oxidative induction time (Identification of correct temperature and maximum processing time)
4. spectrometric determination of colour, translucency and the parts response to polarised light (sample consistency)
5. accurate three-dimensional (3D) measurements using automated equipment to ensure micrometer tolerances (3D scanning equipment to monitor wrapage and shrinkage).
These steps implemented by moulders are only able to monitor the properties of either the raw materials or of the produced parts and do not provide direct information about the materials properties or moulding conditions.
Only a limited number and types of on- and in-line control systems are currently available. The use of sensors to monitor machine variables such as the barrel temperature, melt front velocity, injection pressure and mould temperature rely on indirect measurements coming from parts of the injection moulder itself and not the actual polymer. An example of this remote monitoring is the processing temperature; the processing temperature of polyethylene terephthalate (PET) is approximately 280 °C so you set the barrel temperature to 280 °C but is it the temperature of the polymer in the barrel? The answer is no! The actual melt temperature is much more complex and is influenced by multiple factors including the injection speed, diameter of the barrel and nozzle, backpressure, additives and polymer viscosity. The actual temperature has been found to change by as much as 25 °C from the barrel walls to its centre therefore the temperature measurement is a measure of the metals temperature that surrounds the polymer; this is also dependent on the accuracy of your thermocouple.
Sensors used in injection moulding have to be tough and hard wearing so are often based on simple thermo-couples or piezoelectric devices which therefore can only supply information that is directly measured or calculated using previously established relationships e.g. apparent viscosity.
These indirect measurements are used to minimise the appearance of moulding defects such as dieseling or flash by providing information about the overriding factors causing the problems such as excessive injection pressure and speed. These sensors are however unable to tell if the material is burning or has an excessive moisture content.
The injection moulder is able to monitor machine conditions but cannot give direct process or part information and because of this does not provide 'real' information about the goods being produced and cannot therefore act against scrap.
In such context, OPTIJECT investigated the online use of NIR spectroscopy. Spectroscopy in general uses radiation from the electromagnetic spectrum as a way to provide energy to a sample. The difference between the radiations in the electromagnetic spectrum for example X-rays and radio waves, is the frequency and the energy of the associated photons. The energy of the radiation used can be found using Planks relationship 'E = h?', where E is the energy of the photon (light source in this instance), h is Planks constant and ? is the frequency of the electromagnetic wave. Basically, different types of radiation have different energies but they are all methods of providing energy.
NIR spectroscopy uses non-visible light to excite the chemical bonds in organic and inorganic materials. A light source with an emissive range between 700 and 5000 nm is used to irradiate a sample and cause excitation of some of the bonds present in the sample which absorb a small part of the incident light. The light that returns to the detector is now missing characteristic wavelengths of light which are known to relate precisely to molecular vibrations and rotations. The identification of specific bonds present in the sample can be used to calculate what molecule or combinations of molecules are in the sample. NIR spectroscopy is particularly good at detecting molecules with a large difference between the atomic weight of the atoms in the bond e.g. hydroxide (OH) and NH.
In addition, OPTIJECT explored the possibility on combining NIR with VIS spectroscopy. VIS spectrometry uses light with a wavelength of 380 to 780 nm to induce electronic excitations in the sample being irradiated. The technique is called VIS spectrometry because it utilises a light source with a wavelength detectable by the human eye. In the same way that NIR spectroscopy works, a sample is irradiated and if a particular wavelength of light matches the energy needed to excite a ground state electron (commonly but could be greater) to a higher energy state this wavelength range will be absent from the returning light to the detector. These packets of light (quanta) or energy correspond to chemical groups with spare orbital's or unpaired electrons in the sample and can be used to identify the molecule. Additionally, the ability of the human eye to detect light in this range can be used to calculate the colour of objects. As an example, an object that appears blue to the eye has had the majority of wavelengths above 480 nm absorbed by the sample and thus appears blue.
CIE have developed a system with which the results from VIS spectrometry can by projected onto a colour space that is easily interpreted by the non-spectroscopist. The most common colour space, the LAB space can be interpreted by:
1. L* is scaled between 0 and 100 and can be interpreted as either the white / brightness of the sample or the transparency of the sample dependant upon the observation mode
2. A* represents the colours green and red who's colour intensity is defined by the magnitude of the A* value e.g. of A* = -3 would represent green and A* = 3 would represent red
3. B* represents blue and yellow colours with blue values being negative and yellow positive.
The CIE LAB space is source independent and provides the same value regardless of the source energy distribution (light temperature).
Overall, the objective of this project centred on the development of an online system for monitoring the injection moulding process with the view of improving the quality of resulting parts. A tailored optical probe combination allows performing both reflectance and transmission measurement in NIR and VIS ranges. In addition, the OPTIJECT system has been calibrated for online monitoring of a number of material parameters both in terms of the properties of the polymer stream and of the process parameters during the part injection.
Firstly, the participating small and medium sized enterprises (SMEs), as well as a representative sample of European plastic converters were consulted and results used to determine their needs and specifications and define the industrial specifications for the OPTIJECT prototype. This bottom-up approach ensured that the prototype which would be tested and validated in a real industrial environment, would meet not only the technological requirements of industry, but socioeconomic requirements.
Secondly, the lab scale tests using NIR and VIS spectroscopy started from the beginning of the project and focussed on monitoring various materials parameters of greatest importance for the SMEPs such as colour, humidity and degradation during a large number of off line tests (e.g. by the study of injected material and pellets). Subsequently, when suitable probes were available to be installed in the injection machine, similar tests were performed online. As a result, calibration curves have been obtained for these different parameters, along with the spectral ranges of interest fuelling into the definition of the features that the prototype needed to meet.
Then, a subsequent goal was to deliver the design of the OPTIJECT prototype including a system processing unit (with VIS and NIR spectrometers), optical fibre and connectors, optical probe and mechanical interface with the injection moulding machine. It has been built and used in thorough functional tests during which various iterations of improvements have been necessary.
In addition, various artificial intelligence techniques have been evaluated for their suitability for the treatment of the chemometric data resulting from the online monitoring of the injection process using the OPTIJECT hardware but also later for the accounting of additional qualitative criteria set by the operator. They have been programmed in specific OPTIJECT software.
The pre-competitive OPTIJECT prototype consisting both of hardware, as part of work package three (WP3) and software (WP4), assembled in WP5 have been delivered and validated at an industrial scale at different participating SMEs in WP6 and demonstrated to the industry in WP7.
A further goal was to carry out demonstration activities proving the viability of the OPTIJECT system, outline its potential economic and environmental advantages. In terms of industrial validation trials, they were done in injection machines of different features and their specific requirements especially in terms of using the visible system for online colour detection.
The overriding goal of this project was to ensure that the pre-competitive OPTIJECT prototype would fulfil the threshold requirements of the industry to ensure its further development post-project into a fully industrial system that would be taken to market, where its beneficial impact on the quality of injected parts could be felt at European level. This was also empowered by the realisation of dissemination activities throughout the project and the layout of an exploitation strategy by the consortium.
Project results:
The prototype system developed during the OPTIJECT project transcends the boundary of the current state-of-the-art in monitoring the injection moulding process, as it is capable of simultaneously measuring real-time parameters that are critical for the quality of the final moulded parts and the concentrations of the different compounds. Should it reach the market, it will overcome many of the limitations of the existing practices which currently do not allow in-line measurements of all the parameters targeted in the project, especially regarding the materials.
The main scientific and technological (S&T) elements of foreground consist in:
1. calibration curves for using the OPTIJECT system. This foreground consists in the set of specific methodologies for the calibration of various plastic parameters, as well as the resulting NIR calibrations in given conditions. This 'library' of conditions results from over 1 000 experiments performed in UNEXE.
2. OPTIJECT probe design: a robust probe withstanding the high temperature and pressure in the injection moulder that can simultaneously work in VIS and NIR range with a version for transmittance and a version for reflection with the ability to work also as a receiver for transmittance mode.
3. OPTIJECT hardware and software: the hardware consist in a system processing unit (with VIS and NIR spectrometers), the above described probes and mechanical interface with the injection moulding machine. Software for processing the VIS and NIR data was also provided with intelligent features to guide the user in the improvement of part quality.
More details are given below regarding the S&T features of each element of foreground.
Parameters that were calibrated
In response to the needs identified in the stakeholder study performed in the past project in terms of factors driving defects during the injection moulding process, the consortium selected several parameters to calibrate the system. The system has been shown to extent the state-of-the-art through the identification of visual information as well as spectral information that was previously achieved in the literature as described in the previous paragraph.
The tests performed have focussed on materials parameters such as moisture, colour concentration and short term polymer degradation indicators, but also on the determination of the machine variables such as pressure, temperature and flow speed (with temperature showing a particular affinity for measurement via NIR), leading to the delivery of the specifications of the system to use and for the conditions to calibrate it.
Industrial processers can benefit from immediate quality improvement potential thanks to moisture detection. Elevated moisture levels cause premature degradation and discolouration of materials throughout processing. Most polymers are hydroscopic; plastics processors therefore dry the material before use according to manufacturer's recommendation. This drying process is seldom followed and may lead to 'wet' material being processed causing a knock on effect to processing and part quality. In addition, moulders face difficulties to identify the cause of defects as they often lack of a moisture meter in their facilities. The use of NIR to detect moisture is well documented for off-line tests but direct identification during processing at elevated temperatures has not been performed outside of the OPTIJECT project. The strong absorption of water in the IR region makes it readily discernible from other spectral features and is somewhat process independent as a consequence. The predicted moisture concentration can be that of the polymer in the hopper or of the moulded part and has a parts-per-million accuracy.
Colour monitoring thanks to OPTIJECT is believed to have the potential to become an online colour quality control and allow parts manufacturers to terminate with a non reliable reliance upon visual inspection.
While being a low cost component, the VIS part of the OPTIJECT system has been shown to be one of the most versatile measuring components due to the ability to deal with large fluctuations in moulding variables. CIE LAB values, which can be calculated after the VIS spectra, are an absolute representation of the tri-stimulus values our eyes use to identify colour projected into a colour space providing a representation of perceived colour. The software developed throughout the project detects whether the part meets the predetermined CIE LAB value. This capability of the system is of international importance as signified by acceptance of our findings to the International Polymer Processing Symposium (PPS 28).
The colorimetric analysis of molten PET and polypropylene (PP) has been performed during injection moulding respectively for red and blue colour in transmission and reflection (further details on the conditions of the tests can be found on the project website, article 'Real-time-vis-spectroscopic-monitoring-of-colour-during-injection-moulding'). Results taken during the injection moulding processing for variable amounts of masterbatch allowed calculating CIE LAB values. Online data were compared to offline data showing good agreement and therefore the potential to provide an absolute colour value for produced parts. The greatest difference between the on- and offline CIE LAB values can be found in the yellow characteristic in the B value. The difference in translucence between solid and molten parts is believed to be the source of this discrepancy, but such difference could be easily taken into account by the OPTIJECT software in case the colour specifications are based on CIELAB values measured offline.
In addition, the NIR system was used to monitor the real colour concentration in the polymer online in the melt. The online PET colour calibration has been generated using 20 repeat spectra collected at 21 different concentrations of masterbatch. Overall, the conversion of VIS spectroscopic data to the CIE LAB colour space has allowed the monitoring of moulded part colour fluctuation shot by shot and the combination with NIR data regarding the colour concentration holds the possible benefit of optimising the colour concentration to reach a target colour.
In addition to the previously presented parameters which were widely studied, the proof of feasibility of monitoring contaminants as well as of chemical blowing agents were made. The system was used during the processing of recycled PET previously used for food packaging which hast strict European Union (EU) guidelines and trace amounts of oxo-degradation additives and contaminants from food stuffs, e.g. limonene, could be detected. Based on a preliminary calibration, the opportunities for the OPTIJECT system are vast when considering the range of moulded products that are produced using injection moulding and via other processes such as extrusion.
Hardware
It should be noted that the proposed system was also prepared and tested in transmission using a second receiving probe.
Software
The OPTIJECT software is based on an artificial intelligence (AI) feature to give process set up recommendations to the moulder. At the moment, it can mainly flag single parameters which are out of operator's expectations such as colour, colour concentration, humidity, but further training is needed to provide solutions in terms of modifications of the process parameters that would lead to parts with better quality.
Thus, OPTIJECT technology can bring the following key advancements to the injection moulding process:
1. monitoring the material directly, which is completely novel in the industry
2. enabling recommendations of readjustment of the injection moulding process to improve part quality
3. most effective positioning of the NIR sensors in the nozzle of the injection moulding machine as opposed to the monitoring of single cavities - monitoring of the flow as opposed to the finished component
4. ability to rapidly measure a broad spectrum of critical parameters with good selectivity and via a single system - in terms of moisture the accuracy is as good as offline devices such as Karl Ficher (ie. ppm range)
5. ability to monitor raw materials batch-to-batch variations causing unexpected defects at the moulder
6. flexibility to shift between NIR and visible, in transmission or reflection depending of the mode that may be best suited for each set of parameter and material monitored, without handling probes.
Potential impact:
The results of the project consist of a precompetitive prototype monitoring system (system hardware, software and interface) with calibration curves for a range of parameters and a methodology highlighting how to implement and calibrate the system in new injection operations (technical service that will belong to the post-project exploitation of the system). As a conclusion of the information gathered in WP1, the general strengths, weaknesses, opportunities and threats of the system were outlined and it was shown that the OPTIJECT system, as a tool for controlling the injection process using NIR and visible spectrometers could meet a real need in the plastic converting industries to assist them in supplying ends users with parts of better quality and reducing scrap. It will also involve the supply and value chain whereby machine and sensors manufacturers will play a key role. As previously mentioned, the SMEs have jointly applied for a demonstration activity to lead the OPTIJECT system to the market.
In terms of economic aspects, one of the key economic requisite from the SMEs regarding the OPTIJECT system was that it should be low cost to open a wider market by making it commercially attractive to prospective clients. This has been one of the key challenges met during the up-scaling of the prototype and need further improvements to optimise the cost of the system as the cheapest version did not allow working online in the injection machine and only a version of greater price worked in those conditions.
There are numerous socioeconomic impacts that will be derived from the results of this OPTIJECT research project. Central to the expected socioeconomic impacts, is the boosting of the competitiveness of companies operating along the injection moulding chain- from raw materials and sensors suppliers to parts moulders - by improving the quality of European injection moulded plastic products and decreasing their scrap. The OPTIJECT system will be introduced in a very large market. Indeed there are well over 13 000 SME injection moulders in EU27 and an average production of 200 million of plastic components per company with an average annual growth of about 5 %. The total revenues are estimated at approximately EUR 280 billion. Directly and indirectly, 1.5 million jobs can be accredited to the European injection moulding industry. Making this industry more competitive will have impacts for the European economy, growth and jobs.
Furthermore, by virtue of their technological features, the OPTIJECT systems could improve working conditions by doing away with time consuming and chemical-based sampling techniques.
There are also environmental benefits associated with the OPTIJECT systems. Quality monitoring of materials will prevent batches unfit for injection from entering the processing chain. This will prevent the waste of energy, raw materials and resources. Furthermore, OPTIJECT will assist European producers to adhere to the various European Commission (EC) guidelines that exist in the plastic industry, such as, for example, for processed items intended for food packaging and other even more stringent guidelines when it comes to medical applications. Moreover, by reducing the energy costs involved in plastics processing, as well as raw materials consumption and waste (due to lower scrap generation), the results of this project will also contribute to assisting the EU in meeting its carbon dioxide (CO2) cutting objectives under the Kyoto protocol.
In addition, further opportunities were disclosed during the OPTIJECT project. In particular, while the system focussed on the validation of the injection moulding process, the part quality depends not only on the injection process itself, but also on the raw material quality and consistency. Therefore, the use of the similar system at the manufacturer of pellets would allow a global management of the quality parameters with a similar system throughout the supply and value chain. The development is reduced and less challenging and will be undertaken here opening up potential in an additional 30 % market share of the plastic converting techniques.
Main dissemination and exploitation activities:
The expected economic and technical benefits of the OPTIJECT technology have served as basis for the dissemination and exploitation activities. As previously mentioned, OPTIJECT is expected to have a significant impact on the quality management in the plastic industry, whereby injection moulding is a very important conversion technique.
The project included a series of other activities targeted at maximising the potential for the OPTIJECT system.
A great importance was given to the management of the intellectual property rights generated and in agreement to the dissemination of non-confidential information throughout the project. A patent review was carried out and concluded that OPTIJECT does not infringe upon existing protected intellectual property (IP). Although the project developed significant know-how exploitable by the SMEs such as specific designs for the probes, the calibration curves for their injection processes and material parameters of interest, their novelty did not allow filling a patent.
The plastic market in which OPTIJECT will be introduced is a highly promising one, following an upward trend in terms of volume produced but also of quality aspirations and scrap reduction. In addition, to date none of the currently available solutions offer the technological innovations or performance of OPTIJECT. The foreground generated by the OPTIJECT project is broad giving rise to the potential exploitation opportunities for each member of the consortium. A preliminary business plan as well as post project development work have been laid out benefitting from the synergistic role of the partners covering the whole supply and value chain. Based on the results of the project, the SMEs have access to a system that they can exploit on the market place. They jointly applied for a demonstration activity project in order to bridge the gap to commercialisation. Findings from the industry questionnaire-based survey and the industry consultations have revealed a willingness to invest in such a system in view of tangible technical and envisaged economic benefits. These figures predict a fast return on investment for OPTIJECT system.
Non-confidential information on the results and principles of the OPTIJECT technology has been disseminated throughout the project. The public website of the project (see http://www.optiject.eu online) informs the public and interested parties on the basics of the technology, latest news and the progress of the project insofar as it can be disseminated without threatening the proper protection of the developed intellectual property. Various press releases were circulated in professional plastic technology magazines and websites, raising the awareness of OPTIJECT both in industry and in the public domain and generating positive feedback. The project partners also took part in number of trade fairs. In total over 50 dissemination efforts were realised with contribution from the whole consortium.
Project website: http://www.optiject.eu
Contact details for further information:
Dr Ing. Elodie Bugnicourt
IRIS Innovació i Recerca Industrial i Sostenible, Spain
ebugnicourt@iris.cat
