CORDIS - EU research results
CORDIS

Automated inspection for sintered parts by non-destructive techniques for improved quality in production

Final Report Summary - AUTOINSPECT (Automated inspection for sintered parts by non-destructive techniques for improved quality in production)

Executive Summary:
Powder Metallurgy (PM) parts are typically intricate and complex in shape, produced in near net shape by compaction of powders into the required geometry (i.e. green state), followed by sintering of the compacts for consolidation (i.e. sintered state), where particles are bonded upon heating.

There is a growing need for efficient use of materials and processes, while delivering improved quality across the whole of manufacturing in Europe, to remain competitive with cheap labour areas of the globe. With its high material utilisation and low energy requirement, PM is currently one of the most effective forms of manufacturing, and has the potential to give significant competitive advantage to European manufacturers. The PM process by nature is suited to high volume production. Therefore, any flaws/defects in the parts can have a significant impact on the production output, for example loss of material and efficiency, as well as potential failures in the end use.

Sintered components can suffer from porosity (hence density variations), cracks and impurities. Such defects negatively affect the mechanical properties of the part hence its performance. As they are produced in their thousands in a production environment, if the defected parts and the cause(s) are not determined at an early stage, the whole production output can be rejected, if considered unsuitable for the intended application.

Increasingly, PM parts are used in the automotive, aerospace and defence industries where there is 100% inspection required. One of the driving developmental forces for PM parts within the automotive industry is for fuel economy achieved through weight saving targets, and this

There is an increasing push in the automotive industry for fuel economy that will create weight savings targets,
which require higher performance requirements and better materials / alloys for sintered parts [3]. PM parts will need to be as defect free as possible for improved integrity to meet such requirements. Common defects i.e. porosity (hence density variations), cracks, impurities etc. significantly affect the mechanical properties of the part and hence, its performance.

Currently there are no true “in-line” NDT techniques that can inspect parts before or after the sintering process. In order to reduce the inspection time, manufacturers of PM parts have no choice but to perform spot testing on batches of parts from a given production run. However, this means that defective parts are even more likely to reach the customer and, due to the high expectations of both primary manufacturers and end consumers, this represents a significant obstacle to the wider use of PM components. The current end-of-line inspection techniques used are not suited for high volume 100% manufactured part inspection because they involve manual or part manual processes, and are therefore slow and require subjective interpretation by an operator.

AutoInspect system fulfils the need for an automated inspection by non-destructive means, to identify and separate the good and bad components during production. The identification of good/bad components needs to be achieved as early as possible, without having to resort to manual destructive examinations that can have a negative impact on the production flow and output. Crucially, any faulty part that is overlooked may cause more problems later on, such as unexpected, premature failures in application.

The intention of the AutoInspect project was to make a number of technological advances combined into a prototype that will take the PM parts manufacturing sector into an area, where the risk of defective PM parts is negligible leading to maximum part integrity. The objectives of the AutoInspect project (www.AutoInspectProject.eu) were to design and develop an in-situ digital radiographic system that can reliably inspect PM components, both in the green and sintered states. This technique would allow fast inspection and application of image processing for the detection of small cracks, flaws and density variations in-situ. The AutoInspect system prototype has met its objectives. The inspection system is fully automated and can inspect and automatically detect defects in real time, on the green and sintered components.

Field trials have been successfully conducted using the AutoInspect prototype and the systems capability has been demonstrated. Field trials have proven that the technology readiness level of the AutoInspect system is from 4-6. With further development needed for a purpose built lead enclosure and further integration, the partners estimate a further 9 to 18 months development requirement based on potential customers and their requirements.
Project Context and Objectives:
Green and sintered components obtained by the powder metallurgy process are employed in several industry sectors. Powder metallurgy is a process whereby typically intricate, complex shaped parts are produced in near net shape by compaction of powders into a geometry (referred to as green state) followed by sintering of the compacts for consolidation, where particles are bonded upon heating.

Due to a growing need for efficient use of materials and processes coupled with improved quality across the whole of manufacturing in Europe, powder metallurgy currently provides one of the most effective forms of manufacturing with its high material utilisation and low energy requirements.

Sintered components typically suffer from porosity (hence density variations), cracks and impurities. Such defects negatively affect the mechanical properties of the part hence its performance. As they are produced in their thousands in a production environment, if the defected parts and the causes are not determined at an early stage, the whole production output can be rejected, if considered unsuitable for the intended application. The need for defect free manufacturing particularly in the automotive sector has motivated manufacturers to seek reliable cost effective inspection methods for eliminating the defect output in production.

Currently, there are no in-line NDT techniques that can inspect parts immediately after sintering. This is an issue because it has been reported that even with a 100% end-of-line inspection (which is a time consuming process) typically 6-8% of production parts are scrapped and there are still field failures returned by customers [Private Partner Communication]. To reduce the inspection time usually manufacturers of PM parts have no choice but to perform spot testing on batches of parts from a given production run. However, this means that defective parts are even more likely to reach the customer and due to the high expectations of both primary manufacturers and end consumers, defects cannot be tolerated even at low levels in million piece quantities. Increasingly, PM parts are used in the aerospace and defence industries and there is 100% inspection required. The current end-of-line inspection techniques used are not suited for high volume 100% manufactured part inspection because they involve manual or part manual processes, are therefore slow and require subjective interpretation by an operator.

Moreover, if any flaws are critical and this is not known by the manufacturer, the manufacturer may for example end up with a poor batch of parts in their thousands returned due to poor quality or simply as a faulty batch. Overlooked faulty parts may cause problems in service such as unexpected premature failures in application. Depending on the component and the criticality of the application, this can have drastic consequences such as accidents. Failure of moving parts of engine components for example valve seats, gears etc. in automobiles, although repairable, can be complicated, and in huge numbers damaging for the manufacturers reputation. Some rotating components (turbines) in aeroplanes, where service conditions and criteria are even more stringent, present an even greater requirement for defect free production.

The recall, of eight million cars by Toyota since winter 2009 to date, highlights the importance of quality control and the need for automated inspection in mass production environments. Products need to be inspected before they are released to the public, to avoid faults being discovered later which can cause failures and accidents that can be disastrous. The recall of eight million cars was to fix vehicle/engine components like accelerator pedals, brakes etc. with dangerous defects. Sales of eight popular models had been suspended and the company’s reputation was significantly damaged as reported by the Financial Times, February 6 2010. This signalled a need for new inspection equipment that was automated, could inspect with a high throughput and could reliably sentence good and bad components. AutoInspect, a digital radiography (DR) system was a very timely approach in the pursuit of an improved quality control in an increasingly mass productive industrial environment.

Digital radiographic (DR) systems are well developed in the medical and dentistry markets and have been for several years. DR systems are more efficient than conventional X-ray films since the exposure time is significantly reduced leading to a faster inspection. Also, digital radiography can be conducted in enclosed lead shielded cabinets so is safer for operators. When performing radiography with films it is always necessary to be able to access the films for development, and typically this kind of inspection takes place in a walk in radiography bay. Although industrial digital radiographic systems are available, their take up has been limited to a few specialist inspection applications and prior to AutoInspect no radiography inspection system has been used to inspect PM parts in-line in the PM production process.

AutoInspect equipment is designed to be an automated system that allows real time imaging and automatic defect detection for different PM components.

The proposed concept was a new approach to the problem of performing 100% inspection of PM components which required significant research effort by the RTDs, way beyond the knowledge of the SME consortium or end-users. The end result was a system which is beyond the state-of-the-art.

The main project objectives were;

1) To create a digital radiography inspection technique able to automatically inspect PM parts (green and sintered state) in seconds.
2) To embed time delay integration (TDI) linear X-ray detectors that allow the supply conveyor to run continuously, while a row of parts is simultaneously scanned. The TDI technique creates very low noise X-ray images with resolution of up to 10 um pixel size, depending on X-ray setup magnification.
3) To develop a dedicated defect detection software for fast and automatic sentencing of the PM components. A component database storing details of known, good, components for comparison will be included as part of the software development.
4) To develop an optimum mechanical arrangement taking into account the conveyor belt throughput speed, repeatability, accuracy, position with respect to the source and detector, and X-ray safety.

The AutoInspect project opens up a completely new inspection market place for SME digital radiography inspection equipment suppliers. Uptake of such an inspection system is likely to be commercialized if the system is low cost, offers fast inspection and component sentencing and finally, is of use in-line, in real time in a production environment.

Project Results:
The AUTOINSPECT project has taken existing radiography technologies used in Industrial NDT, Medical and
dentistry markets and further developed their capability into new PM component inspection application where there is a genuine requirement based on safety, environmental, economic issues.

The proposed S&T objectives were to develop a new real time digital radiography inspection system for the detection of defects in PM components. The inspection system is to be automated and perform each component inspection in a few seconds. The techniques, systems and technology needed to realise the project included:

1) Development of an appropriate stabilised energy X-ray source system fully controllable by computer. Technical development work surveyed a number of different sources to check whether a mini or micro focus source will be acceptable. The connected X-ray source parameter settings like the tube current, voltage, etc. can be adjusted using the user interface on the PC. Graphical LEDs are displayed on the software to show the status of the X-ray source cooling, vacuum parameters.
Also, the source controller electronics and software module in order to control and communicate with the X-ray source, via the Ethernet link from a PC has been developed and integrated into the main LabVIEW software user interface. The results for this are the software solution for controlling the X-ray source to facilitate its remote operation.

2) Development of a linear detector array of small physical size and suited to imaging typical PM components with high resolution. This type of detectors can be found in the dentistry sectors. Various preliminary tests were conducted by the lead project partner AP2K in their laboratory and a most suitable Time Delay Integration (TDI) X-ray linear detector was tested and implemented, for image acquisition in the AutoInspect project (TDI is a technology in which a detector produces a continuous video image of a moving object by means of a stack of linear arrays). The technical result was a custom built Teledyne Dalsa TDI detector with a scintillator capable of giving 6-7 lpm (in 2x2 binning mode) resolution at 160 keV and that would allow scanning at high speed.

3) Development of novel radiography setup. After detector and source procurement, investigations were carried out to establish the optimised radiography inspection setup. Control and correction software was written using LabVIEW for the selected detector. Testing was carried out using the development software which allows control of the detectors together with fast image acquisition.
Moreover, the AutoInspect system is designed to use two detectors to acquire two orthogonal images of the parts, equally visualizing parts sides for normally hidden or overlapped defects. This foreground IP is currently being investigated as a patent application.

4) Development of a suitable purpose-built computer software created using LabVIEW.
The main program interface operates the entire system automatically for the acquisition and storing of the radiographic images. The graphical user interface (GUI) is designed so as to allow the user to select and change a variety of parameters. Different tabs on the interface provide the functionality to perform different operations such as initializing the system, setting up the system parameters (conveyor speed, magnification, TDI rate of acquisition) and set up the X-ray source parameters. The images obtained using this interface are stored in a user-specified folder.
5) Separate software was also developed using LabVIEW’s vision library, to enhance the images using a number of algorithms, detect any anomalies or defects and sentence the PM part as either good or bad. When executed on the AutoInspect main control computer, component sentencing can be performed in a few seconds. The results for this are the software solution.

6) Development of a component feeding mechanism. A conveyor belt was specifically designed for feeding samples to the AutoInspect DR system. The conveyor is a motorised belt which uses a VF-S11 inverter to control the operation parameters like speed, forward, reverse etc. The conveyor can be controlled remotely and has been integrated in the user interface software for control. The results for this are the hardware (control electronics) and software solution.

7) Integration of the various components making the AutoInspect system. The project started with a computer aided design (CAD) model in order to create an optimum and accurate design of the steel frame which would hold the X-ray source, the collimator and the two TDI detectors taking into account the detector movement. Although a purpose built radiography shielded enclosure was not created, CAD manufacture drawings have been created for a shielded enclosure to increase the technology readiness level.

After the system was mechanically integrated, extensive calibration work was conducted in order to ensure that the AutoInspect system operation was optimised taking into account the

• Magnification: Due to the fact that the radiographic component’s frame is independent of the conveyor, and therefore their relative heights are independent, in the prototype’s current state the magnification calibration had to be carried out each time the prototype was moved. This involved measuring the no. of pixels in the width of the imaged sample and knowledge of the detector pixel width (54μm in 2x2 binning mode) using which, the width of the projected sample image was calculated. The ratio of the width of the projected image to the actual width of the samples provided the magnification of the system.
• Synchronised movement of the conveyor belt with respect to the TDI detector: A delay in software needed to be added before the image acquisition command to ensure that images captured fully encompassed the sample, and that automated runs are repeatable. The conveyor being designed to run at different speeds, consequently the delay time was a function of the conveyor speed. Hence, a relationship between the delay time and conveyor speed needed to be experimentally derived for both the detectors.
• Line rate synchronisation calibration: One of the results needed from the AutoInspect system is the quality of the produced TDI image generated by the synchronisation between the TDI line rate and the conveyor speed. This was to ensure that the when the TDI line rate is too fast or slow, the image didn’t appear squashed or stretched because of multiple parts of the sample appearing on one pixel of the image or that a single part of the sample appeared over multiple pixels of the image. An equation was deduced to relate line rate, magnification and conveyor speed to develop the automation software to set the correct TDI line speed based on the user-inputted (or loaded) magnification and conveyor speed, in order to obtain the best DR images.

The scientific and technology result is a digital radiography based inspection prototype which is ready to automatically obtain radiographic images of supplied PM parts, sentence any defective PM components in an accurate and repeatable manner.

The AutoInspect system is able to scan for defects (such as cracks and pores down to sub mm scale) and perform the inspection in seconds.

Potential Impact:
* Socio-economic impact:

The AutoInspect project has made significant technological progress in providing a real time automatic DR inspection solution for the detection of defects in PM parts. The successful outcome to the project will benefit the EU potential business on a global scale with a technology that could benefit every PM component manufacturer and end user.

Commercialisation of the AutoInspect project will allow greater use of sintering for a variety of batch sizes suitable to the SME manufacturing area and this will result in more efficient production, reduced waste, less rejects, and minimal failures in a range of products.

Depending on the criticality of application, an overlooked faulty component being used in a larger device may result in serious failures This may result in high concern over the quality of the PM parts and the manufacturer’s reputation in the relevant PM or other industry sector where the parts are used.

For SME’s participating in the project, there is a significant impact since the automation in inspection aspect of the manufacturing line will increase productivity, reliability, provide safe working conditions and bring the manufacturing costs down. PM parts are increasingly used in the automotive, aerospace and defence industries and hence, the PM manufacture business forms an important part of the European high tech industries. Moreover, the project participants will find increased business opportunities based on their pooled knowledge for mutually beneficial co-operation on products for other applications and market sectors like the food industry.

AutoInspect will enable early and automatic detection in real time of defects like cracks, flaws, density variation etc. This will decrease reject rates of PM parts and increase quality production whilst providing the EU with a competitive edge over other manufacturers outside Europe.

It is expected that, after successful commercialisation of the AutoInspect prototype, and due to its unique design and fast automatic real time detection, the system has the potential to become the industry standard. As such, it would be a standard requirement at goods inwards, both at the component distributor and the company receiving the components. With many European companies subcontracting parts manufactured outside Europe, sales are also expected for manufacturing sites outside Europe.

* Dissemination activities:

Exploitation and dissemination: All AutoInspect partners have been active in dissemination. This included dissemination of project activities and results through scientific journal and conference publications. Also, dissemination activities were conducted via news items, magazine articles and seminars. The AutoInspect website hosted a blog which was regularly updated, and reported on global news items related to the PM industry, as well as news specific to the AutoInspect project.
A video about the AutoInspect project has been made and uploaded to the project website and on YouTube. A final plan for use and dissemination has been developed giving a clear indication of past and future dissemination and exploitation activities.

Of significant note are the following activities:

Conference presentations: Scientific work from the AutoInspect project was presented as a poster and an oral presentation at the EPMA 2012 conference, held at Switzerland in September 2012. This resulted in two conference proceedings and a Journal paper published in the Powder Metallurgy Journal.

Further, AutoInspect findings were also presented at the EPMA 2013 conference held in Sweden which has resulted in three conference proceedings.

Flyers: Two flyers were produced at different stages in the project for distribution. The first flyer was distributed at the EPMA 2012 conference and the second flyer was distributed at the AutoInspect final demo day in December 2013.

News Article: In November 2013, TWI advertised a news story on IPMD.net E-news highlighting the features of the AutoInspect system and also welcoming people to attend the AutoInspect demonstration day.
Demonstration: A demonstration of the finished AutoInspect inspection system was given on 12 December 2013 at TWI in Port Talbot, UK. All the consortium partners were in attendance. The AutoInspect inspection system is currently available for demonstration at TWI, Port Talbot.

Final Video: TWI has scripted and shot a professional video of the AutoInspect system. This was recorded a day before the final demonstration of the AutoInspect system in December 2013. The collected video footage has subsequently been edited in to a short information film. This film will be uploaded to YouTube and presented on the AutoInspect website blog and is available as a DVD for interested parties. However at the time of writing the video has not been made public. The coordinator is currently investigating applying for a patent and as such public material about the AutoInspect project is currently being put on hold so as not to affect the patent application.

* Exploitation of the results

The novel radiography inspection setup used in AutoInspect is currently the subject of a patent investigation/application.

The AutoInspect project will deliver an integrated inspection solution to meet the requirements of the EU PM parts industry. Commercialisation of the AutoInspect inspection system as a whole is expected to be a joint effort between the SMEs. Lead AP2K will be the main exploiter of the AutoInspect system as a whole. They are already currently making an initial patent application. AP2K will lead and take charge of manufacturing and subcontracting.
Licensing of the technology to third parties outside the consortium will be considered by the Exploitation Manager and approved by the PSC as necessary.

The partners believe that the technology readiness level of the AutoInspect system is in the range 4-6. Further development for the system will include a purpose built lead enclosure and further integration. The partners estimate a further 9 to 18 months development is required depending on possible customer requirements. The project partners are currently examining further funding possibilities.

List of Websites:
A project website was set up at the start of the project to facilitate and act as a communication tool for the consortium. The website consists of two main areas: one accessible to the public, and one only accessible by the members of the consortium. The project website address is: http://www.AutoInspectProject.eu The website is used also for dissemination of the project results.

For general enquires regarding the AutoInspect project please contact Mihai.Iovea@accent.ro
Accent Pro 2000 s.r.l.
Nerva Traian 1, Bl K6, Sc 1, Ap 26, Bucharest S3, 031041 – ROMANIA
Tel/Fax/Answering machine: 0040 745 182660/ 0040 213 204759
Web Site: http://www.accent.ro
E-mails: office@accent.ro ; miovea@pcnet.ro

Contact details according to the specific technologies developed in AutoInspect are listed below:

Topic: The AutoInspect system, Image processing and analysis software
Mihai Iovea
Accent Pro 2000 s.r.l.
Nerva Traian 1, Bl K6, Sc 1, Ap 26, Bucharest S3, 031041 – ROMANIA
Tel/Fax: 0040 745 182660/ 0040 213 204759
Website: http://www.accent.ro
E-mails: office@accent.ro ; miovea@pcnet.ro

Topic: Digital radiography and the graphical user interface software
Ian Nicholson
TWI Ltd,
Granta Park, Great Abington, Cambridge, CB21 6AL, UK.
Tel: +44 (0)1639 873100
Email: ian.nicholson@twi.co.uk
Website: www.twi.co.uk