Final ReportSummary - NOZZLEINSPECT (Autonomous Robot for an Automatic Inspection of Nozzle Welds in Nuclear Environment)
Regular, in-service inspection is important to verify the integrity of welded nozzle sections in nuclear and other safety critical facilities. Nozzle sections made from austenitic and ferritic steels can be susceptible to rapid crack growth due to thermal fatigue and stress corrosion. Early detection of cracks is therefore essential to ensure the continued safe operation of the facility in question. The nozzles form a critical part of the nuclear reactor and therefore their structural integrity is of the uttermost importance to the nuclear plant operators. The nozzles are located at the vicinity of the nuclear reactor where there is a high ionising radiation. The inspection personnel have very limited time to enter the nuclear reactor area to set up and calibrate the inspection equipment and often more than two personnel need to enter the reactor in order to minimise the individual radiation dose. If the exposure to the radiation of the inspection personnel exceeds the annual maximum allowable dose, then the employee will not be allowed to carry out any inspections for the remaining of the year. In that case, the employee's cost is passed from the inspection provider to the nuclear plant operator. Therefore, apart from health implications there are also financial implications associated with the inspection of complex nozzle welds in the high radiation areas.
In order to reduce the time and cost of such inspections, there is a need to develop a system capable of performing a full inspection of nozzles without the need to frequently change ultrasonic probes. The aim of NOZZLEINSPECT project is to design a semi-autonomous and automatic inspection robotic system that is able to reduce the inspection and setup times, improve defect detectability and sizing and reduce human intervention. This will reduce workforce radiation exposure and reduce the requirement for complex robotic manipulation and consequently, reduce the size and cost of the robotic deployment system. This is achieved through the use of advanced 2D phased array probe that provides 3D steering capabilities and therefore some of the manipulation required can be achieved through the setup of the ultrasonic probe via specialised phased array software.
Project context and objectives:
The main goal of the NOZZLEINSPECT project is to improve the reliability of the inspection of the welded nozzle sections in nuclear and other safety critical facilities. The scanner and manipulator assembly will carry a new and novel adaptable phased array probe to allow a full inspection of nozzle weld areas, and an advanced navigation system that will follow the weld around the nozzle. The presence of defects in these parts could lead to component failure.
NOZZLEINSPECT will develop an advanced scanning / manipulator robotic system with a new generation of phased array technology providing advantages for the inspection of nozzles in nuclear power plants including:
1. inspection of nozzles in 25 % of the time currently taken, thus reducing the extent of operator exposure to radiation;
2. navigation of complex weld profiles;
3. inspection of nozzle welds without the need to use a large number of probes, calibrations and procedures;
4. improvements in the detection and sizing of defects in the challenging and complex nozzle welds.
Nozzle inspections require a variety of inspection techniques and equipment. The process necessitates that the plant be shut down for several weeks, which leads to lost production and profitability. Although automatic or semi-automatic inspection equipment has been introduced for many applications, operators are nevertheless still exposed to radiation during the equipment set-up and during certain operational activities.
Inspections are currently undertaken using a number of conventional ultrasonic probes. Since numerous probes are required to enable full inspections of all the nozzle weld areas, and each change of probe demands a recalibration to be completed, defect detection and sizing capabilities are not optimised. Furthermore, as the current probes have no beam steering capability, a large and expensive robot is required to provide accurate manipulation.
In order to reduce the time and cost of such inspections there is an urgent need to develop a system capable of performing a complete inspection of nozzles without the need to change probes. Such a system will offer the following clear benefits, including:
- faster inspection times;
- improved defect detectability and sizing;
- reduced human intervention which will reduce workforce radiation exposure;
- reduced requirement for robotic manipulation and, consequently, reduced size and cost of robotic deployment systems.
The objective of this project is therefore to develop a system that will use a semi-autonomous scanner / manipulator system for an automated inspection of nozzle welds. The robot will carry a novel adaptable matrix phased array probe which, combined with the 3D beam steering capability, will enable the entire volume of the weld to be inspected in a single operation.
NOZZLEINSPECT is a collaborative project consisting of the following organisations: TWI Ltd, Optel, Vermon, PeakNDT, Iberdrola Generation, Phoenix, KTU and Cereteth. The project is coordinated and managed by TWI Ltd and is partly funded by the EC under the Research for the benefits of SMEs project reference No. FP7-SME-2008-1-GA-232523.
Project results:
Provision of samples
The project consortium used a very detailed specification to understand and agree the approach to be adopted for the robotic design and this in turn needed some initial kinematic simulation to be carried out. It was apparent that the structural mock-up needed to be large but light enough to be portable and mounted on wheels and at the same time it should be capable of being dismantled for shipment to the project partners. In addition, the mock up must not topple over when the robot is being manoeuvred and must be stiff enough to resist the clamping and other reaction forces.
This nozzle geometric mock up was designed and fabricated for the robotic scanning system and control algorithms development and the final testing of the complete system and its functionalities.
The design requirements of the nozzle mock-up that will be used for the development and testing of the scanner and probe manipulator assembly were as follows:
The mock up consists of two main components, which are the pipe nozzle and back plate fabrication and the stand.
Pipe, nozzle and back plate fabrication requirements:
- one single steel fabrication preferred to achieve the desired tolerances;
- machine steel pipe > 5 mm wall thickness;
- pipe wall thickness suitable for induced bending moments based in assumptions;
- welded end flanges to allow connection to nozzle fabrication and stand;
- concentricity/roundness +/- 1.0 mm;
- surface finish< 3.2 µm or better on both 345 mm and 605 mm pipe outer diameter;
- replicates outer surface profile of nozzle to allow functionality of manipulator to be tested;
- fabricated to nominal fabrication tolerances as stated in BS22768;
- rolled plate to form curvature of vessel reactor and bolts to nozzle fabrication;
- replicates outer reactor radius to an area of 1.6 m x 1.6 m;
- vessel inside radius 2785 mm;
- provides a structural support for the nozzle feature although has additional stiffeners to inhibit flexure.
Mock up stand fabrication:
- fabricated steel frame structure;
- lockable castor wheels to aid moving around;
- stable arrangement to allow manipulator functional tests to be performed;
- allows close access to the pipe.
Another research aspect of the project was the ultrasonic technique development for the inspection of the nozzle to vessel weld. Due to the nozzle to vessel geometry, the weld profile and volume changes with circumferential position around the nozzle to vessel weld. Therefore, three welded reference samples with known defects that represent the weld profile at 0º, 45º and 90º were designed and manufactured. These samples were used to develop the ultrasonic inspection technique. A total of 18 defects were introduced in the nozzle to vessel weld that provides a wide range of sizes, skew and tilt angles. The defects in the nozzle to vessel weld will be inner and outer surface breaking defects as well as internal defects in order to assess the full volume coverage of the developed ultrasonic technique. The defects have different sizes, skew and tilt angles to represent the type of defects that are expected in the nozzle weld structure. All these different parameters assisted the development of the inspection technique and evaluation of the detectability and sensitivity achieved as well as understanding the effect of the defect parameters to the detectability.
Three ultrasonic validation welded reference test samples with introduced defects were required to be manufactured for the ultrasonic technique development purposes. As part of the project, a trial sample to demonstrate the dimensional accuracy of the flaws implanted by electrode discharge machining (EDM) was prepared.
In order to qualify the flaw implant procedure, a test coupon was manufacture applying the same welding and machining processes as would subsequently have been used to manufacture the ultrasonic reference samples. The test coupon was obtain by butt-welding two mild steel plates 300 mm long x 150 mm wide x 50 mm thick, using submerged arc welding (SAW) process and using the exact welding parameters employed during the fabrication of the nozzle.
In order to obtain embedded flaws, the weld bevel was filled up to mid-thickness, four notches with different height and length were then implanted by EDM, using skew and tilt angles typical of the UT validation blocks.
The ultrasonic reference samples included surface breaking notches at the weld root and cap, on the base metal (nozzle inner radius), as well as embedded notches at mid-thickness. Considering that EDM notching is a fully automatic process, it was not considered necessary to qualify the procedure used to obtain the surface-breaking notches. On the other hand, embedded notches are obtained by EDM and subsequently buried by manual metal arc welding (MMA), therefore, the qualification activities focused on these type of flaws.
For the final demonstration of the developed automatic inspection system, the consortium has secured access to the N5 nozzle mock up that was manufactured for the qualification of new inspection techniques and the training of operators. The full nozzle mock-up fully represents the geometry of the actual component and has a number of different defects introduced at the nozzle weld and the inner radius.
Design of an adaptable phased array probe
In order to design and manufacture a suitable phased array probe for the inspection of nozzle weld, a number of different designs on the element configurations were simulated to find the optimum design. The pulser-receiver that will be used for the inspection is capable of pulsing 128 elements simultaneously and therefore all the phased array designs are working around this parameter and limited to 128 elements.
In order to optimise the geometry of 2D phased arrays, various configurations were considered including 2D matrix, 2D annular segmented and 2D matrix annular phased arrays. The beam profiles generated by these arrays were simulated using CIVA 9.2 and the results were compared with respect to beam amplitude, focal depth, focal zone, etc. An algorithm using an apodisation technique was used in order to reduce the level of side lobes that result to ultrasonic energy loses of the primary beam and potential unwanted echoes. An ultrasonic field modelling software called SimulUS developed by the project partner that was used to apply the apodisation function to the 2D phased arrays and the beam profiles generated by various 2D phased arrays were simulated. Also this software was used to compare the ultrasonic field generated by annular matrix arrays and matrix arrays. One of the novelties and innovation of this project is to use flexible coupling material (instead of using wedge) which allows the 2D phased arrays to be placed on both flat and curved surfaces, which are both present during the nozzle to vessel weld inspection. Therefore, in order to investigate the effect of using the flexible coupling material and the wedge, their corresponding ultrasonic field were simulated and compared.
The matrix-annular phased arrays are constructed using square or rectangular elements in a matrix configuration. By removing a number of elements from the corners of the matrix phased arrays or manually setting, these elements in the matrix arrays as not functioning or sending and receiving any ultrasonic waves. That way an annular shape array can be obtained using the elements square or rectangular shape elements. Therefore, when giving the geometry details of the matrix annular phased arrays the geometry of matrix array is given as well. The reason using smaller gap is to obtain larger elements with reduced probe footprint and then to extract more ultrasonic energy from the array and therefore introduce more ultrasonic energy into the material. The final probe size is 31.9 mm by 29.9 mm, and the element pitch in X and Y directions is more than one wavelength, which are 2 mm and 3 mm, respectively. The reason for using this configuration is that even though the large pitch size results in the generation of grating lobes, if the level of grating lobes is acceptable this configuration is a good option since it is able to provide larger amplitude of the ultrasonic beam.
Overall, the phased array probe configuration selected as the best configuration among the 2D phased arrays due to the following reasons:
1. It gives the highest amplitude of ultrasonic beam when steering the beam at 40 in YZ plane. Note that the linear array (unable to steer beam in XS plane) and AS2 (difficult to implement in reality) are not considered.
2. The level of the grating lobes generated by MA3 is small and acceptable.
3. The focal depths of -3 dB and -6 dB are the largest and the focal zone is able to reach the bottom of the inspected component.
A mechanical probe holder has been designed to house the phased array probe the selected probe. The probe holder incorporates a flexible membrane such that the surface in front of the phased array probe adapts to the shape of the surface of the vessel or nozzle being inspected. The probe holder also incorporates specific features to assist with both the ultrasonic and mechanical performance when connected to the robot scanner.
Having developed a phased array probe specific to the nozzle weld inspection task a method was required for holding the probe at a known stand off from the surface at a prescribed angle and a method for coupling the ultrasound to the surface through a liquid or solid medium. Conventional inspections methods often use solid wedge transducers manufactured from plastics such as Perspex. In this instance, solid wedges were not an option. This was due to the fact that the footprint of the developed phased array probe was relatively large (31.9 x 29.9 mm) and that the profile of the vessel surface would vary from flat to curved as the transducer was rotated around the nozzle. It was therefore necessary to develop an adaptable probe holder that would conform to the varying surface whilst still maintaining transducer orientation relative to the surface. The solution was to design a fluid-filled probe holder with a flexible front that would provide a bath of water between the transducer and the vessel surface.
The probe holder incorporated four water hose fittings that are attached to two separate delivery systems. Two of the hose fittings go into the body in order to fill the cavity with water for immersion couplant. One should be used as an inlet and the other a bleed hole (outlet) to ensure the body is filled. The other two hose fittings go straight though the body to two spray nozzles that wet the surface of the vessel ahead of the probe holder as it is scanned. This couples the ultrasound from the membrane to the vessel.
Two polyurethane membranes with different stiffness were manufactured and evaluated. A lower stiffness membrane (shore hardness 42) was found to be very flexible but also exhibited high friction when scanning on the surface. A higher stiffness membrane (shore hardness 60) had a much lower friction coefficient and was determined to be of adequate flexibility for the surface variation to be encountered on the vessel. The friction coefficient was measured to be around 0.9 for dry sliding contact. However when the surface was wetted through the spray nozzles the sliding of the polyurethane membrane improved significantly and the friction coefficient was measured to be around 0.3 for wet sliding contact.
Phased array ultrasonic technique development
As part of the ultrasonic technique development, the signal responses from 42 defects were simulated using an ultrasonic modelling software package. The orientation and location of each defect is varied: the angle is skewed from 0° to 10° and tilted from -15° to 15°; the position is in the middle or bottom of the vessel welding zone. The modelling software is capable of displaying the simulated inspection phased array results in sectorial scan, B-scan and C-scan of the defect. The half-skip method was used to inspect the defect located in the bottom as well as all the defects located in the nozzle weld. The stainless steel cladding present at the inner surface of the vessel, and at a distance of 50mm from the weld area, will make the application of the full-skip method difficult to be implemented. The full skip method relies on the ultrasonic beam to reflect on the component backwall and then reflect at the same angle as the incidence angle to the area of interest in the nozzle weld. However, due to the presence of the stainless steel cladding, the full skip method cannot be used because the ultrasonic beam will be distorted, and mode converted at the steel-stainless steel interface. Therefore, only the half-skip method is applicable where reliance is placed of direct reflected ultrasonic signals from the defects. The maximum signal response directly from each defect located in the middle was simulated and used to determine the optimal location.
For each defect, in order to receive the maximum amplitude of the signal, the location of the probe and the beam angle in relation to the probe were determined to identify the location of the probe (X-distance and Z-distance) and the beam angle (beam steer angle and beam skew angle).
Following the completion of the simulation task, the 2D matrix annular probe mounted into a flexible probe holder filled with water was used to test the manufactured ultrasonic reference samples with known defects. The ultrasonic beam was transmitted from the probe into the water and then passed through the flexible polyurethane membrane. The probe holder was placed on the vessel of the nozzle. The gel coupling was used between the membrane and the sample in the laboratory development work, and water coupling will be applied to the inspection of the real sample through the use of miniature spray nozzles.
The experimental results from the developed phased array probe and flexible probe holder shown that the introduction of 3D steering abilities has a significant effect to the detection of defects misorientated to the incident ultrasonic beam. Furthermore, the addition of flexible interface at the probe holder resulted to improved ultrasonic coupling at curved surfaces.
An ultrasonic inspection technique for the testing of nozzle welds was developed in the laboratory using the 2D matrix annular phased array probe, probe holder and NDT reference samples designed and manufactured during the project. The ultrasonic technique developed has successfully detected the majority of the introduced defects with good sizing capabilities. There were some internal reflections observed due to internal reflections and reverberations in the probe holder immersion bath. However, these internal reflections did not affect the inspection capabilities of the technique because these reflections are located outside of the ultrasonic inspection range where the defects are located.
Development of motion control and navigation system
A number of requirements have been taken into account during the design process of the navigation system for the scanner / manipulator scanning system. These are summarised below:
Reduced human intervention which will minimise the time spent by personnel at risk of exposure to harmful radiation levels
- Minimisation of set-up and removal time of the system from nozzle.
- Can inspect nozzles ideally in up to 25 % of time currently taken, thus reducing the extent of operator exposure to radiation.
- Can inspect nozzle welds without the need to use a large number of probes, calibrations and procedures.
Navigate complex weld profiles
- Improve the reliability of the inspection of the welded nozzle sections in nuclear and other safety critical facilities.
- The NDT robotic system must be able to support adaptability to inspect various nozzle shapes, geometries and configuration with respect to the vessel (welded to vessel in an non perpendicular angle).
- Provide faster and more reliable concept for fixing coordinates of the system, thus ensuring repeatability of inspections.
- Develop an advanced profile following system, with improved reliability and sufficient accuracy, to determine the position of the robot on the nozzle components.
Additional specifications derived from the inspection requirements are:
- The scanner must scan 360° around the nozzle without operator intervention.
- The scanner must scan up to 400 mm radially on the vessel from the blend radius.
- The manipulator shall include means to lock each drive axis at any position within the available movement range.
- The manipulator shall be capable of scanning the probe smoothly along each axis with a minimum speed of 5 mm/s and a maximum speed of 50 mm/s.
- The manipulator shall be able to be registered relatively to a marked datum on the nozzle / vessel surface.
- Circumferential, axial, and radial positions shall be encoded to an accuracy of < 0.05 mm, i.e. more than 20 encoder counts / mm.
The manipulator shall be capable of positional accuracy of ±2 mm on all axes. This error includes all play and backlash and suitable measures to eliminate these shall be implemented.
- The probe holder shall provide adequate gimbal rotation and sprung loading such that the probe has firm contact on the inspection surface at the prescribed positions. This must account for minor variations in the scanning surface enabling smooth motion of the probe.
The overall goal of the system design was to automate the procedures for:
(a) calibration and
(b) inspection.
The navigation system incorporates suitable sensors to collect the data needed to calibrate and perform the inspection with the desired accuracy and repeatability. The design achieves the following goals:
- Identification of the position and orientation of the robotic scanner's base
The robotic scanner base can be put at a random position on the nozzle, with the distance from the vessel not calibrated. The overall calibration time that is needed for the worker to position, align and fasten the inspection system on the nozzle is greatly reduced. Using the data rapidly collected from the sensors, the robotic scanner identifies its position and orientation through advanced algorithms and uses this information for positioning the probe to the desired position to initiate the inspection process.
- Initialisation of the system's components
Another time-consuming procedure that currently requires the worker to be on the vicinity of the nuclear reactor is the initialisation of the components that comprise the inspection system. With the proposed design, this initialisation is carried out automatically, so that human intervention is reduced and greater repeatability can be achieved. The automatic initialisation takes place as soon as assembly has been completed by using absolute encoders and thus the workers radiation exposure is minimised during the system installation on the nozzle in dangerous ionizing conditions.
- Awareness of the system state during inspection
In order to follow the inspection path on the vessel surface, the system needs to have information of its state. The probe is constantly in contact with the vessel and applies a vertical pressure to follow the varying curvature. Failing to apply the required pressure, due to bad positioning for instance, may result to probe flipping or pure compliance which could severely affect the ultrasonic behaviour. At the same time, accuracy and repeatability of ultrasonic measurements is a major requirement from an inspection system. Therefore, the accurate knowledge of systems position is needed during the inspection, not only for knowing the exact location of a possible defects, but also in order to adjust the ultrasonic parameters to perform, as efficiently as possible, while moving on the curved vessel surface.
- Correction of the navigation algorithms with real-time feedback of the actual state of the system
No matter how rigid a system may be, there will always be some deflections caused by either the mechanical flexibility of system structural parts or the navigation and control systems' performance. Other parameters that can affect the navigation accuracy are the manufacturing tolerances of the nozzle and vessel or structural deformations that may have developed through the plant operation time. In order to compensate these faults and ensure that the probe will operate with accuracy, the navigation system uses real time measurements to adjust path planning based on the state of the robotic scanner. The parameter that is believed to be the most critical for the probe's compliance is the distance between the gimbal joint and the vessel surface.
The application of scanning a surface on a determined path by applying a force (vertical probe compliance pressure) during inspection requires controlling both position and force. Such control systems are described as hybrid. The hybrid position / force controller must solve three problems [1]:
1. position control of a manipulator along directions in which a natural force constraint exists;
2. force control of a manipulator along directions in which a natural position constraint exists;
3. a scheme to implement the arbitrary mixing of these modes along orthogonal degrees of freedom of an arbitrary frame.
Development of a low cost automatic scanning system
The developed robotic scanner's main subsystems are the following:
1. the collar base
2. the main manipulator
3. the main end effector
4. the secondary manipulator with its end effector
5. the electrical and electronics subsystem
6. the control software.
The collar base comprises three main subsystems:
a. The main frame: It supports all components and equipment, creating a rigid yet lightweight base for the safe and reliable operation of the manipulator.
b. The locking and adjusting mechanism around the nozzle: It provides adjustability with reference to its optimum compliance on various nozzle diameters. Finally, it ensures safe clamping of the main frame on the nozzle
c. The rotational degree of freedom: It provides precise rotation, of a range of 360°, to a carriage where on it the two manipulators are mounted.
The main manipulator is comprised by two links that are both driven by a DC motor and two gearboxes. Each joint is designed to be rigid without any backlash. Power transmission in the case of the second link is done by means of a timing pulley stage. The design and implementation of it is presented in detail in the following paragraphs.
Links and joints
The two links are selected to be of a rectangular profile in order to achieve high stiffness to torque and bending. According to the loading of the mechanism the first link is imposed to torque as well as bending. In the same fashion the second link is only subjected to bending. Obviously, the maximum bending forces applied on these are when the manipulator is fully extended (820 mm distance / 400 mm stand-off) and totally horizontal.
The first link is selected with dimensions 100mm x 50mm and the second one 60 mm x 40 mm with a thickness wall of 4 mm. Their material is a relatively strong Al grade, 6082-T6 so that threading is possible for small diameter screws. Their dimensional tolerances are considered sufficiently tight, since they are produced by extrusion. Finally, the links can be anodised for optimum endurance to water spraying.
The two joints are designed for zero play and high rigidity. In joint 1, 2 angular contact bearings are used in order to tight them back to back and thus fully stabilise the shaft. On joint 2, low backlash bearings are used in combination with precision collars and shims to nullify play of the shaft. In general an attempt was made to minimise the weight of the two link mechanism.
An end effector was designed to mount the developed probe and probe holder. The main purposes that the main end effector should fulfil are:
- support the probe during inspection or when the probe is not restricted and may move freely;
- provide to the probe the necessary degrees of freedom to be able to perform a successful inspection while moving within the determined limits;
- apply vertical force to the probe in order to comply on the varying vessel surface and give the necessary ultrasonic measurements conditions;
- include passive means for taking account for minor variations in the scanning surface enabling smooth motion of the probe.
Degrees of freedom and structure
During inspection, in order for the probe to comply on the varying surface of the vessel, it will need to have two degrees of freedom. A gimbal joint was designed for that purpose, which allows two axis of rotation perpendicular to each. The main design consideration was to fulfil the system specifications with the less possible weight. This would reduce the required load from the motors of the manipulator and lead to lighter, more compact and rigid scanner design.
In addition, during inspection it is required that a certain vertical force is applied to the probe. Two springs that will be put on linear shafts will be providing this force. This mechanism will also allow small displacements away from trajectory, which will ensure the probe compliance on the vessel and smooth forces from surface variations.
Potential impact:
The SME partners consider that the NOZZLEINSPECT system will deliver a step change in technology through its further development and optimisation for full commercialisation. The main drive of the nuclear industry is to reduce the inspection personnel exposure to radiation during the set up and calibration of the system that is being carried out at a high radiation activity environment. This requirement was used throughout the design and development process. The mechanical design and motion control philosophy followed to meet this requirement by introducing features such as integrated system, easy mounting and installation, remote automatic calibration and automatic identification of the datum point. All these features reduce significantly (at least 25 - 30 %) the time required to install the automated inspection system compared to the existing system. Furthermore, a number of sensors have been incorporated into the inspection system to ensure a reliable and accurate remote monitoring of the inspection. Also safety features and sensors that protect the system from damage during the transportation, installation and scanning have been added that are directly linked to the developed control system. In addition to the mechanical and control developments, an innovative phased array probe with 3D steering capabilities and combined with a flexible water irrigation coupling system is a great advantage and significantly improves the detection capabilities of the developed system.
Socio-economic impact
As the current nuclear reactors are ageing and their working life is extended, regular inspections will be required to ensure the structural integrity and avoid any accidents that will lead to radiation leaks. The nuclear plant operators are extremely matriculated to ensure the smooth running of their assets and on the same time looking after the wellbeing of the personnel. The nuclear reactor area is a high radiation activity area and the personnel are required to enter to carry out the installation and calibration of the scanning systems for the testing of the nozzles. Therefore, the inspection personnel are exposed to high levels of radiation and their radiation dosage absorbed is closely monitored. If the radiation dosage of the personnel exceeds that annual permissible, then the person(s) are not allowed to work for the remaining of the year and that has direct financial implications to their employers and in turn to the nuclear power plant operators that subcontract the inspection tasks.
Therefore, the NOZZLEINSPECT system provides an automated scanning system that significantly reduces the personnel exposure to ionising radiation by offering an integrated system that can be transferred to the inspection area and easily installed. The full set-up and calibration can be carried out remotely and outside the nuclear reactor vicinity and outside the containment concrete wall where there is limited radiation activity.
Dissemination activities
Scientific articles
A number of articles have been published in engineering and scientific magazines during the last six months of the project. They have successfully disseminated the project's results and achievements.
Electronic copies of the published articles can be found by following the links below. Page numbers are included to ease navigation:
- The Engineer
http://www.theengineer.co.uk/sectors/electronics/news/robot-inspects-structural-damage-in-nuclear-reactors/1009031.article
- EU 22
http://edition.pagesuite-professional.co.uk/launch.aspx?referral=mypagesuite&pnum=&refresh=Qq901Ty28a1J&EID=82bccec1-b05f-46f9-b085-701afc238b42&skip=
Pp. 256-257.
- Central Government 23
http://edition.pagesuite-professional.co.uk/launch.aspx?EID=0f269dd8-3e03-44f6-8e77-fc1d495cfd07
Pp. 126 and 143.
- International Innovation
http://www.research-europe.com/magazine/EUROFOCUS/2011-5/pageflip.html
Pp. 34-36
Conferences and meetings
The project results have been presented to a number of high quality conferences. Furthermore, the project results will be presented after the project end. A list of main conferences attended or planned to be attended are:
- British Institute of NDT 2010
- British Institute of NDT 2011
- 8th International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurised Components in 2010
- 9th International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurised Components in 2012
- 18th World Conference of NDT in 2012
- SPIE Smart structures / NDE (Non Destructive Evaluation) Conference
- 20th Mediterranean Conference on Control and Automation (MED'12)
- Annual Spanish nuclear meeting in 2012
- Exploitation of the results.
The NOZZLEINSPECT project delivered an integrated inspection solution to meet a range of requirements for inspecting nozzle welds in nuclear power stations. It is hoped that this shall be commercialised within two years of project completion in full. Additionally, there are many components of the solution that can be exploited in the more immediate future by the SME partners. There are several industrial and commercial routes for exploitation of the deliverables for the SMEs.
The aim of the project was to produce a working prototype of the inspection system to understand and overcome any technical issues with the hardware, and enable the development and investigation of software control and inspection algorithms. The route to full commercialisation will require further work involving aesthetics and production engineering of the individual systems to satisfy cost, standards and quality criteria. The control system will need to be adapted to ensure radiation protection.
Key developments that are exploitable by the SME partners include the following:
- a novel 2D phased array transducer that maximises inspection power and resolution within the constraints of a limited number of elements - available immediately;
- contour following motion control software incorporating inverse kinematic algorithms integrated with commercial motor drive hardware - concept proven and software modules available immediately for development of alternative configurations;
- 2D phased array beam steering software algorithms for controlling the transducer inspection direction, integrated with position feedback to target specific zones on a complex geometry - concept proven and modules available immediately. Further development required on graphical user interface, if the product is to be commercialised to provide user-friendly environment for more generic applications - estimated development time two years;
- data presentation for 2D phased array ultrasonic inspections (including defect evaluation and measurement tools) - limited concept demonstrated, estimated development time two years;
- flexible probe holder providing consistent coupling performance over a varying surface geometry - concept proven for low frequency inspection and available immediately. Further development required to provide higher frequency and/or zero couplant loss versions if required - estimated development time six months;
- automated centralising unit for quick clamping and positioning of the scanner mounting ring covering a range of pipe diameters - concept proven and available immediately for integration into alternative configurations;
- automated identification of datum reference using distance measurement sensors - concept proven and available immediately;
- fully integrated NOZZLE-INSPECT system for inspection of pipe to vessel saddle welds - concept proven, estimated further development time two years.
Individual subsystem inspection tools are ready for market (some immediately, others ready within one year). A number of these items have the potential to be of significant commercial benefit to the SMEs, with the project presenting opportunities for useful developments to be made available within their portfolio, for integration into other company products.
The overall inspection system requires testing to evaluate the design in terms ruggedness and durability, to identify areas of robustness and potential weakness. The data from this testing must be fed into a design review that should also incorporate modification for production engineering, as well as a compliance review to ensure that the system meets the operational requirements of a nuclear power plant. It is believed that only one review cycle should be necessary to provide a commercial robotic scanner suitable for the inspection of nozzles.
The software requires further work to integrate the developed modules into a generic, user-friendly package that links in with the mechanised scanner and is suitable for sale. The majority of the software development work would comprise providing visualisation methods for showing the ultrasonic data within the 3D part geometry and developing the corresponding analytical tools for the interrogation of the 3D data. Defect measurement becomes significantly more complex when dealing with large amounts of data. Hence, intelligent methods for simplifying the presentation of this information and thereby simplifying its interpretation would be of great commercial benefit.
A demonstration of the prototype system has been filmed by EuroNews channel as part of the project dissemination along with interviews with project partners to discuss how the system was integrated. This news item shall be freely available to disseminate through links to it on partner websites and can be directly promoted to target personnel within the nuclear industry. The aim is to market the system to a number of potential end-users including plant operators and inspection service companies. This activity is intended as a market survey to assess the interest level in the scanner to evaluate the level of investment that partners may require to commercialise the system. The secondary goal of this activity is to identify parties to invest in the further development of the equipment and to fund or partially fund the commercialisation. Identification of one or more committed customers for the system would provide the necessary qualification in the development and provide the confidence for all partners to complete a fully integrated system.
Project website: http://www.nozzleinpsect.eu
In order to reduce the time and cost of such inspections, there is a need to develop a system capable of performing a full inspection of nozzles without the need to frequently change ultrasonic probes. The aim of NOZZLEINSPECT project is to design a semi-autonomous and automatic inspection robotic system that is able to reduce the inspection and setup times, improve defect detectability and sizing and reduce human intervention. This will reduce workforce radiation exposure and reduce the requirement for complex robotic manipulation and consequently, reduce the size and cost of the robotic deployment system. This is achieved through the use of advanced 2D phased array probe that provides 3D steering capabilities and therefore some of the manipulation required can be achieved through the setup of the ultrasonic probe via specialised phased array software.
Project context and objectives:
The main goal of the NOZZLEINSPECT project is to improve the reliability of the inspection of the welded nozzle sections in nuclear and other safety critical facilities. The scanner and manipulator assembly will carry a new and novel adaptable phased array probe to allow a full inspection of nozzle weld areas, and an advanced navigation system that will follow the weld around the nozzle. The presence of defects in these parts could lead to component failure.
NOZZLEINSPECT will develop an advanced scanning / manipulator robotic system with a new generation of phased array technology providing advantages for the inspection of nozzles in nuclear power plants including:
1. inspection of nozzles in 25 % of the time currently taken, thus reducing the extent of operator exposure to radiation;
2. navigation of complex weld profiles;
3. inspection of nozzle welds without the need to use a large number of probes, calibrations and procedures;
4. improvements in the detection and sizing of defects in the challenging and complex nozzle welds.
Nozzle inspections require a variety of inspection techniques and equipment. The process necessitates that the plant be shut down for several weeks, which leads to lost production and profitability. Although automatic or semi-automatic inspection equipment has been introduced for many applications, operators are nevertheless still exposed to radiation during the equipment set-up and during certain operational activities.
Inspections are currently undertaken using a number of conventional ultrasonic probes. Since numerous probes are required to enable full inspections of all the nozzle weld areas, and each change of probe demands a recalibration to be completed, defect detection and sizing capabilities are not optimised. Furthermore, as the current probes have no beam steering capability, a large and expensive robot is required to provide accurate manipulation.
In order to reduce the time and cost of such inspections there is an urgent need to develop a system capable of performing a complete inspection of nozzles without the need to change probes. Such a system will offer the following clear benefits, including:
- faster inspection times;
- improved defect detectability and sizing;
- reduced human intervention which will reduce workforce radiation exposure;
- reduced requirement for robotic manipulation and, consequently, reduced size and cost of robotic deployment systems.
The objective of this project is therefore to develop a system that will use a semi-autonomous scanner / manipulator system for an automated inspection of nozzle welds. The robot will carry a novel adaptable matrix phased array probe which, combined with the 3D beam steering capability, will enable the entire volume of the weld to be inspected in a single operation.
NOZZLEINSPECT is a collaborative project consisting of the following organisations: TWI Ltd, Optel, Vermon, PeakNDT, Iberdrola Generation, Phoenix, KTU and Cereteth. The project is coordinated and managed by TWI Ltd and is partly funded by the EC under the Research for the benefits of SMEs project reference No. FP7-SME-2008-1-GA-232523.
Project results:
Provision of samples
The project consortium used a very detailed specification to understand and agree the approach to be adopted for the robotic design and this in turn needed some initial kinematic simulation to be carried out. It was apparent that the structural mock-up needed to be large but light enough to be portable and mounted on wheels and at the same time it should be capable of being dismantled for shipment to the project partners. In addition, the mock up must not topple over when the robot is being manoeuvred and must be stiff enough to resist the clamping and other reaction forces.
This nozzle geometric mock up was designed and fabricated for the robotic scanning system and control algorithms development and the final testing of the complete system and its functionalities.
The design requirements of the nozzle mock-up that will be used for the development and testing of the scanner and probe manipulator assembly were as follows:
The mock up consists of two main components, which are the pipe nozzle and back plate fabrication and the stand.
Pipe, nozzle and back plate fabrication requirements:
- one single steel fabrication preferred to achieve the desired tolerances;
- machine steel pipe > 5 mm wall thickness;
- pipe wall thickness suitable for induced bending moments based in assumptions;
- welded end flanges to allow connection to nozzle fabrication and stand;
- concentricity/roundness +/- 1.0 mm;
- surface finish< 3.2 µm or better on both 345 mm and 605 mm pipe outer diameter;
- replicates outer surface profile of nozzle to allow functionality of manipulator to be tested;
- fabricated to nominal fabrication tolerances as stated in BS22768;
- rolled plate to form curvature of vessel reactor and bolts to nozzle fabrication;
- replicates outer reactor radius to an area of 1.6 m x 1.6 m;
- vessel inside radius 2785 mm;
- provides a structural support for the nozzle feature although has additional stiffeners to inhibit flexure.
Mock up stand fabrication:
- fabricated steel frame structure;
- lockable castor wheels to aid moving around;
- stable arrangement to allow manipulator functional tests to be performed;
- allows close access to the pipe.
Another research aspect of the project was the ultrasonic technique development for the inspection of the nozzle to vessel weld. Due to the nozzle to vessel geometry, the weld profile and volume changes with circumferential position around the nozzle to vessel weld. Therefore, three welded reference samples with known defects that represent the weld profile at 0º, 45º and 90º were designed and manufactured. These samples were used to develop the ultrasonic inspection technique. A total of 18 defects were introduced in the nozzle to vessel weld that provides a wide range of sizes, skew and tilt angles. The defects in the nozzle to vessel weld will be inner and outer surface breaking defects as well as internal defects in order to assess the full volume coverage of the developed ultrasonic technique. The defects have different sizes, skew and tilt angles to represent the type of defects that are expected in the nozzle weld structure. All these different parameters assisted the development of the inspection technique and evaluation of the detectability and sensitivity achieved as well as understanding the effect of the defect parameters to the detectability.
Three ultrasonic validation welded reference test samples with introduced defects were required to be manufactured for the ultrasonic technique development purposes. As part of the project, a trial sample to demonstrate the dimensional accuracy of the flaws implanted by electrode discharge machining (EDM) was prepared.
In order to qualify the flaw implant procedure, a test coupon was manufacture applying the same welding and machining processes as would subsequently have been used to manufacture the ultrasonic reference samples. The test coupon was obtain by butt-welding two mild steel plates 300 mm long x 150 mm wide x 50 mm thick, using submerged arc welding (SAW) process and using the exact welding parameters employed during the fabrication of the nozzle.
In order to obtain embedded flaws, the weld bevel was filled up to mid-thickness, four notches with different height and length were then implanted by EDM, using skew and tilt angles typical of the UT validation blocks.
The ultrasonic reference samples included surface breaking notches at the weld root and cap, on the base metal (nozzle inner radius), as well as embedded notches at mid-thickness. Considering that EDM notching is a fully automatic process, it was not considered necessary to qualify the procedure used to obtain the surface-breaking notches. On the other hand, embedded notches are obtained by EDM and subsequently buried by manual metal arc welding (MMA), therefore, the qualification activities focused on these type of flaws.
For the final demonstration of the developed automatic inspection system, the consortium has secured access to the N5 nozzle mock up that was manufactured for the qualification of new inspection techniques and the training of operators. The full nozzle mock-up fully represents the geometry of the actual component and has a number of different defects introduced at the nozzle weld and the inner radius.
Design of an adaptable phased array probe
In order to design and manufacture a suitable phased array probe for the inspection of nozzle weld, a number of different designs on the element configurations were simulated to find the optimum design. The pulser-receiver that will be used for the inspection is capable of pulsing 128 elements simultaneously and therefore all the phased array designs are working around this parameter and limited to 128 elements.
In order to optimise the geometry of 2D phased arrays, various configurations were considered including 2D matrix, 2D annular segmented and 2D matrix annular phased arrays. The beam profiles generated by these arrays were simulated using CIVA 9.2 and the results were compared with respect to beam amplitude, focal depth, focal zone, etc. An algorithm using an apodisation technique was used in order to reduce the level of side lobes that result to ultrasonic energy loses of the primary beam and potential unwanted echoes. An ultrasonic field modelling software called SimulUS developed by the project partner that was used to apply the apodisation function to the 2D phased arrays and the beam profiles generated by various 2D phased arrays were simulated. Also this software was used to compare the ultrasonic field generated by annular matrix arrays and matrix arrays. One of the novelties and innovation of this project is to use flexible coupling material (instead of using wedge) which allows the 2D phased arrays to be placed on both flat and curved surfaces, which are both present during the nozzle to vessel weld inspection. Therefore, in order to investigate the effect of using the flexible coupling material and the wedge, their corresponding ultrasonic field were simulated and compared.
The matrix-annular phased arrays are constructed using square or rectangular elements in a matrix configuration. By removing a number of elements from the corners of the matrix phased arrays or manually setting, these elements in the matrix arrays as not functioning or sending and receiving any ultrasonic waves. That way an annular shape array can be obtained using the elements square or rectangular shape elements. Therefore, when giving the geometry details of the matrix annular phased arrays the geometry of matrix array is given as well. The reason using smaller gap is to obtain larger elements with reduced probe footprint and then to extract more ultrasonic energy from the array and therefore introduce more ultrasonic energy into the material. The final probe size is 31.9 mm by 29.9 mm, and the element pitch in X and Y directions is more than one wavelength, which are 2 mm and 3 mm, respectively. The reason for using this configuration is that even though the large pitch size results in the generation of grating lobes, if the level of grating lobes is acceptable this configuration is a good option since it is able to provide larger amplitude of the ultrasonic beam.
Overall, the phased array probe configuration selected as the best configuration among the 2D phased arrays due to the following reasons:
1. It gives the highest amplitude of ultrasonic beam when steering the beam at 40 in YZ plane. Note that the linear array (unable to steer beam in XS plane) and AS2 (difficult to implement in reality) are not considered.
2. The level of the grating lobes generated by MA3 is small and acceptable.
3. The focal depths of -3 dB and -6 dB are the largest and the focal zone is able to reach the bottom of the inspected component.
A mechanical probe holder has been designed to house the phased array probe the selected probe. The probe holder incorporates a flexible membrane such that the surface in front of the phased array probe adapts to the shape of the surface of the vessel or nozzle being inspected. The probe holder also incorporates specific features to assist with both the ultrasonic and mechanical performance when connected to the robot scanner.
Having developed a phased array probe specific to the nozzle weld inspection task a method was required for holding the probe at a known stand off from the surface at a prescribed angle and a method for coupling the ultrasound to the surface through a liquid or solid medium. Conventional inspections methods often use solid wedge transducers manufactured from plastics such as Perspex. In this instance, solid wedges were not an option. This was due to the fact that the footprint of the developed phased array probe was relatively large (31.9 x 29.9 mm) and that the profile of the vessel surface would vary from flat to curved as the transducer was rotated around the nozzle. It was therefore necessary to develop an adaptable probe holder that would conform to the varying surface whilst still maintaining transducer orientation relative to the surface. The solution was to design a fluid-filled probe holder with a flexible front that would provide a bath of water between the transducer and the vessel surface.
The probe holder incorporated four water hose fittings that are attached to two separate delivery systems. Two of the hose fittings go into the body in order to fill the cavity with water for immersion couplant. One should be used as an inlet and the other a bleed hole (outlet) to ensure the body is filled. The other two hose fittings go straight though the body to two spray nozzles that wet the surface of the vessel ahead of the probe holder as it is scanned. This couples the ultrasound from the membrane to the vessel.
Two polyurethane membranes with different stiffness were manufactured and evaluated. A lower stiffness membrane (shore hardness 42) was found to be very flexible but also exhibited high friction when scanning on the surface. A higher stiffness membrane (shore hardness 60) had a much lower friction coefficient and was determined to be of adequate flexibility for the surface variation to be encountered on the vessel. The friction coefficient was measured to be around 0.9 for dry sliding contact. However when the surface was wetted through the spray nozzles the sliding of the polyurethane membrane improved significantly and the friction coefficient was measured to be around 0.3 for wet sliding contact.
Phased array ultrasonic technique development
As part of the ultrasonic technique development, the signal responses from 42 defects were simulated using an ultrasonic modelling software package. The orientation and location of each defect is varied: the angle is skewed from 0° to 10° and tilted from -15° to 15°; the position is in the middle or bottom of the vessel welding zone. The modelling software is capable of displaying the simulated inspection phased array results in sectorial scan, B-scan and C-scan of the defect. The half-skip method was used to inspect the defect located in the bottom as well as all the defects located in the nozzle weld. The stainless steel cladding present at the inner surface of the vessel, and at a distance of 50mm from the weld area, will make the application of the full-skip method difficult to be implemented. The full skip method relies on the ultrasonic beam to reflect on the component backwall and then reflect at the same angle as the incidence angle to the area of interest in the nozzle weld. However, due to the presence of the stainless steel cladding, the full skip method cannot be used because the ultrasonic beam will be distorted, and mode converted at the steel-stainless steel interface. Therefore, only the half-skip method is applicable where reliance is placed of direct reflected ultrasonic signals from the defects. The maximum signal response directly from each defect located in the middle was simulated and used to determine the optimal location.
For each defect, in order to receive the maximum amplitude of the signal, the location of the probe and the beam angle in relation to the probe were determined to identify the location of the probe (X-distance and Z-distance) and the beam angle (beam steer angle and beam skew angle).
Following the completion of the simulation task, the 2D matrix annular probe mounted into a flexible probe holder filled with water was used to test the manufactured ultrasonic reference samples with known defects. The ultrasonic beam was transmitted from the probe into the water and then passed through the flexible polyurethane membrane. The probe holder was placed on the vessel of the nozzle. The gel coupling was used between the membrane and the sample in the laboratory development work, and water coupling will be applied to the inspection of the real sample through the use of miniature spray nozzles.
The experimental results from the developed phased array probe and flexible probe holder shown that the introduction of 3D steering abilities has a significant effect to the detection of defects misorientated to the incident ultrasonic beam. Furthermore, the addition of flexible interface at the probe holder resulted to improved ultrasonic coupling at curved surfaces.
An ultrasonic inspection technique for the testing of nozzle welds was developed in the laboratory using the 2D matrix annular phased array probe, probe holder and NDT reference samples designed and manufactured during the project. The ultrasonic technique developed has successfully detected the majority of the introduced defects with good sizing capabilities. There were some internal reflections observed due to internal reflections and reverberations in the probe holder immersion bath. However, these internal reflections did not affect the inspection capabilities of the technique because these reflections are located outside of the ultrasonic inspection range where the defects are located.
Development of motion control and navigation system
A number of requirements have been taken into account during the design process of the navigation system for the scanner / manipulator scanning system. These are summarised below:
Reduced human intervention which will minimise the time spent by personnel at risk of exposure to harmful radiation levels
- Minimisation of set-up and removal time of the system from nozzle.
- Can inspect nozzles ideally in up to 25 % of time currently taken, thus reducing the extent of operator exposure to radiation.
- Can inspect nozzle welds without the need to use a large number of probes, calibrations and procedures.
Navigate complex weld profiles
- Improve the reliability of the inspection of the welded nozzle sections in nuclear and other safety critical facilities.
- The NDT robotic system must be able to support adaptability to inspect various nozzle shapes, geometries and configuration with respect to the vessel (welded to vessel in an non perpendicular angle).
- Provide faster and more reliable concept for fixing coordinates of the system, thus ensuring repeatability of inspections.
- Develop an advanced profile following system, with improved reliability and sufficient accuracy, to determine the position of the robot on the nozzle components.
Additional specifications derived from the inspection requirements are:
- The scanner must scan 360° around the nozzle without operator intervention.
- The scanner must scan up to 400 mm radially on the vessel from the blend radius.
- The manipulator shall include means to lock each drive axis at any position within the available movement range.
- The manipulator shall be capable of scanning the probe smoothly along each axis with a minimum speed of 5 mm/s and a maximum speed of 50 mm/s.
- The manipulator shall be able to be registered relatively to a marked datum on the nozzle / vessel surface.
- Circumferential, axial, and radial positions shall be encoded to an accuracy of < 0.05 mm, i.e. more than 20 encoder counts / mm.
The manipulator shall be capable of positional accuracy of ±2 mm on all axes. This error includes all play and backlash and suitable measures to eliminate these shall be implemented.
- The probe holder shall provide adequate gimbal rotation and sprung loading such that the probe has firm contact on the inspection surface at the prescribed positions. This must account for minor variations in the scanning surface enabling smooth motion of the probe.
The overall goal of the system design was to automate the procedures for:
(a) calibration and
(b) inspection.
The navigation system incorporates suitable sensors to collect the data needed to calibrate and perform the inspection with the desired accuracy and repeatability. The design achieves the following goals:
- Identification of the position and orientation of the robotic scanner's base
The robotic scanner base can be put at a random position on the nozzle, with the distance from the vessel not calibrated. The overall calibration time that is needed for the worker to position, align and fasten the inspection system on the nozzle is greatly reduced. Using the data rapidly collected from the sensors, the robotic scanner identifies its position and orientation through advanced algorithms and uses this information for positioning the probe to the desired position to initiate the inspection process.
- Initialisation of the system's components
Another time-consuming procedure that currently requires the worker to be on the vicinity of the nuclear reactor is the initialisation of the components that comprise the inspection system. With the proposed design, this initialisation is carried out automatically, so that human intervention is reduced and greater repeatability can be achieved. The automatic initialisation takes place as soon as assembly has been completed by using absolute encoders and thus the workers radiation exposure is minimised during the system installation on the nozzle in dangerous ionizing conditions.
- Awareness of the system state during inspection
In order to follow the inspection path on the vessel surface, the system needs to have information of its state. The probe is constantly in contact with the vessel and applies a vertical pressure to follow the varying curvature. Failing to apply the required pressure, due to bad positioning for instance, may result to probe flipping or pure compliance which could severely affect the ultrasonic behaviour. At the same time, accuracy and repeatability of ultrasonic measurements is a major requirement from an inspection system. Therefore, the accurate knowledge of systems position is needed during the inspection, not only for knowing the exact location of a possible defects, but also in order to adjust the ultrasonic parameters to perform, as efficiently as possible, while moving on the curved vessel surface.
- Correction of the navigation algorithms with real-time feedback of the actual state of the system
No matter how rigid a system may be, there will always be some deflections caused by either the mechanical flexibility of system structural parts or the navigation and control systems' performance. Other parameters that can affect the navigation accuracy are the manufacturing tolerances of the nozzle and vessel or structural deformations that may have developed through the plant operation time. In order to compensate these faults and ensure that the probe will operate with accuracy, the navigation system uses real time measurements to adjust path planning based on the state of the robotic scanner. The parameter that is believed to be the most critical for the probe's compliance is the distance between the gimbal joint and the vessel surface.
The application of scanning a surface on a determined path by applying a force (vertical probe compliance pressure) during inspection requires controlling both position and force. Such control systems are described as hybrid. The hybrid position / force controller must solve three problems [1]:
1. position control of a manipulator along directions in which a natural force constraint exists;
2. force control of a manipulator along directions in which a natural position constraint exists;
3. a scheme to implement the arbitrary mixing of these modes along orthogonal degrees of freedom of an arbitrary frame.
Development of a low cost automatic scanning system
The developed robotic scanner's main subsystems are the following:
1. the collar base
2. the main manipulator
3. the main end effector
4. the secondary manipulator with its end effector
5. the electrical and electronics subsystem
6. the control software.
The collar base comprises three main subsystems:
a. The main frame: It supports all components and equipment, creating a rigid yet lightweight base for the safe and reliable operation of the manipulator.
b. The locking and adjusting mechanism around the nozzle: It provides adjustability with reference to its optimum compliance on various nozzle diameters. Finally, it ensures safe clamping of the main frame on the nozzle
c. The rotational degree of freedom: It provides precise rotation, of a range of 360°, to a carriage where on it the two manipulators are mounted.
The main manipulator is comprised by two links that are both driven by a DC motor and two gearboxes. Each joint is designed to be rigid without any backlash. Power transmission in the case of the second link is done by means of a timing pulley stage. The design and implementation of it is presented in detail in the following paragraphs.
Links and joints
The two links are selected to be of a rectangular profile in order to achieve high stiffness to torque and bending. According to the loading of the mechanism the first link is imposed to torque as well as bending. In the same fashion the second link is only subjected to bending. Obviously, the maximum bending forces applied on these are when the manipulator is fully extended (820 mm distance / 400 mm stand-off) and totally horizontal.
The first link is selected with dimensions 100mm x 50mm and the second one 60 mm x 40 mm with a thickness wall of 4 mm. Their material is a relatively strong Al grade, 6082-T6 so that threading is possible for small diameter screws. Their dimensional tolerances are considered sufficiently tight, since they are produced by extrusion. Finally, the links can be anodised for optimum endurance to water spraying.
The two joints are designed for zero play and high rigidity. In joint 1, 2 angular contact bearings are used in order to tight them back to back and thus fully stabilise the shaft. On joint 2, low backlash bearings are used in combination with precision collars and shims to nullify play of the shaft. In general an attempt was made to minimise the weight of the two link mechanism.
An end effector was designed to mount the developed probe and probe holder. The main purposes that the main end effector should fulfil are:
- support the probe during inspection or when the probe is not restricted and may move freely;
- provide to the probe the necessary degrees of freedom to be able to perform a successful inspection while moving within the determined limits;
- apply vertical force to the probe in order to comply on the varying vessel surface and give the necessary ultrasonic measurements conditions;
- include passive means for taking account for minor variations in the scanning surface enabling smooth motion of the probe.
Degrees of freedom and structure
During inspection, in order for the probe to comply on the varying surface of the vessel, it will need to have two degrees of freedom. A gimbal joint was designed for that purpose, which allows two axis of rotation perpendicular to each. The main design consideration was to fulfil the system specifications with the less possible weight. This would reduce the required load from the motors of the manipulator and lead to lighter, more compact and rigid scanner design.
In addition, during inspection it is required that a certain vertical force is applied to the probe. Two springs that will be put on linear shafts will be providing this force. This mechanism will also allow small displacements away from trajectory, which will ensure the probe compliance on the vessel and smooth forces from surface variations.
Potential impact:
The SME partners consider that the NOZZLEINSPECT system will deliver a step change in technology through its further development and optimisation for full commercialisation. The main drive of the nuclear industry is to reduce the inspection personnel exposure to radiation during the set up and calibration of the system that is being carried out at a high radiation activity environment. This requirement was used throughout the design and development process. The mechanical design and motion control philosophy followed to meet this requirement by introducing features such as integrated system, easy mounting and installation, remote automatic calibration and automatic identification of the datum point. All these features reduce significantly (at least 25 - 30 %) the time required to install the automated inspection system compared to the existing system. Furthermore, a number of sensors have been incorporated into the inspection system to ensure a reliable and accurate remote monitoring of the inspection. Also safety features and sensors that protect the system from damage during the transportation, installation and scanning have been added that are directly linked to the developed control system. In addition to the mechanical and control developments, an innovative phased array probe with 3D steering capabilities and combined with a flexible water irrigation coupling system is a great advantage and significantly improves the detection capabilities of the developed system.
Socio-economic impact
As the current nuclear reactors are ageing and their working life is extended, regular inspections will be required to ensure the structural integrity and avoid any accidents that will lead to radiation leaks. The nuclear plant operators are extremely matriculated to ensure the smooth running of their assets and on the same time looking after the wellbeing of the personnel. The nuclear reactor area is a high radiation activity area and the personnel are required to enter to carry out the installation and calibration of the scanning systems for the testing of the nozzles. Therefore, the inspection personnel are exposed to high levels of radiation and their radiation dosage absorbed is closely monitored. If the radiation dosage of the personnel exceeds that annual permissible, then the person(s) are not allowed to work for the remaining of the year and that has direct financial implications to their employers and in turn to the nuclear power plant operators that subcontract the inspection tasks.
Therefore, the NOZZLEINSPECT system provides an automated scanning system that significantly reduces the personnel exposure to ionising radiation by offering an integrated system that can be transferred to the inspection area and easily installed. The full set-up and calibration can be carried out remotely and outside the nuclear reactor vicinity and outside the containment concrete wall where there is limited radiation activity.
Dissemination activities
Scientific articles
A number of articles have been published in engineering and scientific magazines during the last six months of the project. They have successfully disseminated the project's results and achievements.
Electronic copies of the published articles can be found by following the links below. Page numbers are included to ease navigation:
- The Engineer
http://www.theengineer.co.uk/sectors/electronics/news/robot-inspects-structural-damage-in-nuclear-reactors/1009031.article
- EU 22
http://edition.pagesuite-professional.co.uk/launch.aspx?referral=mypagesuite&pnum=&refresh=Qq901Ty28a1J&EID=82bccec1-b05f-46f9-b085-701afc238b42&skip=
Pp. 256-257.
- Central Government 23
http://edition.pagesuite-professional.co.uk/launch.aspx?EID=0f269dd8-3e03-44f6-8e77-fc1d495cfd07
Pp. 126 and 143.
- International Innovation
http://www.research-europe.com/magazine/EUROFOCUS/2011-5/pageflip.html
Pp. 34-36
Conferences and meetings
The project results have been presented to a number of high quality conferences. Furthermore, the project results will be presented after the project end. A list of main conferences attended or planned to be attended are:
- British Institute of NDT 2010
- British Institute of NDT 2011
- 8th International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurised Components in 2010
- 9th International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurised Components in 2012
- 18th World Conference of NDT in 2012
- SPIE Smart structures / NDE (Non Destructive Evaluation) Conference
- 20th Mediterranean Conference on Control and Automation (MED'12)
- Annual Spanish nuclear meeting in 2012
- Exploitation of the results.
The NOZZLEINSPECT project delivered an integrated inspection solution to meet a range of requirements for inspecting nozzle welds in nuclear power stations. It is hoped that this shall be commercialised within two years of project completion in full. Additionally, there are many components of the solution that can be exploited in the more immediate future by the SME partners. There are several industrial and commercial routes for exploitation of the deliverables for the SMEs.
The aim of the project was to produce a working prototype of the inspection system to understand and overcome any technical issues with the hardware, and enable the development and investigation of software control and inspection algorithms. The route to full commercialisation will require further work involving aesthetics and production engineering of the individual systems to satisfy cost, standards and quality criteria. The control system will need to be adapted to ensure radiation protection.
Key developments that are exploitable by the SME partners include the following:
- a novel 2D phased array transducer that maximises inspection power and resolution within the constraints of a limited number of elements - available immediately;
- contour following motion control software incorporating inverse kinematic algorithms integrated with commercial motor drive hardware - concept proven and software modules available immediately for development of alternative configurations;
- 2D phased array beam steering software algorithms for controlling the transducer inspection direction, integrated with position feedback to target specific zones on a complex geometry - concept proven and modules available immediately. Further development required on graphical user interface, if the product is to be commercialised to provide user-friendly environment for more generic applications - estimated development time two years;
- data presentation for 2D phased array ultrasonic inspections (including defect evaluation and measurement tools) - limited concept demonstrated, estimated development time two years;
- flexible probe holder providing consistent coupling performance over a varying surface geometry - concept proven for low frequency inspection and available immediately. Further development required to provide higher frequency and/or zero couplant loss versions if required - estimated development time six months;
- automated centralising unit for quick clamping and positioning of the scanner mounting ring covering a range of pipe diameters - concept proven and available immediately for integration into alternative configurations;
- automated identification of datum reference using distance measurement sensors - concept proven and available immediately;
- fully integrated NOZZLE-INSPECT system for inspection of pipe to vessel saddle welds - concept proven, estimated further development time two years.
Individual subsystem inspection tools are ready for market (some immediately, others ready within one year). A number of these items have the potential to be of significant commercial benefit to the SMEs, with the project presenting opportunities for useful developments to be made available within their portfolio, for integration into other company products.
The overall inspection system requires testing to evaluate the design in terms ruggedness and durability, to identify areas of robustness and potential weakness. The data from this testing must be fed into a design review that should also incorporate modification for production engineering, as well as a compliance review to ensure that the system meets the operational requirements of a nuclear power plant. It is believed that only one review cycle should be necessary to provide a commercial robotic scanner suitable for the inspection of nozzles.
The software requires further work to integrate the developed modules into a generic, user-friendly package that links in with the mechanised scanner and is suitable for sale. The majority of the software development work would comprise providing visualisation methods for showing the ultrasonic data within the 3D part geometry and developing the corresponding analytical tools for the interrogation of the 3D data. Defect measurement becomes significantly more complex when dealing with large amounts of data. Hence, intelligent methods for simplifying the presentation of this information and thereby simplifying its interpretation would be of great commercial benefit.
A demonstration of the prototype system has been filmed by EuroNews channel as part of the project dissemination along with interviews with project partners to discuss how the system was integrated. This news item shall be freely available to disseminate through links to it on partner websites and can be directly promoted to target personnel within the nuclear industry. The aim is to market the system to a number of potential end-users including plant operators and inspection service companies. This activity is intended as a market survey to assess the interest level in the scanner to evaluate the level of investment that partners may require to commercialise the system. The secondary goal of this activity is to identify parties to invest in the further development of the equipment and to fund or partially fund the commercialisation. Identification of one or more committed customers for the system would provide the necessary qualification in the development and provide the confidence for all partners to complete a fully integrated system.
Project website: http://www.nozzleinpsect.eu