Final Report Summary - POLYTANK (Development and Validation of an automated Ultrasonic system for the Non-Destructive Evaluation (NDE) of welded joints in thermoplastic storage tanks)
Executive Summary:
Plastics, such as high density polyethylene (HDPE), polypropylene (PP), polyvinylidene fluoride (PVDF) and polyvinylchloride (PVC) offer significant advantages over metals for the storage of chemicals such as acids, caustic soda, detergents and other corrosive liquids. They are chemically inert, tough, flexible and lightweight. However, plastics also have relatively low maximum operating temperatures and loads, which vary from plastic to plastic, as does chemical resistance. For this reason it is important that plastics tanks and storage vessels are designed and built to the relevant European standards.
It is estimated that there are over 50,000 plastics tanks in service in Europe, ranging in size from 150 litres to 100m3. Many of these store materials that are classified as hazardous and therefore have the potential to cause injury or death to personnel and damage to property and the environment in the event of a tank failure. Indeed, there have been a number of reported failures of plastics tanks in service.
Plastics tanks are normally designed for a finite life, usually between 15 and 25 years. However, due to economic pressure, many of these tanks are still in operation beyond their design life, often with little or no engineering justification. It is also not uncommon for plastics tanks to be used for storing chemicals that they were not designed to contain. For these reasons it is very important that operators of plastics tanks and vessels inspect them throughout their life. An issue at hand is that there are currently no standards for the in-service inspection of plastics tanks. There is also very limited expertise available on the visual examination of plastics tanks and virtually no use of non-destructive examination (NDE).
The majority of visual inspections are external and can therefore only identify cracks that break the outside surface of the tank. Since many of the cracks initiate from the inside of the tank there is already a leak path through the tank wall if and when the crack is detected. Internal inspections are carried out less frequently, if at all, because they are expensive, potentially dangerous to the inspector, and result in a shut-down because the tank has to be emptied.
The PolyTank project has determined the potential failure mechanisms in plastics tanks and storage vessels and developed ultrasonic NDE procedures, techniques and systems to be able to identify these. The project developed an inspection system that is able to be deployed on-site and is relatively simple to operate.
As part of the project, welded joints representative of those used to fabricate plastics tanks and vessels were manufactured containing known flaws. These were inspected and the NDE data analysed to determine the limits of flaw detection. In parallel, the significance of flaw size and quantity was established in relation to service requirements. This was achieved by long-term mechanical testing of joints containing known flaws, and comparison with results for welds containing no flaws.
The prototype NDE equipment, designed and built as part of this project, was successfully assessed under both laboratory and field conditions, and the inspection procedures developed will be submitted to CEN as the basis for producing a European standard on this subject.
Project Context and Objectives:
There are around 300,000 medium-to-large industrial storage tanks in the EU, mainly in the petrochemical, process plant, pharmaceutical and food processing sectors, of which about 60,000 are made of thermoplastic materials containing safety critical welds that are currently impossible to fully examine through the full body of the weld because there is no suitable NDT technique commercially available. Despite their economic, safety and environmental advantages the market growth for thermoplastic tanks into new and more demanding applications will not be realised unless a suitable NDT method can be commercially developed to examine the full volume of the weld. Virtually all medium to large thermoplastic storage tanks are fabricated using containment welds that include the base to the shell of the tank, shell helix welds or circumferential welds (depending on the design type) and inlet/outlet nozzles and instrumentation penetrations. All of the welds (except the shell helix welds) are made manually and are therefore more likely to contain manufacturing defects that threaten the structural integrity of the finished tank either in the short term or during its service life.
The current practice of visual inspection includes the tank inner surfaces, which requires tank preparation (opening, cleaning, purging etc.) that can take up to two weeks and means that normally only one or two tanks can be taken out of service at any one time, thus limiting the number of tanks that can be realistically inspected at one location.
Thermoplastics are relatively new structural materials and they provide significant challenges for NDT. In particular, these materials are acoustically very opaque. For more demanding service applications, such as higher operating temperatures and pressures, there has been a lack of confidence in the long-term reliability of these materials and this has restricted the use of welded thermoplastic tank systems in these applications, despite their obvious benefits, including their inertness to most chemicals, corrosion resistance, ease of fabrication and cost. However, because of the benefits for use in-service, there is a renewed interest in the use of thermoplastic tanks for these more critical applications provided that applied inspection can be shown to give a high confidence level in the welded joint integrity and that of the adjacent parent material. Added to this is the need to develop a realistic defect acceptance standard for a range of in-service thermoplastic welds and that the defect can be reliably detected to the acceptance criteria.
The two main techniques for welding thermoplastic tanks are extrusion and hot gas welding, both of which are manual processes. Extrusion welding is used for joining thermoplastic sheets of wall thicknesses greater than around 6mm or for long joint lines (greater than 5m), whereas hot gas welding is mainly used for joining thinner sheets or short, complex joint geometries. Extrusion welding (Figure 1.2) involves continuously extruding molten thermoplastic (extrudate), of the same material as the parts being welded, into a prepared joint between the parts, using an extrusion “gun”. The joint is preheated to its softening temperature by a stream of hot air before the extrudate is forced into the joint under pressure. This ensures that the extrudate and the parts fuse together and produce a weld.
The inspection technique proposed in PolyTank is for an ultrasonic examination of the full weld volume to take place from the outside surfaces of the tank, which means that it is therefore not necessary to open up the tank and prepare the inside for examination. Furthermore, the proposed ultrasonic inspection is far more searching as it is a full volumetric examination of the tank welds that, until now, have not been possible to inspect. The potential EU annual savings on tank preparation alone that would result from the project are €120million, based on a five-year inspection interval for 50% of the available tanks. Add in the other potential savings to EU industry and generated profit for the EU SMEs as a result of the project, then the total potential annual benefit could be as great as €165million per year, giving a “return on investment‟ [ROI] of 110:1 per annum based on project eligible costs. The ability to inspect critical welds in thermoplastic tanks in-situ and without disruption to normal operations will facilitate an increased market for the use of thermoplastic storage tanks in the process industry in preference to the more expensive and problematic steel tanks.
Three previous EC projects, WINDEPP, Polytec Systems and TestPEP, developed the basic concept of the ultrasonic inspection of butt fusion and EF welds in plastics pipes, respectively and went on to develop an NDE system for inspecting pipes in a wide range of sizes and between pipes and fittings as well as pipe-to-pipe joints. However, these projects only investigated the inspection of welds in PE pipes. The PolyTank project builds on the work carried out in these earlier projects and applies the techniques developed for pipe joints to the type of welded joints used in storage tanks.
This project has developed and optimised methods for inspecting tanks made from HDPE. The ultrasonic property of this material is unknown for the transmission of ultrasonic energy at angles non-normal to the outside surface of the tank scanning surface. This has been restricting the application of ultrasonic inspection and required a concentrated research programme to enable suitable ultrasonic techniques, procedures and equipment to be developed that will enable tank owner/operators to have a high level of confidence in the integrity of the welded joints. Furthermore, the welds in thermoplastic chemical storage tanks, can generate planar flaws with complex geometric problems to solve. Several angled ultrasonic beams (normally refracted shear wave angles of 45°, 60° and/or 70°) are generally used in weld testing in order to align the beam as near as possible to be at right angles to the main face of the defect in order to get the maximum reflection from the defect possible thus maximising the probability of detection. In specifying the angles to be used it is usual to consider the most likely orientations of anticipated potential defects, e.g. weld side wall fusion defects, root defects etc. However, the low velocity of the ultrasound creates a serious technical challenge in generating angled ultrasonic beams in thermoplastic materials. Angled ultrasonic beams are generated by refraction from a slower medium. There are very few materials that will support low velocity ultrasound and this challenge has restricted the application of ultrasonic NDE for thermoplastics.
Scientific Objective
Increased scientific understanding of the transmission of ultrasonic waves through thermoplastic materials to enable the detection of the required defect sizes in all joint types found in thermoplastic tanks with wall thicknesses in the range of 10mm up to 80mm.
Technical Objectives
A high power ultrasonic transducer array(s) with a system gain of 120dB and a signal/noise of 80 - 100dB with signal processing, using ultrasonic angled compression wave techniques at a frequency range of 0.5MHz to 5MHz that will be significantly attenuated than the more conventional and commonly used shear waves.
Creation of a knowledge database of up to 100 datasets allowing unique failure curves of defect size against creep rupture life to be plotted for different tank materials and joint configurations.
Development of flaw acceptance criteria for welded joints in storage tanks made PE thermoplastic materials.
Integration Objectives
Development of a pre-production prototype ultrasonic inspection system capable of inspecting welded joints in thermoplastic tanks up to 10m diameter and 80mm wall thickness.
Project Results:
Introduction
The PolyTank project is divided into six technical work packages:
Work Package 1: Manufacture of Welded Joints
• To produce a range of welded samples in the materials and types of joint that are representative of those present in thermoplastic chemical tanks currently used to store bulk liquids in industry.
• To insert into these samples a range of simulated flaws of known size and shape that are representative of real flaws found in service.
Work Package 2: Development of Ultrasonic NDE techniques
• To measure the physical ultrasonic properties of the thermoplastic materials specified in WP1.
• To develop inspection procedures and determine the limits of detection for each joint geometry and material specified in WP1.
• To develop data processing and analysis algorithms and software.
• To design and manufacture ultrasonic probes and probe wedges.
Work Package 3: Development of Flaw acceptance criteria
• To determine the effect of different types, sizes and concentrations of flaws on the long-term performance of welded joints in thermoplastic bulk liquid chemical tanks by carrying out accelerated mechanical creep rupture test.
• Based on the mechanical test results, develop a flaw acceptance procedure for the agreed range of flaws likely to be present during the service of the target thermoplastic welded joints.
• Compare the results with the NDT ultrasonic results to confirm that the flaw sizes detailed in the acceptance criteria can be detected using the developed ultrasonic technique using an array of transducers.
Work Package 4: Integration and Validation of Ultrasonic Inspection System
• To design and build a modular tank inspection for easy deployment on thermoplastic tank joints.
• Integrate the complete ultrasonic inspection system including the manipulator holding and scanning the transducer array probe head assembly, remote ultrasonic site, ultrasonic data acquisition and evaluation instrument and, control software, data analysis software and software for delay law formulation for the ultrasonic beam angle generated by the ultrasonic phased array.
• Validate the completed PolyTank system in the laboratory.
Work Package 5: Field trials using validated Ultrasonic Inspection System
• Prepare On-site Ultrasonic Inspection Procedure and Technique Sheets for specific weld geometries.
• Carry out Field Trials to assess the performance of the validated prototype inspection system and on-site ultrasonic inspection procedures.
WP1 - Manufacture of Welded Joints
Welds to be made in the project
A list of the welds to be made in the project was developed, taking into account the following:
• Limited budget available for welding and testing
• Limited budget available for materials
• Weld geometries and material thicknesses representative of the plastic tank industry
The extruded sheet material was HDPE, (Polystone® G-Natur HD), supplied by Röchling, Germany. Three material thicknesses were studied, 15mm (natural), 25mm (natural) and 40mm (black). It was originally planned to study 50mm, but this was not felt to be representative of the tank industry. The pipe was PE100, 110mm SDR11 (black), and was manufactured by Fusion Provida. The welding rod was the same grade as the sheet, 3 and 4mm diameter and yellow in colour. The granules were supplied by Röchling and the rod was manufactured by Hessel Ingenieurtechnik.
Four joint geometries were studied.
• Double-V butt
• T-Joint
• Cruciform
• Pipe outlet
Three flaw types were investigated:
• Lack of fusion between extrudate and parent material.
• Cracks in parent material and through the extrudate.
• Cold weld between extrudate and parent material.
TWI has produced a range of welded joints containing various idealised flaws (cold welds, embedded aluminium discs and partially embedded aluminium discs) using the procedures below. The joints were made in various configurations: double–V butt (Figure 1), T-Joint (Figure 2), cruciform (Figure 3) and pipe-in-sheet (Figure 4). Some of the joints contained no deliberate flaws and will be used as the reference welds. A number of duplicate welds were also made to enable the limits of detection and the critical flaw sizes to be determined more accurately. The total number of welds to be made is 120 and this is shown in Table 1.
An aluminium base plate, end plates and support bars were designed and manufactured to clamp and support the sheet for manufacturing T-joints and cruciform joints (Figure 5). A wooden base plate and steel clamping bars were designed and manufactured to hold the sheets for manufacturing the double-V butt joints (Figure 6). A fixture was manufactured to ensure accurate pipe positioning and support for manufacturing the pipe-in-sheet joints. This is shown in Figure 7.
The list of welds made is given in Table 2.
The manufacture of pipe-in-sheet joints is shown in Figure 8 and a photograph of some of the completed joints is shown in Figure 9. A photograph of a 40mm thick double-V butt joint being manufactured is shown in Figure 10 and photographs of completed double-V butt joint samples are shown in Figure 11. A photograph of 40mm thick T-joints being made are shown in Figure 12 and a photograph of completed 25mm cruciform joints is shown in Figure 13.
Double-V Butt Joints
A common type of edge preparation for a butt weld is a double-V configuration. This is shown in Figure 14. For all material thicknesses, the double-V edge of the sheet was prepared using a CNC-milling machine. The angle of the preparation was 60° inclusive.
T-joint and cruciform joint
T-joints are widely used in the fabrication of plastic tanks. For example, a T-joint would be used for joining a tank body to a base plate. Cruciform joints are not common in the fabrication of plastic tanks. However, they are being made in the project because the long-term mechanical test to assess the effect of flaws in the fillet welds requires a cruciform joint rather than a T-joint. The T-joints will therefore be used to assess the NDE technique and the cruciform joints will be used to determine the flaw acceptance criteria. An example of a cruciform joint is shown in Figure 15.
Pipe-in-sheet
The pipe-in-sheet weld configuration was chosen to represent a pipe inlet/outlet in a tank. This is a common configuration for joining pipework to tank bodies. An example of a pipe-in-sheet trial weld (with partially embedded flaws) is shown in Figure 16.
Flaw Insertion Procedures
Planar flaws in the weld were simulated using aluminium discs, which were made out of 0.025mm thick aluminium foil with an adhesive backing, and with a purity of 99%. They were supplied by Goodfellow UK. 2, 3, 4, 8, 15 and 25-mm discs were punched out from foil.
Hole punches and dies were manufactured so that aluminium discs with repeatable and consistent dimensions could be punched out of the foil. The hole punches and dies are shown in Figures 17 and 18.
Procedure for inserting embedded aluminium discs
1. Lay two pre-machined lengths of HDPE sheet on to a flat bench and secure in place.
2. Using vernier callipers, measure from one edge of the HDPE sheet along the preparation to determine the location of the aluminium disc placement.
3. Use a scalpel blade to carefully remove the backing paper from a pre-punched aluminium disc to expose the adhesive layer (Figure 19).
4. Place the aluminium disc in the centre of the preparation using a pair of tweezers as shown in Figure 20.
5. Using the rubber end of a pencil, gently press the aluminium disc against the HDPE surface until the adhesive adheres to the surface producing as smooth a finish as possible (Figure 21).
6. For each of the aluminium discs, mark its position and size using a white correction pen (on black sheet) or permanent marker (on natural sheet) on the adjacent surface of the HDPE sheet in a location that will not be obscured after welding. This is shown in Figure 22.
Procedure for inserting partially embedded aluminium discs
1. Place the pre-cut lengths of sheet into the welding jig as shown in Figure 23.
2. Tack the sheets together using a hot air welding torch as shown in Figure 24. The sheets should be tacked along their full length to ensure that they don’t move during the welding operation. The welding torch temperature should be calibrated to 315°C using a Calstik temperature calibration device.
3. Once all of the joints have been tacked, carefully remove the tacked assembly from the welding jig.
4. Using a steel ruler and the sharp point of a compass and mark the predicted distance of the extrusion weld width from the weld root tack on the HDPE sheet. Draw a line between the points using a permanent marker or white correction pen Figure 25.
5. Using a steel ruler measure the embedded distance from the marked weld width line (i.e. towards the weld root), depending upon the required disc insertion depth. Mark this position with the sharp point of a compass (the mark should be visible without causing significant indentation) and draw a line parallel to the first line.
6. Use a scalpel blade to carefully remove the backing paper from a pre-punched 25mm diameter aluminium disc to expose the adhesive layer (Figure 19).
7. Place the aluminium disc so that its edge is on the insertion depth marked line, as shown in Figure 26.
8. Using the rubber end of a pencil, gently press the aluminium disc against the surface of the HDPE sheet until it adheres, producing as smooth a finish as is possible.
Transverse cracks
To simulate cracking across fillet welds in polyethylene tanks, which can be generated by chemical attack, the following flaw insertion procedure should be carried out:
Procedure for producing Transverse cracks
1. Mark on the sheet, with a white correction pen or permanent marker, the required location of a crack.
2. Load a 0.15mm slitting wheel (Figure 27) on to a CNC milling machine as shown in Figure 28.
3. Secure the welded sample to the CNC milling machine using machine clamps.
4. Insert the rotating slitting wheel into the fillet weld to the required depth using the CNC display (Note: this operation should only be performed by a skilled machine operator).
5. Once the crack has been made, mark on the sheet, using a white correction pen or permanent marker, the final location of the crack (as the crack can be hard to detect visually).
Cold weld
Preheating the parent material using hot air is essential for ensuring that a good quality extrusion weld is achieved as shown in Figure 29. However, due to a lack of knowledge, extrusion welded tanks are sometimes fabricated without the correct preheat being applied and in some circumstances with no preheat. To simulate this type of flaw a preheat deflector plate was manufactured and retro-fitted to the extrusion gun. This is shown in Figure 30.
Procedure for producing cold welds
1. Before carrying a cold weld attach the deflector plate to the Leister S2 extrusion gun nozzle using screws (see Figure 31).
2. Set the pre-heat (260°C) and extrudate (230°C) temperatures on the extrusion gun.
3. When the pre-heat and extrudate temperatures have been reached, carefully check that the hot airflow is being directed away from the weld preparation.
4. Produce the weld.
WP2 – Development of Ultrasonic NDE Techniques
Basic Material Properties
In Non-Destructive Testing (NDT) using ultrasound an important task is to create the correct inspection configuration. This task involves selecting the probe, the coupling medium, the scan plan, the scanning system, etc. These decisions are based on the specimen to be investigated, both geometry and material. The polyethylene plastic materials which were investigated in this project are known for their very slow acoustic velocity and highly attenuating nature. However, the exact properties of the material will vary depending on for example the material grade. Therefore, these properties need to be determined for each material to be tested. The determined information about the ultrasonic properties will then be used to design the inspection configuration; selecting probe parameters; and selecting coupling solution.
Acoustic properties
In ultrasound there many different modes in which the sound travels. The two most common modes in NDT are longitudinal and transverse waves. In plastic pipes only longitudinal waves propagate significantly. Transverse waves are highly attenuated due to the properties of the material. The acoustic properties of PE, such as velocity and attenuation of longitudinal waves and their dependency on frequency need to be well known. These ultrasonic parameters are necessary for development of NDT techniques, configuration of NDT equipment, detecting and measuring wall thickness and flaws, and for modelling which is used during inspection technique development. The schematic drawing of the setup used to acquire the data required for the determination of the acoustic properties can be seen in Figure 32. A pulse-echo configuration was used where the same probe was used to transmit and receive the ultrasonic signal through a 30mm thick plastic sheet. A number of phased array probes with different frequencies were used to get the acoustic properties and their dependency on frequency. The frequencies used were 2MHz, 5MHz and 10MHz.
Velocity
In this project plastic sheets from High Density Polyethylene (HDPE) were used. The ultrasonic properties needed to be determined for this material to be able to develop and calibrate the inspection techniques. The ultrasonic A-scan was recorded for each probe used; see Figure 33 for the measured A-scan using the 2MHz probe. By measuring the time it takes for the ultrasonic signal to travel through the material and mechanically measure the thickness of the material the velocity can be calculated as
c = dm/tu
where dm is the mechanically measured thickness of the sample and tu is the ultrasonically measured time taken for the ultrasound to travel through the sample. For all the frequencies the velocity was determined to be 2400m/s. For the probes used and the material under investigation, no effect of frequency dependency on the velocity was established.
The possible difference in velocity in the parent plastic material and the weld was also investigated. Figure 34 shows the response received looking through a 70mm thick sample where some part of the signal travels through the weld and some part travels through the parent material. The signal from the back wall when travelling through the weld is slightly delayed compared to the signal travelling through the parent material. This indicates that the velocity is lower in the weld. The velocity in the weld was calculated to be 2356m/s. However, the decrease in velocity is less than 2% and will not have any effect on selecting the inspection configuration.
Attenuation
The same experimental configuration that was used to get the velocity was used for the determination of the attenuation. Figure 35 shows the measured ultrasonic attenuation in relation to the expected probe frequency. The attenuation is measured as decibel per millimetre. It can be seen that the attenuation increases rapidly with increasing frequency. Furthermore, usually the actual frequency of the probe is less than the labelled value for higher frequency probes. This would make the attenuation curve even steeper.
The results from the attenuation measurements show that selecting the probe frequency is important. When inspecting thicker components, where the propagation distance increases, a lower frequency is required for the sound to propagate the required distance.
Effect from ageing
Theoretical calculations indicate that the effect on acoustic properties due to ageing is very small. The density and the elastic modulus could vary but the resulting variation in the velocity will be very small. This variation will have no effect when designing the inspection techniques and experimental configurations. To verify the theory, initial measurements have been taken on a specimen that will be subjected to ageing before final measurements can be taken.
Phased Array Transducers
The geometrical structure of the different welds in the polyethylene tank and the acoustic properties of the material, together with the requirements of a rapid and robust system, demand a phased array device. The fact that shear waves do not propagate in polyethylene means that we are restricted to using longitudinal waves. When utilising phased array technology there are options to use linear, circular, annular, etc. arrays, but in this application a linear array is most appropriate considering the measurement area and also enabling steering capabilities. A one dimension (1D) array has been used, where the benefits over a 2D array are; less computationally heavy; rapid calculations; lower cost; and not as demanding on the ultrasonic instrument.
The different joint configurations require different inspection solutions and hence different acoustic devices (phased array transducers). Designing phased array transducers is always a compromise between selecting the proper pitch, element width, aperture, and number of elements. A high number of small elements increases steering, reduces side lobes and provides focusing but the drawbacks are the cost of manufacturing and potential instrument complexity. Separating elements with a greater distance will increase the aperture size, but this creates unwanted grating lobes. Figure 36 shows an illustration of a phased array and some of its properties.
The transducer features have been selected based on results regarding the attenuation and velocity in the materials, the geometrical limitations from the weld to inspect and practical limitation from the array controller. Some of the transducer features are the dimensional parameters in Figure 36, and others are centre frequency, bandwidth, etc.
The total number of channels available in the array controller is 128. However, for this project the numbers of elements used were between 8 and 32. The pitch, gap and width were defined to be able to create the proper steering angle, focal depth, focal strength and aperture size. These properties were defined after experimental and modelling work. The objective was that the centre frequency was in the range 1-5 MHz The height of the aperture must be large enough to create the focal point at the desired depth.
Figure 37 and 38 shows curves that can be used as guidelines when determining the values of some of the parameters. One important property is the near field range, which is the greatest distance as a focal point can be created using the array, and be approximated as
,
where; f is the frequency, c is the speed of sound in the material and D is the aperture size. For a linear phased array, the aperture size is D=n’p, where n’ is the number of active elements and p is the pitch. Since the near field is proportional to the array aperture and a fixed number of active elements will be used, the pitch will be the deciding parameter. A larger pitch will create a higher focal length. However, the pitch is also determining the focal strength and the prevention of grating lobes. Ideally, the pitch should be less or equal to half the wavelength, λ/2, where λ=c/f. Using 32 elements result in the possibility to choose a pitch size from the dark areas in Figure 37 and 38. Fortunately, the requirement of half the wavelength can be relaxed, increasing the possibilities slightly. For a smaller tank with a thinner wall thickness, only requiring a near field range of about 40mm, frequencies up to 5MHz can easily be used without violating the requirement of half the wavelength, see Figure 38. Therefore, the selection of frequencies can be based on attenuation properties described above.
Optimising Transducer Parameters
Another factor that needs to be taken into consideration is the propagation distance. This is highly affected by the attenuation in the material and the frequency selected. To cope with this beam modelling was used together with information about the geometrical constraints. Beam field modelling is conducted to assure that the beams generated by the array in the housing are performing in a desired way. In order to do this, the SimulUS software was used. This software can be used to minimise side lobes and to make sure the desired propagation distance is achieved. Steering at higher angles will always push the limits of the transducer and side lobes will be generated. However, these need to be minimised in order to get more energy to the desired inspection region, see Figure 39 for examples of high angle steering for both the 4MHz and the 2.25MHz transducers.
An important factor when developing inspection techniques and optimising the hardware, i.e. transducers and wedges, to be used, is the geometrical constraints. The developed inspection technique will tell where the transducer will be placed and, hence, give indication on the physical footprint of the transducer. For the butt weld, the inspections will be carried out with the transducer placed on the side of the weld. The area next to the weld is smooth and does not have any direct limitations in size, which help the transducer parameter optimisation. However, for the fillet welds, the optimal position of the transducer is on the weld cap. This limits the maximum size of the transducer, which needs to be taken into consideration. Several linear 1D transducers with frequencies between 1-5MHz have been used in the project. The parameters for the transducers are presented in Table 3. The number of transducers for each joint type has been kept at a minimum to make the system as flexible as possible. The optimum would have been one transducer for each joint size, but the wide range of sizes and type of defects to detect covered in this project did not allow that. The transducer parameters have then been optimized in terms of resolution and coverage and the optimum solution is presented in Table 3.
Transducer Specifications
The two transducers TJ1 and BW1 have been optimised to inspect the weld joints identified in WP1 to detect the types of flaws inserted into the welded samples made during WP1. The transducer BW1 is available at TWI and has been used in the project to cover the BW joints. The transducer is shown in figure 40. The transducer TJ1 was not available from TWI stock and, following the principles outlined above, a specification for the required transducer was developed. This specification was sent to the relevant suppliers of phased array transducers listed below for quotes for manufacture:
• Sonaxis
• Vermon
• GE
• Zetec
• Imasonic
• Olympus
Quotes were received from three of the suppliers. However, the lead times to supply these bespoke transducers was not acceptable within the project timescales and so an in stock transducer with acceptable parameters was identified and procured. This transducer is shown in figure 41. During the course of the inspection procedure development it was identified by the project steering committee that it would be beneficial to detect an additional type of defect to those already detailed in WP1. This resulted in two additional techniques, one of which could be accomplished with the use of transducer TJ1, but the other technique required a transducer with different parameters to those already detailed. Fortunately another transducer already available at TWI could fulfil those parameters and is detailed in table 3 as TJ2 and shown in figure 42.
Prototype Ultrasonic Probe Wedges
In order to transmit ultrasound from a phased array ultrasonic transducer into the test material, the transducer needs to be coupled with the material using one or more mediums which are acoustically compatible with the material under examination. This will facilitate the transmission of ultrasound and reduce the amount of energy which is reflected at the interface. If the component material surface is flat and smooth and the required angle of the ultrasonic beam is approximately normal to the surface then the probe need only be coupled by the use of a suitable fluid, such as Ultragel, to eliminate air at the interface. However, if a beam of any significant angle from the normal to the surface is required then it is highly advantageous to introduce the sound from the transducer at a pre-defined angle by the use of a wedge made from a material with suitable acoustic properties. Traditionally, for metallic material inspection, the wedge material is often a type of plastic called Rexolite. However, this is not always suitable for HDPE. If the component material surface is uneven then the probe can often be coupled at any angle by using a water path between the probe and the material. A simple way to achieve a water path is to submerge the component and transducer in a bath of water, but this is not always practicable. For site work a new type of wedge has been developed consisting of a wedge with a cavity to hold water for the water path, see figure 43. This may be either open faced, with the water in direct contact with the material surface and a flexible skirt to maintain a seal, see figure 44, or alternatively the wedge could be fitted with a flexible membrane to contain the water path within the wedge.
The butt welds produced have a very similar profile to a conventional steel double V butt weld. Initially an open faced water wedge was proposed, using a technique similar to that developed for HDPE pipe inspection. However, it was found that a standard Rexolite wedge, as used for steel weld inspection, produced very good results. The wedge allowed the transducer to produce compressional beams in the sample material of between 30 and 40 degrees and is shown with the probe in figure 40. Because the plate material is flat and smooth the fact that the Rexolite is inflexible presents no problems in coupling. The technique used to inspect the fillet welds for production flaws is to introduce ultrasound into the weld cap to examine the weld fusion faces. This technique is novel to weld inspection as it is not normally employed in conventional metal weld inspection. Although the HDPE weld surface is usually fairly smooth, it is not very even, in some cases being convex and in others concave. Furthermore, it is usually desirable when applying the inspection technique to introduce the UT beams at a mean angle approaching the normal to the fusion face, which is often at an angle of approximately 35 degrees to the weld cap surface. In order to determine the best design for a water wedge the optimum position of the required probe, BW1, in relation to the fillet welds for both T-joint and pipe insert was modelled and is shown in figures 45 and 46. To accomplish the correct probe angle and coupling a membrane water wedge was designed to accommodate the probe, BW1, and to fit the fillet weld geometry. The wedge was manufactured by a plastic prototyping technique and the final product is shown in figure 47. Although this wedge was designed to closely fit the probe BW1, it was found that it was just slightly too large to successfully couple with the smaller profile of the fillet welds in the 15mm thick sheet and the pipe in sheet welds. To overcome this shortfall a more compact membrane wedge was designed. This wedge was machined from aluminium and is shown in figure 48.
During the course of the project an additional type of defect was identified as being desirable to detect. This is in-service cracking at the toes of the internal fillet welds at the base of the tank. To detect these, two different techniques were developed, one for the weld toe in the tank wall and one for the weld toe in the tank base. To inspect the weld toe in the tank wall the technique involves introducing the sound beam at an angle into the tank wall. The method of construction of the tank wall results in a smooth but uneven surface. To accommodate this an angled membrane water wedge was employed as shown in Figure 49. This wedge was found from existing TWI stock. The technique to examine the base plate weld toe involves direct contact of the PAUT transducer onto the edge of the base plate and therefore no wedge is required.
Inspection Procedures
This section contains a description of the work carried out to develop the basic ultrasonic inspection procedures and techniques to inspect the weld geometries specified earlier. Work has been conducted to understand the geometry of the joints to be inspected. This is important for the inspection configuration to achieve appropriate coverage of the fusion zones, and also for the interpretation of the data from the scanned welded joint samples. Test specimens for each joint, tank material and joint thickness were created for complete development of the inspection techniques. The inspection results of these specimens have been interpreted to evaluate the performance of the inspection techniques.
Ultrasonic probes
Several linear 1D probes with frequencies between 1-5MHz were used in the project, depending on the joint type and the pipe size. The butt welded joints have good accessibility for the phased array probes and the physical size of the probe is not a significant problem. This means that the element size, pitch and gap can be optimised for the best resolution and penetration without considering any physical constraints. a big fusion area and normal 0-degree beams are to be used for the inspection. The T-joints will have restricted accessibility for a phased array probe. The base plate will be cut close to the fillet weld and the tanks usually have an irregular surface, not optimal for a phased array probe. The only remaining location for a probe is then on the weld itself. The probes then need to be physically smaller. The frequency dependent attenuation has been described earlier. The results indicated that if a longer propagation distance is required, a lower frequency of the probe should be used. In the T-joints, due to the long propagation distance, a low frequency probe needs to be used. However, when decreasing the frequency of a probe, usually the physical size of the elements increases. Therefore, the number of elements needs to be reduced. The pipe outlet joints have similar welds as the ones in the T-joints. The same ultrasonic probe will be used for these welds. The parameters for the probes desired in the project are presented in Table 3. The number of probes for each joint type has been kept to one each to make the system as flexible as possible. The probe parameters have then been optimized in terms of resolution, coverage and physical size and the appropriate solution is presented in Table 3.
Wedge
Water wedges have been employed for the inspection of all different joints in PE storage tanks. The advantages with a water wedge are the low attenuation; good acoustic matching with the PE material and the velocity ratio enabling the steering of angled beams to the fusion zone. The main challenges with a water wedge are possible air bubbles and to keep the water between the elements and the material. When designing and manufacturing the wedges, the aim is to make them as physically small as possible, but still allowing for the desired range of angles to be transmitted into the material. The possibility to use membrane water wedges is investigated. A membrane water wedge will have two major advantages; the water can be maintained in the wedge without any loss; the wedge surface is flexible and can be adapted for the surface profile.
Instrument
The variety of ultrasonic techniques used for the inspection in this project sets high demands on the ultrasonic instrument. The instrument that was used to inspect the welded samples in the project was the Dynaray from Zetec with their Ultravision 3 software. This is a very powerful instrument with 256x256 channels, high voltage and good signal-to-noise ratio. All ultrasonic techniques in this project can be implemented with this instrument. However, when integrating the system in the field trials, the Omniscan MX II PA by Olympus with its Tomoview software was employed. This is a portable phased array system that enables site inspections.
Inspection Techniques
Due to the different joint geometries described above, different inspection solutions are required. This includes phased array probe, coupling solution (wedge) and focal law settings in the phased array controller.
Butt welds
The inspection technique used for double V butt welds is shown in Figure 50. A standard sectorial scan is used, where the beam angle is swept between lower angles up to higher angles to ensure appropriate coverage of the fusion zones. By inspecting the weld from both sides, the sector pulse-echo configuration is believed to give a complete coverage of the weld. However, due to the complexity of the structure, caution must be taken when interpreting the scans. Direct pulse-echoes will be received together with signals form a skip on the back wall, which could be hard to interpret. Figure 50 shows the inspection from one side, but to get full coverage the welds need to be inspected from both sides. The lower fusion face is covered by the direct angles and the upper fusion face is cover through a skip on the back wall. A sector scan, using all the elements in the array to create an aperture, sweeping the beam from the lower angle to the higher angle is used. The step between each beam could be 0.5° to increase the resolution. The transmitted beams are focussed at the inner surface distance and Dynamic Depth Focussing (DDF) is used when receiving the beams.
The standoffs for the sector pulse-echo technique are depending on the plate thickness and the weld cap. The ultrasonic probe is positioned with an offset from the weld centreline due to the weld cap. The geometry of the weld cap and the weld will help defining the focal law settings to use. When the thickness of the plate increases the weld cap increases as well and the probe have to be positioned further away from the weld centreline. Therefore, different focal law settings are needed for different thicknesses of plastic plates. The focal law settings for the three different plate thickness used in this project are given in Table 4.
T-joints and pipe outlets
Due to the geometry of T-joints, different inspection techniques are required. Several different solutions, including the probe placed on the base, the tank and finally the weld itself, have been considered to be able to access all possible areas where a defect could be present, see Figures 51 and 52.
Development Results
A 20mm thick plate with aluminium discs inserted was used for the development of the inspection technique for the butt welds. Phased array scans using the setups described above were conducted on the plate. Figure 53 shows a typical S-scan on one location on the plate, with the weld overlay and some descriptive text. Figures 54-56 show three images of different features. In all three figures: A-scan (top left), S-scan (top left), and B-scan (bottom) are shown. The A-scan displays what the cursor in the S-scan indicates at the position where the cursor in the B-scan is located. The B-scan shows the data from the cursor in the S-scan all across the weld. Figure 57 shows the B-scan with all the detected aluminium discs highlighted. Inspection techniques have been developed for the inspection of welds in PE storage tanks. The developed inspection technique for double-V butt welds has proved capable of detecting all inserted aluminium discs in the test piece.
PAUT Detection Limits
Table 5 details the results of the inspections carried out on the Butt weld and T-joint Weld samples. The Pipe-in-sheet samples were not inspected due to the inability to successfully position the probe on the fillet weld. Further probe wedge development is required to achieve an inspection of this complex geometry. There were four types of weld flaw produced:
• Welds with no flaws.
• Welds with discrete flaws of varying sizes ranging from 2 to 25mm diameter.
• Welds with 25mm diameter flaws at varying distances from the root ranging from 2 to 10mm
• Welds with a cold weld element.
Figure 58 shows a typical scan of a butt welded sample with no flaw; Figure 59 shows a typical scan of a butt welded sample with discrete flaws and Figure 60 shows a typical scan of a butt welded sample with a cold weld flaw.
Figure 61 shows a typical scan of a T-joint welded sample with no flaw; Figure 62 shows a typical scan of a T-joint welded sample with discrete flaws; Figure 63 shows a typical scan of a T-joint welded sample with 25mm flaws and Figure 64 shows a typical scan of a T-joint welded sample with a cold weld flaw.
It can be seen from Table 5 that for all of the T-joint welded samples which were produced as fault free no flaws were detected. However, of the butt weld samples which were produced as fault free, four out of six (the thicker plates) contained 100% lack of root fusion in the centre of the double V joint. In addition, sample 42 contained 35mm of no weld and sample 82 contained 60mm of lack of side wall fusion.
For all of the butt welds produced with discrete flaws, all of the flaws were detected, some as small as 2mm. For all of the T-joint welds produced with discrete flaws, some smaller flaws were not detected, as few as 50% of the 2mm flaws. The percentage of discrete flaws detected is illustrated in Figure 65. Only the T-joint samples were produced with 25mm flaws at varying distances from the root, and unfortunately, the 40mm thick sample was produced without any of these in error. Therefore only two samples with this type of flaw produced any results, the 15mm thick sample and the 25mm thick sample. All of the flaws were detected in the 25mm thick sample, but for the 15mm thick sample the detection rate reduced from 100% for the discs at 2mm from the root to 25% for the discs at 10mm from the root. Of the samples produced with cold welds the current PAUT technique was unable to detect any signs that the welds were cold. However the inspection revealed that all of the double V butt welds contained up to 100% lack of root fusion.
The PAUT techniques developed were able to detect most of the artificial flaws inserted into the welded samples with the following results:
• The fault free samples were successfully inspected and some unintentional flaws were detected.
• The samples containing discrete flaws were inspected with a detection rate of 50% for the 2mm size flaw in the T-joint welds rising to 100% for the 25mm flaw.
• The detection rate for the 25mm size flaws at varying distances from the root was from 25% for the 10mm distance to 100% for the 2mm distance.
• Evidence of cold welds was not detected, however up to 100% lack of root fusion was detected.
WP3 – Development of Flaw Acceptance Criteria
WP4 – Integration and Validation of Ultrasonic Inspection System
A scanning system has been designed to encompass the full range of weld geometries specified in WP1. Currently available scanners which may be suitable for tanks are designed for metallic tanks and are frequently held in place with magnets, which will not be applicable for plastics tanks. It was the aim that a single scanner should encompass all the sizes of tank and the base fillet weld, pipe outlet fillet weld and base butt weld configurations. To extend the use of the manipulator and minimise costs it has been designed in a modular form with simple changes for different configurations. A complex part of the manipulator has been the probe holder, which is required to hold the probe in a very precise position around the fillet welds or butt welds. A system has been designed and manufactured in a modular format, with the capability of accommodating the different types of joints. The scanner system is integrated with an encoder for positional information.
Aim
The aim was to develop a 3-axis portable scanner that can be put in place by a single operator. It should be able to scan the weld at the bottom of the tank and also the weld around a nozzle pipe. The scanner should be:
• As lightweight as possible.
• As low a profile in z direction as possible to be able to operate in constrained spaces.
• Able to measure the force applied by the probe to enable the scanner to apply a constant contact force with a surface.
• Detect obstacles at the end of axis 2 to prevent collision with the ground and the nozzle pipe.
Assuming a minimum tank diameter of 2m, then a 600mm long scanner with three axis should be able to inspect the weld in segments of at least 500mm length at each pass. The three joints of the scanner should be prismatic (3 linear slides actuated by motor/gearbox combinations). The scanner will be attached to the tank wall by four or more suction cups actuated mechanically using levers. If required, an air compressor with venturi valves can actuate the suction cups for long periods. To inspect the weld at the bottom of a tank, attach the scanner above the weld. Axis 2 moves the probe over the weld and is then locked. Axis 3 brings the probe into contact with the weld with the desired contact force. Axis one then deploys the probe along the weld. Axis 3 makes any adaptation to the curvature by maintaining a constant contact force. To inspect a nozzle weld, the scanner is placed vertically on the tank wall. Axis 2 and 3 is actuated to bring the probe in contact with the weld at a starting point. Axis 3 is locked.(assuming that there will be little change required in the depth when the probe is moved around a half circumference of the pipe weld). Axes 1 and 2 then execute a curved trajectory to follow the weld line. The proximity sensor at the end of axis 2 can be used to guide the following of the trajectory.
In concept the scanner was envisaged to consist of 3 linear slides, 3 motor/gearbox combinations, 3 servo amplifiers and controllers, one supervisory controller.
• The motor/gearboxes on board the scanner should be splash proof.
• The servo amps and controllers will be in a separate splash proof rack quite close to the tank. Initially the system will not be designed for intrinsic safety operation in an explosive environment. The splash proof control rack should be wheeled for easy transportation to a test site.
• The Phased array probe should be spring loaded with a dual ball bearing, dual gimbal arrangement to keep the probe normal to the surface on slightly contoured surfaces.
• The control system should be easy to setup with programmed trajectories e.g. operator specifies points A and B by placing the probe on two points. The controller automatically performs a scan between A and B by maintaining the desired contact force and distance from the ground or pipe.
A review of commercially available NDT scanners with a similar concept was undertaken to aid in the design process. Taking into account the design considerations in above and using some of the ideas generated from the review of commercially available NDT scanners, several concept designs were produced as shown in figures 66, 67 & 68. After careful consideration and discussions with the interested SME, Innotech, and the project steering committee it was decided to take forward the concept shown in Figure 68. This concept was further refined into two options shown in figures 69 & 70. Both of these options will satisfy the project requirements and it was the option shown in figure 70 which was the one taken forward for assembly. However, this option can easily be modified to satisfy the option shown in figure 69. The scanner has been designed around commercially available components. Figure 71 shows the fully assembled scanner.
Validation Work
Table 2 details the joints which were produced for PAUT testing by the PolyTank system. All of the Double V butt welds and all of the T-joint fillet welds were inspected, however, due to problems with the membrane water wedge development; it was not possible to inspect the Pipe in sheet fillet welds. This shortcoming was accepted by the consortium as these were considered to be of a lower priority to the other welds. Two techniques were used to inspect the samples, an angle beam PAUT sectorial scan to inspect the butt welds and a zero degree PAUT sectorial scan to inspect the fillet welds. For each weld, the phased array system electronically focuses a compression wave acoustic beam at user specified depth within the material, which is determined by a series of “focal laws”. The focused beam is then electronically swept through a range of angles to fully cover the inspection volume. An image of the configuration for the technique used on butt fusion joints is shown in Figure 72. Butt fusion welds were scanned from both sides of the weld and both top and bottom of the plate. The datum point was the plate edge and linear scan measurements were taken from this point. Figure 73 shows the probe positions and scan direction. Table 6 gives the relevant probe and focal law parameters for the BF joints. The array controller software display screen was set up to show the A-scan, a B-scan and an S-scan (sectorial scan). Figure 74 shows an example for the BF joint inspection. An image of the configuration for the technique used on the T-joint fillet weld is shown in Figure 75. The T-joint fillet welds were scanned from the weld cap with the probe orientated to produce an angle between the central beam and the respective fusion face approaching 90 degrees. The datum point was the end of the welded section of each sample and linear scan measurements were taken from this point. Figure 76 shows the probe position. Table 7 gives the relevant probe and focal law parameters for the T joints. The array controller software display screen was set up to show the A-scan, a B-scan and an S-scan (sectorial scan). Figure 77 shows an example for the T-joint inspection.
Results
Table 5 details the results of the inspections carried out on the Butt weld and T-joint Weld samples. The Pipe-in-sheet samples were not inspected due to the inability to successfully position the probe on the fillet weld. Further probe wedge development is required to achieve an inspection of this complex geometry. There were four types of weld flaw produced:
• Welds with no flaws.
• Welds with discrete flaws of varying sizes ranging from 2 to 25mm diameter.
• Welds with 25mm diameter flaws at varying distances from the root ranging from 2 to 10mm
• Welds with a cold weld element.
The PAUT system was assembled and successfully tested in the laboratory. The Butt welded sheets and fillet welded T-joint samples were inspected and the results presented above. The pipe-in-sheet samples were not inspected due to limitations of the membrane water wedge.
WP5 – Field trials using validated Ultrasonic Inspection System
A finalised inspection procedure has been produced and was used during the field trials. The field trials were undertaken in three parts, firstly initial trials at Chemresist to demonstrate the capabilities of the techniques developed in the laboratory, then trials at Univar on live tanks to demonstrate the in-service flaw detection capabilities and finally blind trials at Chemresist to demonstrate the in-service flaw detection capabilities.
Initial Trials
The initial trials were carried out at the premises of Chemresist in Dewsbury, UK on the 24/25 June 2014. The objective of the initial trials was to demonstrate the capabilities of three inspection techniques which had been developed in the laboratory in a site environment. The three techniques were:
• T-joint fillet weld inspection for lack of fusion.
• T-joint fillet weld tank wall toe inspection for cracking.
• T-joint fillet weld tank base toe inspection for cracking.
The PolyTank probe fitted to the specially designed membrane water wedge was used to inspect the sample T-joint fillet welds cut from a sample tank provided by HSL identified as ‘Base-T1’. Figure 78 shows an image of the scan being carried out and Figure 79 shows a snap shot of the resultant scan. The resultant scans from this inspection gave indications from the nearside fillet weld root and the opposite fillet weld toes. There were no lack of fusion flaws present to be detected and the technique was unable to distinguish between a signal from the remote weld toes and the crack which was present. It was concluded from this that this technique is capable of inspecting the nearside fillet weld for fusion defects but not able to detect in-service cracking in the remote fillet weld toes.
A 4MHz probe fitted with an angle beam membrane water wedge was used to inspect the tank sample Base-T1 for cracking in the inner (remote) fillet weld to tank wall toe. An image of this scan being carried out is shown in Figure 80 and the resultant scan is shown in Figure 81. The resultant scans from this demonstration have shown that the technique using the 4MHz probe fitted with an angle beam membrane water wedge was able to successfully detect cracking in the fillet weld toe on the tank wall.
The PolyTank probe fitted to the specially designed compact membrane water wedge was used to inspect the sample T-joint fillet welds cut from a sample tank provided by HSL identified as ‘Base-T1’ to detect cracking in the remote fillet weld toe in the tank base. An image of this scan being carried out is shown in Figure 82 and the resultant scan is shown in Figure 83. This inspection technique provided a possible indication of the cracking in the remote fillet weld toe in the tank base plate, however the indication could not be positively confirmed and therefore further technique development is required.
The 4MHz probe fitted with an angle beam membrane water wedge was used to inspect the full tank sample T2 for cracking in the inner (remote) fillet weld to tank wall toe. An image of this scan being carried out is shown in Figure 84. There was no resultant scan as there was a significant inner and outer weld toe mismatch which created interpretation difficulties.
In-service Trials
The in-service trials were carried out at the premises of Univar in Wellingborough, UK. The objective of these trials was to apply the in-service crack detection techniques demonstrated in the initial trials to a small number of fully functional in-service tanks in a live tank farm. The tanks inspected are detailed in Table 8.
Tank WL23
This was the first tank identified for a trial inspection. Initial examination revealed that the construction of this tank was at variance with those inspected to date. Instead of the tank wall resting on the base in an inverted T form, the base was a sloping disc inserted into the tubular tank wall. Figure 85 illustrates the different construction methods. Due to the construction of the tank the only inspection technique which could be applied was the T-joint fillet weld tank wall toe inspection. Before this could be applied it was necessary to locate the exact position of the internal fillet welds. This was achieved by the use of a zero degree probe to plot the absence of the tank wall signal. Where there was no response from the inner side of the tank wall it was presumed that this was the position of the fillet weld. During this investigation the wall thickness was determined to be approximately 25mm thick. Figure 86 shows the weld position plotted onto the outer surface of the tank wall. The sample length of weld plotted out as shown in Figure 86 was scanned using the 4MHz probe fitted with an angle beam membrane water wedge and an example of the resultant scan is shown in Figure 87. No crack indications were detected at the weld toe location however several irregular indications were detected mid-way through the wall thickness as shown in Figure 88. The manufacturer (Chemresist) was able to confirm that these were due to the manufacturing process causing a certain amount of inter-wrap lack of fusion which is acceptable.
Tank WL 2R
The construction of this tank is the same as WL 23 and so the weld position was plotted in the same way. During this investigation the wall thickness was determined to be approximately 35mm thick. Figure 89 shows the weld position plotted onto the outer surface of the tank wall. The sample length of weld plotted out as shown in Figure 89 was scanned using the 4MHz probe fitted with an angle beam membrane water wedge and an example of the resultant scan is shown in Figure 90. No crack indications were detected at the weld toe location however several irregular indications were detected mid-way through the wall thickness as shown in Figure 91. The manufacturer (Chemresist) was able to confirm that these were due to the manufacturing process causing a certain amount of inter-wrap lack of fusion which is acceptable.
Tank WL 22
The construction of this tank was as expected as shown in the left side diagram in Figure 85. A zero degree UT probe was used to confirm the wall thickness as approximately 17mm. The tank is shown in Figure 92. This tank was scanned using both the 4MHz probe fitted with an angle beam membrane water wedge to inspect the tank wall and the PolyTank 2.25MHz probe without a wedge to inspect the tank base plate. An example scan of the tank wall is shown in Figure 93, no defects were found. The scan of the tank base plate was not recorded, however a small indication at a range of 62mm was noted, which is beyond the expected range for the base plate fillet weld toe (45mm) and therefore unlikely to be due to cracking. Univar took note of this indication.
Blind trials
The blind trials were carried out at the premises of Chemresist in Dewsbury, UK. The objective of the blind trials was to demonstrate the capabilities of two inspection techniques in a site environment to successfully detect discrete artificially induced cracking in the tank wall to base plate fillet welds. The two techniques were:
• T-joint fillet weld tank wall toe inspection for cracking.
• T-joint fillet weld tank base toe inspection for cracking.
To simulate discrete cracks in the inner fillet weld toes (tank wall and base) slots of approximately 3mm deep and between 30mm and 120mm in length were cut into the weld toe as shown in Figure 94. The defects were introduced into two different areas, each containing 3 defects in the wall weld toe and 2 defects in the base weld toe. Area 1 is shown in Figure 95 and Area 2 is shown in Figure 96. Each area was marked out for scanning as shown in Figures 97 and 98. Table 9 shows the results of the inspections and figures 99 to 107 show the indication for each defect. All the artificial defects were detected.
The series of field trials successfully demonstrated the capabilities of the Polytank PAUT system to detect in-service cracking in the toes of the internal fillet welds. The techniques to detect production fusion flaws were not demonstrated during the field trials; however, the system validation inspections have successfully demonstrated these.
Potential Impact:
The expected impacts that the PolyTank project will have on the EU can be divided into three categories:
- Economic
- Societal
- Environmental
Economic impacts
The implementation of the results of the PolyTank project will result in reduced capital cost of many tanks and tank systems by facilitating greater use of thermoplastics as opposed to steel and other metals. It will also reduce repair costs by enabling the identification of defective joints before entry into service. Maintenance costs will be reduced by facilitating greater use of thermoplastics, which generally have a greater resistance to the stored chemicals and consequently longer operating lives than currently used materials. Losses of product from tank failure will be avoided as well as costs of emergency repairs and consequential losses from tank leaks. There will be a reduced risk of leakage of environmentally damaging fluids and consequent costly repairs and clean-up costs and reduced risk of catastrophic failure resulting in loss of life and consequential economic costs particularly for hydrocarbon and acid carrying tanks.
Societal impacts
There are advantages to the community in the use of thermoplastic tanks over the conventional metal storage tanks. Many corrosive chemicals (e.g. hydrofluoric acid) are more safely stored in thermoplastic tanks, often in areas of high population density. Chemicals of this type are not only safer to store in thermoplastic tanks but it is also cheaper and requires less maintenance. The societal value is therefore in getting a safer and cheaper storage of toxic and corrosive chemicals when compared to storage in steel tanks. Also, because the developed PolyTank systems and techniques will be applied from the outside of the tank, there will not be a need for man entry into confined spaces for pre-cleaning or testing of tanks that can often hold hydrocarbons or toxic chemicals. It is anticipated that the project will create new EU employment opportunities (through new knowledge generated from PolyTank), which will be passed on to the EU SMEs operating in the thermoplastic tank manufacture, installation, repair/maintenance and inspection markets. In addition, funding from the FP7 will propel EU SMEs to the forefront of in-service monitoring of thermoplastic tanks and other similar components, thereby enabling other industries to benefit from increased product reliability (e.g. tanks used for liquids transportation).
The results of this project will lead to an improvement in the level of skills and knowledge for European SMEs in NDT, thermoplastic welded products (including storage tanks) and materials technology. The introduction of ultrasonic NDT technology in the thermoplastic storage tank industry will represent a step change. This will necessitate training of operators of the ultrasonic NDT equipment and procedures. Currently, there is a shortage of NDT inspectors in the EU and member states are investing heavily in this area. The results of PolyTank will contribute to this effort.
Environmental impacts
The use of ultrasonic NDT on welded joints (circumferential and helix welds and nozzle welds) in thermoplastic tanks as a quality assurance method will lessen the risk of failure and increase confidence in the use of these materials. Due to the demonstrable evidence of thermoplastic tank weld integrity that will result from the PolyTank project, this will lead to the more widespread application of thermoplastic storage tanks within the process industry sector, and will allow thermoplastic tanks to be used for more stringent service applications in the future. Failures in storage tanks can range from the environmentally unacceptable to the catastrophic (Stonehill J, Bainbridge H and Heyes PF: “Specification and inspection of thermoplastic storage tanks‟. HSL Report HSL/2006/21, 2002). The referenced HSL document describes failure and likely defect mechanisms that include in no small part welding defects, and the report recommends that thermoplastic tanks should be inspected (at this point in time by vessel entry) within the first 2 years of operation, this being a clear indicator of the concern of the HSL report as to the integrity of this type of tank for service. Furthermore the HSL report states that there is considerable scope for further examination and consideration of methods of inspecting and monitoring thermoplastic tanks. Leaks of hydrocarbons, chemicals and processed liquids from storage tanks can be potentially harmful to the environment, local workers and residents and is therefore totally unacceptable. Leaks of this type must be rendered as extremely unlikely events and therefore the integrity of storage tank welds throughout their life must be carefully monitored and understood against realistic acceptance criteria that will be developed as part of the project. The direct impact of the project on working and living conditions is limited, but will accrue from reduced disruption to industry, making the workplace safer by reducing hydrocarbon toxic chemical spillages. Minor leakage from water storage tanks for example has been accepted in the past, but is increasingly penalised as water demand rises and concerns grow for the amount of water abstracted from rivers and boreholes.
Main dissemination activities
Project website
The project website, http://www.polytank.eu/(odnośnik otworzy się w nowym oknie) was launched on 26 March 2013 and contains the following pages:
Public facing area
• Home – Welcome
• About the project
o Project background
o Project concept
o Strategic objectives
o Work packages
• Project Partners
o CHEM RESIST
o TAB
o ACUTECH
o INNOTECUK
o TWI
o HESSEL
o HSL
o UNIVAR
• Publications
• News
• Videos
• Contact us
Partners’ area (Consortium only)
• Project notices
• Project documents
• Project contacts
• My account
• Sign out
Admin only menu
• Logs: User access
• Logs: Database usage
• Add new item
The website also contains the project logo, which is shown in Figure 108.
Project flyer, video and banner
The PolyTank flyer was produced in 2013 (Figures 109 and 110) and is available as a download from the project website (http://www.polytank.eu/publications.jsp(odnośnik otworzy się w nowym oknie)). The text for the flyer is available in English, Dutch, French and German. A video has been produced outlining the project and introducing the project partners. This is hosted on YouTube and is linked from the project website http://www.polytank.eu/videos.jsp(odnośnik otworzy się w nowym oknie). A final project video has been produced, including footage to illustrate the welding procedures, inspection procedures and the equipment design. A freestanding banner has also been produced (Figure111).
Conferences
The PolyTank project has been presented at the following conferences and events:
• 53rd Annual Conference of the British Institute for Non-Destructive Testing, 9-11 September 2014, Manchester, UK.
• European Conference on Non Destructive Testing. 6-10 October 2014, Prague, Czech Republic.
Articles
The following papers and articles about the PolyTank project have been published:
• http://www.sensorsportal.com/HTML/DIGEST/january_2013/PolyTank_TAB.htm(odnośnik otworzy się w nowym oknie) Sensorsportal e-digest newsletter January 2013
• http://www.twi.co.uk/news-events/news/2013-04-polytank-project-to-develop-safety-critical-approach-for-testing-welded-joints-in-thermoplastic-storage-tanks/(odnośnik otworzy się w nowym oknie) Published on TWI website April 2013
• http://www.mundoplast.com/noticia/proyecto-polytank-sobre-testeo-juntas-tanques-plastico/70151(odnośnik otworzy się w nowym oknie) Published in ETD Prensa Profesional (MUNDOPLAST)
• http://www.europeanplasticsnews.com/subscriber/headlines2.html?cat=1&id=2791(odnośnik otworzy się w nowym oknie) European plastics news April 2013
• http://www.sensorsportal.com/HTML/PolyTank.htm(odnośnik otworzy się w nowym oknie) Web page
• http://inspectioneering.com/tag/storage+tanks(odnośnik otworzy się w nowym oknie) Inspectioneering
• Facebook May 2013
• Google +
• LinkedIn
• Welding and cutting. Technical Journal. Issue 03, 2013
Sensors & Transducers e-Digest, Vol. 148, Issue 1, January 2012: Products News http://www.sensorsportal.com/HTML/DIGEST/january_2013/PolyTank_TAB.htm(odnośnik otworzy się w nowym oknie)
• http://www.polytank.eu/publications/Flyer_PolyTank_v3.pdf(odnośnik otworzy się w nowym oknie) Flyer
• Introduction to PolyTank video http://www.polytank.eu/videos.jsp(odnośnik otworzy się w nowym oknie)
• Flyer Distributed at SENSORCOMM' 2013 conference, 25-29 August 2013, Barcelona, Spain.
• TWI connect magazine. Issue 184 – May/June 2013. http://www.twi.co.uk/news-events/connect/connect-may-june-2013/(odnośnik otworzy się w nowym oknie)
• Cordis Wire https://cordis.europa.eu/wire/index.cfm?fuseaction=main.viewrelease
• Silobreaker http://news.silobreaker.com/polytank-project-to-develop-safetycritical-approach-for-testing-welded-joints-in-thermoplastic-storage-tanks-5_2267122038646767662(odnośnik otworzy się w nowym oknie)
• Flyer Circulated to ten UK tank manufacturers
• March 2014 website updates include: addition of new partner info, NIKK, changes to existing partner info, changes relating to EU guidance on logos and funding statement
• 48 companies contacted using the project flyer
Dissemination seminars
• “Vorstellung des Forschungsprojektes Polytank zur zerstörungs-freien Überprüfung von Schweißnähten mittels Ultarschall” (Presentation of the Polytank research project for non destructive inspection of welding seams by ultrasonic testing) Thermoplastics for the chemical processing industry Seminar organised by Rochling. 6 June 2013, Haren, Germany.
• Kunststoffe im Anlagenbau – Plastics in Process Plant. TÜV Süd Seminar. 25-26 February 2014. Munich, Germany.
• NDT of Welds in Thermoplastic Tanks - NDT von Schweißnähten in Thermoplastbehältern. DVS AG W4.3b 9 April 2014, Minden, Germany
• New developments in NDT of Plastics and Composites. TWI Technical Group Meeting. 23 May 2014, Chesterfield, UK
Exploitation of results
The main exploitable products from the PolyTank project are:
• Phased array probes
o Ultrasonic low frequency probes for testing of high attenuation thermoplastic welds.
• Ultrasonic instrument
o Ultrasonic remote site data acquisition/display instrument and techniques for volumetric inspection of thermoplastic welds.
• In-service inspection of PE tanks
o Flaw acceptance criteria.
• Prototype inspection system
o Automated manipulator scanning system.
List of Websites:
http://www.polytank.eu/(odnośnik otworzy się w nowym oknie)
Plastics, such as high density polyethylene (HDPE), polypropylene (PP), polyvinylidene fluoride (PVDF) and polyvinylchloride (PVC) offer significant advantages over metals for the storage of chemicals such as acids, caustic soda, detergents and other corrosive liquids. They are chemically inert, tough, flexible and lightweight. However, plastics also have relatively low maximum operating temperatures and loads, which vary from plastic to plastic, as does chemical resistance. For this reason it is important that plastics tanks and storage vessels are designed and built to the relevant European standards.
It is estimated that there are over 50,000 plastics tanks in service in Europe, ranging in size from 150 litres to 100m3. Many of these store materials that are classified as hazardous and therefore have the potential to cause injury or death to personnel and damage to property and the environment in the event of a tank failure. Indeed, there have been a number of reported failures of plastics tanks in service.
Plastics tanks are normally designed for a finite life, usually between 15 and 25 years. However, due to economic pressure, many of these tanks are still in operation beyond their design life, often with little or no engineering justification. It is also not uncommon for plastics tanks to be used for storing chemicals that they were not designed to contain. For these reasons it is very important that operators of plastics tanks and vessels inspect them throughout their life. An issue at hand is that there are currently no standards for the in-service inspection of plastics tanks. There is also very limited expertise available on the visual examination of plastics tanks and virtually no use of non-destructive examination (NDE).
The majority of visual inspections are external and can therefore only identify cracks that break the outside surface of the tank. Since many of the cracks initiate from the inside of the tank there is already a leak path through the tank wall if and when the crack is detected. Internal inspections are carried out less frequently, if at all, because they are expensive, potentially dangerous to the inspector, and result in a shut-down because the tank has to be emptied.
The PolyTank project has determined the potential failure mechanisms in plastics tanks and storage vessels and developed ultrasonic NDE procedures, techniques and systems to be able to identify these. The project developed an inspection system that is able to be deployed on-site and is relatively simple to operate.
As part of the project, welded joints representative of those used to fabricate plastics tanks and vessels were manufactured containing known flaws. These were inspected and the NDE data analysed to determine the limits of flaw detection. In parallel, the significance of flaw size and quantity was established in relation to service requirements. This was achieved by long-term mechanical testing of joints containing known flaws, and comparison with results for welds containing no flaws.
The prototype NDE equipment, designed and built as part of this project, was successfully assessed under both laboratory and field conditions, and the inspection procedures developed will be submitted to CEN as the basis for producing a European standard on this subject.
Project Context and Objectives:
There are around 300,000 medium-to-large industrial storage tanks in the EU, mainly in the petrochemical, process plant, pharmaceutical and food processing sectors, of which about 60,000 are made of thermoplastic materials containing safety critical welds that are currently impossible to fully examine through the full body of the weld because there is no suitable NDT technique commercially available. Despite their economic, safety and environmental advantages the market growth for thermoplastic tanks into new and more demanding applications will not be realised unless a suitable NDT method can be commercially developed to examine the full volume of the weld. Virtually all medium to large thermoplastic storage tanks are fabricated using containment welds that include the base to the shell of the tank, shell helix welds or circumferential welds (depending on the design type) and inlet/outlet nozzles and instrumentation penetrations. All of the welds (except the shell helix welds) are made manually and are therefore more likely to contain manufacturing defects that threaten the structural integrity of the finished tank either in the short term or during its service life.
The current practice of visual inspection includes the tank inner surfaces, which requires tank preparation (opening, cleaning, purging etc.) that can take up to two weeks and means that normally only one or two tanks can be taken out of service at any one time, thus limiting the number of tanks that can be realistically inspected at one location.
Thermoplastics are relatively new structural materials and they provide significant challenges for NDT. In particular, these materials are acoustically very opaque. For more demanding service applications, such as higher operating temperatures and pressures, there has been a lack of confidence in the long-term reliability of these materials and this has restricted the use of welded thermoplastic tank systems in these applications, despite their obvious benefits, including their inertness to most chemicals, corrosion resistance, ease of fabrication and cost. However, because of the benefits for use in-service, there is a renewed interest in the use of thermoplastic tanks for these more critical applications provided that applied inspection can be shown to give a high confidence level in the welded joint integrity and that of the adjacent parent material. Added to this is the need to develop a realistic defect acceptance standard for a range of in-service thermoplastic welds and that the defect can be reliably detected to the acceptance criteria.
The two main techniques for welding thermoplastic tanks are extrusion and hot gas welding, both of which are manual processes. Extrusion welding is used for joining thermoplastic sheets of wall thicknesses greater than around 6mm or for long joint lines (greater than 5m), whereas hot gas welding is mainly used for joining thinner sheets or short, complex joint geometries. Extrusion welding (Figure 1.2) involves continuously extruding molten thermoplastic (extrudate), of the same material as the parts being welded, into a prepared joint between the parts, using an extrusion “gun”. The joint is preheated to its softening temperature by a stream of hot air before the extrudate is forced into the joint under pressure. This ensures that the extrudate and the parts fuse together and produce a weld.
The inspection technique proposed in PolyTank is for an ultrasonic examination of the full weld volume to take place from the outside surfaces of the tank, which means that it is therefore not necessary to open up the tank and prepare the inside for examination. Furthermore, the proposed ultrasonic inspection is far more searching as it is a full volumetric examination of the tank welds that, until now, have not been possible to inspect. The potential EU annual savings on tank preparation alone that would result from the project are €120million, based on a five-year inspection interval for 50% of the available tanks. Add in the other potential savings to EU industry and generated profit for the EU SMEs as a result of the project, then the total potential annual benefit could be as great as €165million per year, giving a “return on investment‟ [ROI] of 110:1 per annum based on project eligible costs. The ability to inspect critical welds in thermoplastic tanks in-situ and without disruption to normal operations will facilitate an increased market for the use of thermoplastic storage tanks in the process industry in preference to the more expensive and problematic steel tanks.
Three previous EC projects, WINDEPP, Polytec Systems and TestPEP, developed the basic concept of the ultrasonic inspection of butt fusion and EF welds in plastics pipes, respectively and went on to develop an NDE system for inspecting pipes in a wide range of sizes and between pipes and fittings as well as pipe-to-pipe joints. However, these projects only investigated the inspection of welds in PE pipes. The PolyTank project builds on the work carried out in these earlier projects and applies the techniques developed for pipe joints to the type of welded joints used in storage tanks.
This project has developed and optimised methods for inspecting tanks made from HDPE. The ultrasonic property of this material is unknown for the transmission of ultrasonic energy at angles non-normal to the outside surface of the tank scanning surface. This has been restricting the application of ultrasonic inspection and required a concentrated research programme to enable suitable ultrasonic techniques, procedures and equipment to be developed that will enable tank owner/operators to have a high level of confidence in the integrity of the welded joints. Furthermore, the welds in thermoplastic chemical storage tanks, can generate planar flaws with complex geometric problems to solve. Several angled ultrasonic beams (normally refracted shear wave angles of 45°, 60° and/or 70°) are generally used in weld testing in order to align the beam as near as possible to be at right angles to the main face of the defect in order to get the maximum reflection from the defect possible thus maximising the probability of detection. In specifying the angles to be used it is usual to consider the most likely orientations of anticipated potential defects, e.g. weld side wall fusion defects, root defects etc. However, the low velocity of the ultrasound creates a serious technical challenge in generating angled ultrasonic beams in thermoplastic materials. Angled ultrasonic beams are generated by refraction from a slower medium. There are very few materials that will support low velocity ultrasound and this challenge has restricted the application of ultrasonic NDE for thermoplastics.
Scientific Objective
Increased scientific understanding of the transmission of ultrasonic waves through thermoplastic materials to enable the detection of the required defect sizes in all joint types found in thermoplastic tanks with wall thicknesses in the range of 10mm up to 80mm.
Technical Objectives
A high power ultrasonic transducer array(s) with a system gain of 120dB and a signal/noise of 80 - 100dB with signal processing, using ultrasonic angled compression wave techniques at a frequency range of 0.5MHz to 5MHz that will be significantly attenuated than the more conventional and commonly used shear waves.
Creation of a knowledge database of up to 100 datasets allowing unique failure curves of defect size against creep rupture life to be plotted for different tank materials and joint configurations.
Development of flaw acceptance criteria for welded joints in storage tanks made PE thermoplastic materials.
Integration Objectives
Development of a pre-production prototype ultrasonic inspection system capable of inspecting welded joints in thermoplastic tanks up to 10m diameter and 80mm wall thickness.
Project Results:
Introduction
The PolyTank project is divided into six technical work packages:
Work Package 1: Manufacture of Welded Joints
• To produce a range of welded samples in the materials and types of joint that are representative of those present in thermoplastic chemical tanks currently used to store bulk liquids in industry.
• To insert into these samples a range of simulated flaws of known size and shape that are representative of real flaws found in service.
Work Package 2: Development of Ultrasonic NDE techniques
• To measure the physical ultrasonic properties of the thermoplastic materials specified in WP1.
• To develop inspection procedures and determine the limits of detection for each joint geometry and material specified in WP1.
• To develop data processing and analysis algorithms and software.
• To design and manufacture ultrasonic probes and probe wedges.
Work Package 3: Development of Flaw acceptance criteria
• To determine the effect of different types, sizes and concentrations of flaws on the long-term performance of welded joints in thermoplastic bulk liquid chemical tanks by carrying out accelerated mechanical creep rupture test.
• Based on the mechanical test results, develop a flaw acceptance procedure for the agreed range of flaws likely to be present during the service of the target thermoplastic welded joints.
• Compare the results with the NDT ultrasonic results to confirm that the flaw sizes detailed in the acceptance criteria can be detected using the developed ultrasonic technique using an array of transducers.
Work Package 4: Integration and Validation of Ultrasonic Inspection System
• To design and build a modular tank inspection for easy deployment on thermoplastic tank joints.
• Integrate the complete ultrasonic inspection system including the manipulator holding and scanning the transducer array probe head assembly, remote ultrasonic site, ultrasonic data acquisition and evaluation instrument and, control software, data analysis software and software for delay law formulation for the ultrasonic beam angle generated by the ultrasonic phased array.
• Validate the completed PolyTank system in the laboratory.
Work Package 5: Field trials using validated Ultrasonic Inspection System
• Prepare On-site Ultrasonic Inspection Procedure and Technique Sheets for specific weld geometries.
• Carry out Field Trials to assess the performance of the validated prototype inspection system and on-site ultrasonic inspection procedures.
WP1 - Manufacture of Welded Joints
Welds to be made in the project
A list of the welds to be made in the project was developed, taking into account the following:
• Limited budget available for welding and testing
• Limited budget available for materials
• Weld geometries and material thicknesses representative of the plastic tank industry
The extruded sheet material was HDPE, (Polystone® G-Natur HD), supplied by Röchling, Germany. Three material thicknesses were studied, 15mm (natural), 25mm (natural) and 40mm (black). It was originally planned to study 50mm, but this was not felt to be representative of the tank industry. The pipe was PE100, 110mm SDR11 (black), and was manufactured by Fusion Provida. The welding rod was the same grade as the sheet, 3 and 4mm diameter and yellow in colour. The granules were supplied by Röchling and the rod was manufactured by Hessel Ingenieurtechnik.
Four joint geometries were studied.
• Double-V butt
• T-Joint
• Cruciform
• Pipe outlet
Three flaw types were investigated:
• Lack of fusion between extrudate and parent material.
• Cracks in parent material and through the extrudate.
• Cold weld between extrudate and parent material.
TWI has produced a range of welded joints containing various idealised flaws (cold welds, embedded aluminium discs and partially embedded aluminium discs) using the procedures below. The joints were made in various configurations: double–V butt (Figure 1), T-Joint (Figure 2), cruciform (Figure 3) and pipe-in-sheet (Figure 4). Some of the joints contained no deliberate flaws and will be used as the reference welds. A number of duplicate welds were also made to enable the limits of detection and the critical flaw sizes to be determined more accurately. The total number of welds to be made is 120 and this is shown in Table 1.
An aluminium base plate, end plates and support bars were designed and manufactured to clamp and support the sheet for manufacturing T-joints and cruciform joints (Figure 5). A wooden base plate and steel clamping bars were designed and manufactured to hold the sheets for manufacturing the double-V butt joints (Figure 6). A fixture was manufactured to ensure accurate pipe positioning and support for manufacturing the pipe-in-sheet joints. This is shown in Figure 7.
The list of welds made is given in Table 2.
The manufacture of pipe-in-sheet joints is shown in Figure 8 and a photograph of some of the completed joints is shown in Figure 9. A photograph of a 40mm thick double-V butt joint being manufactured is shown in Figure 10 and photographs of completed double-V butt joint samples are shown in Figure 11. A photograph of 40mm thick T-joints being made are shown in Figure 12 and a photograph of completed 25mm cruciform joints is shown in Figure 13.
Double-V Butt Joints
A common type of edge preparation for a butt weld is a double-V configuration. This is shown in Figure 14. For all material thicknesses, the double-V edge of the sheet was prepared using a CNC-milling machine. The angle of the preparation was 60° inclusive.
T-joint and cruciform joint
T-joints are widely used in the fabrication of plastic tanks. For example, a T-joint would be used for joining a tank body to a base plate. Cruciform joints are not common in the fabrication of plastic tanks. However, they are being made in the project because the long-term mechanical test to assess the effect of flaws in the fillet welds requires a cruciform joint rather than a T-joint. The T-joints will therefore be used to assess the NDE technique and the cruciform joints will be used to determine the flaw acceptance criteria. An example of a cruciform joint is shown in Figure 15.
Pipe-in-sheet
The pipe-in-sheet weld configuration was chosen to represent a pipe inlet/outlet in a tank. This is a common configuration for joining pipework to tank bodies. An example of a pipe-in-sheet trial weld (with partially embedded flaws) is shown in Figure 16.
Flaw Insertion Procedures
Planar flaws in the weld were simulated using aluminium discs, which were made out of 0.025mm thick aluminium foil with an adhesive backing, and with a purity of 99%. They were supplied by Goodfellow UK. 2, 3, 4, 8, 15 and 25-mm discs were punched out from foil.
Hole punches and dies were manufactured so that aluminium discs with repeatable and consistent dimensions could be punched out of the foil. The hole punches and dies are shown in Figures 17 and 18.
Procedure for inserting embedded aluminium discs
1. Lay two pre-machined lengths of HDPE sheet on to a flat bench and secure in place.
2. Using vernier callipers, measure from one edge of the HDPE sheet along the preparation to determine the location of the aluminium disc placement.
3. Use a scalpel blade to carefully remove the backing paper from a pre-punched aluminium disc to expose the adhesive layer (Figure 19).
4. Place the aluminium disc in the centre of the preparation using a pair of tweezers as shown in Figure 20.
5. Using the rubber end of a pencil, gently press the aluminium disc against the HDPE surface until the adhesive adheres to the surface producing as smooth a finish as possible (Figure 21).
6. For each of the aluminium discs, mark its position and size using a white correction pen (on black sheet) or permanent marker (on natural sheet) on the adjacent surface of the HDPE sheet in a location that will not be obscured after welding. This is shown in Figure 22.
Procedure for inserting partially embedded aluminium discs
1. Place the pre-cut lengths of sheet into the welding jig as shown in Figure 23.
2. Tack the sheets together using a hot air welding torch as shown in Figure 24. The sheets should be tacked along their full length to ensure that they don’t move during the welding operation. The welding torch temperature should be calibrated to 315°C using a Calstik temperature calibration device.
3. Once all of the joints have been tacked, carefully remove the tacked assembly from the welding jig.
4. Using a steel ruler and the sharp point of a compass and mark the predicted distance of the extrusion weld width from the weld root tack on the HDPE sheet. Draw a line between the points using a permanent marker or white correction pen Figure 25.
5. Using a steel ruler measure the embedded distance from the marked weld width line (i.e. towards the weld root), depending upon the required disc insertion depth. Mark this position with the sharp point of a compass (the mark should be visible without causing significant indentation) and draw a line parallel to the first line.
6. Use a scalpel blade to carefully remove the backing paper from a pre-punched 25mm diameter aluminium disc to expose the adhesive layer (Figure 19).
7. Place the aluminium disc so that its edge is on the insertion depth marked line, as shown in Figure 26.
8. Using the rubber end of a pencil, gently press the aluminium disc against the surface of the HDPE sheet until it adheres, producing as smooth a finish as is possible.
Transverse cracks
To simulate cracking across fillet welds in polyethylene tanks, which can be generated by chemical attack, the following flaw insertion procedure should be carried out:
Procedure for producing Transverse cracks
1. Mark on the sheet, with a white correction pen or permanent marker, the required location of a crack.
2. Load a 0.15mm slitting wheel (Figure 27) on to a CNC milling machine as shown in Figure 28.
3. Secure the welded sample to the CNC milling machine using machine clamps.
4. Insert the rotating slitting wheel into the fillet weld to the required depth using the CNC display (Note: this operation should only be performed by a skilled machine operator).
5. Once the crack has been made, mark on the sheet, using a white correction pen or permanent marker, the final location of the crack (as the crack can be hard to detect visually).
Cold weld
Preheating the parent material using hot air is essential for ensuring that a good quality extrusion weld is achieved as shown in Figure 29. However, due to a lack of knowledge, extrusion welded tanks are sometimes fabricated without the correct preheat being applied and in some circumstances with no preheat. To simulate this type of flaw a preheat deflector plate was manufactured and retro-fitted to the extrusion gun. This is shown in Figure 30.
Procedure for producing cold welds
1. Before carrying a cold weld attach the deflector plate to the Leister S2 extrusion gun nozzle using screws (see Figure 31).
2. Set the pre-heat (260°C) and extrudate (230°C) temperatures on the extrusion gun.
3. When the pre-heat and extrudate temperatures have been reached, carefully check that the hot airflow is being directed away from the weld preparation.
4. Produce the weld.
WP2 – Development of Ultrasonic NDE Techniques
Basic Material Properties
In Non-Destructive Testing (NDT) using ultrasound an important task is to create the correct inspection configuration. This task involves selecting the probe, the coupling medium, the scan plan, the scanning system, etc. These decisions are based on the specimen to be investigated, both geometry and material. The polyethylene plastic materials which were investigated in this project are known for their very slow acoustic velocity and highly attenuating nature. However, the exact properties of the material will vary depending on for example the material grade. Therefore, these properties need to be determined for each material to be tested. The determined information about the ultrasonic properties will then be used to design the inspection configuration; selecting probe parameters; and selecting coupling solution.
Acoustic properties
In ultrasound there many different modes in which the sound travels. The two most common modes in NDT are longitudinal and transverse waves. In plastic pipes only longitudinal waves propagate significantly. Transverse waves are highly attenuated due to the properties of the material. The acoustic properties of PE, such as velocity and attenuation of longitudinal waves and their dependency on frequency need to be well known. These ultrasonic parameters are necessary for development of NDT techniques, configuration of NDT equipment, detecting and measuring wall thickness and flaws, and for modelling which is used during inspection technique development. The schematic drawing of the setup used to acquire the data required for the determination of the acoustic properties can be seen in Figure 32. A pulse-echo configuration was used where the same probe was used to transmit and receive the ultrasonic signal through a 30mm thick plastic sheet. A number of phased array probes with different frequencies were used to get the acoustic properties and their dependency on frequency. The frequencies used were 2MHz, 5MHz and 10MHz.
Velocity
In this project plastic sheets from High Density Polyethylene (HDPE) were used. The ultrasonic properties needed to be determined for this material to be able to develop and calibrate the inspection techniques. The ultrasonic A-scan was recorded for each probe used; see Figure 33 for the measured A-scan using the 2MHz probe. By measuring the time it takes for the ultrasonic signal to travel through the material and mechanically measure the thickness of the material the velocity can be calculated as
c = dm/tu
where dm is the mechanically measured thickness of the sample and tu is the ultrasonically measured time taken for the ultrasound to travel through the sample. For all the frequencies the velocity was determined to be 2400m/s. For the probes used and the material under investigation, no effect of frequency dependency on the velocity was established.
The possible difference in velocity in the parent plastic material and the weld was also investigated. Figure 34 shows the response received looking through a 70mm thick sample where some part of the signal travels through the weld and some part travels through the parent material. The signal from the back wall when travelling through the weld is slightly delayed compared to the signal travelling through the parent material. This indicates that the velocity is lower in the weld. The velocity in the weld was calculated to be 2356m/s. However, the decrease in velocity is less than 2% and will not have any effect on selecting the inspection configuration.
Attenuation
The same experimental configuration that was used to get the velocity was used for the determination of the attenuation. Figure 35 shows the measured ultrasonic attenuation in relation to the expected probe frequency. The attenuation is measured as decibel per millimetre. It can be seen that the attenuation increases rapidly with increasing frequency. Furthermore, usually the actual frequency of the probe is less than the labelled value for higher frequency probes. This would make the attenuation curve even steeper.
The results from the attenuation measurements show that selecting the probe frequency is important. When inspecting thicker components, where the propagation distance increases, a lower frequency is required for the sound to propagate the required distance.
Effect from ageing
Theoretical calculations indicate that the effect on acoustic properties due to ageing is very small. The density and the elastic modulus could vary but the resulting variation in the velocity will be very small. This variation will have no effect when designing the inspection techniques and experimental configurations. To verify the theory, initial measurements have been taken on a specimen that will be subjected to ageing before final measurements can be taken.
Phased Array Transducers
The geometrical structure of the different welds in the polyethylene tank and the acoustic properties of the material, together with the requirements of a rapid and robust system, demand a phased array device. The fact that shear waves do not propagate in polyethylene means that we are restricted to using longitudinal waves. When utilising phased array technology there are options to use linear, circular, annular, etc. arrays, but in this application a linear array is most appropriate considering the measurement area and also enabling steering capabilities. A one dimension (1D) array has been used, where the benefits over a 2D array are; less computationally heavy; rapid calculations; lower cost; and not as demanding on the ultrasonic instrument.
The different joint configurations require different inspection solutions and hence different acoustic devices (phased array transducers). Designing phased array transducers is always a compromise between selecting the proper pitch, element width, aperture, and number of elements. A high number of small elements increases steering, reduces side lobes and provides focusing but the drawbacks are the cost of manufacturing and potential instrument complexity. Separating elements with a greater distance will increase the aperture size, but this creates unwanted grating lobes. Figure 36 shows an illustration of a phased array and some of its properties.
The transducer features have been selected based on results regarding the attenuation and velocity in the materials, the geometrical limitations from the weld to inspect and practical limitation from the array controller. Some of the transducer features are the dimensional parameters in Figure 36, and others are centre frequency, bandwidth, etc.
The total number of channels available in the array controller is 128. However, for this project the numbers of elements used were between 8 and 32. The pitch, gap and width were defined to be able to create the proper steering angle, focal depth, focal strength and aperture size. These properties were defined after experimental and modelling work. The objective was that the centre frequency was in the range 1-5 MHz The height of the aperture must be large enough to create the focal point at the desired depth.
Figure 37 and 38 shows curves that can be used as guidelines when determining the values of some of the parameters. One important property is the near field range, which is the greatest distance as a focal point can be created using the array, and be approximated as
,
where; f is the frequency, c is the speed of sound in the material and D is the aperture size. For a linear phased array, the aperture size is D=n’p, where n’ is the number of active elements and p is the pitch. Since the near field is proportional to the array aperture and a fixed number of active elements will be used, the pitch will be the deciding parameter. A larger pitch will create a higher focal length. However, the pitch is also determining the focal strength and the prevention of grating lobes. Ideally, the pitch should be less or equal to half the wavelength, λ/2, where λ=c/f. Using 32 elements result in the possibility to choose a pitch size from the dark areas in Figure 37 and 38. Fortunately, the requirement of half the wavelength can be relaxed, increasing the possibilities slightly. For a smaller tank with a thinner wall thickness, only requiring a near field range of about 40mm, frequencies up to 5MHz can easily be used without violating the requirement of half the wavelength, see Figure 38. Therefore, the selection of frequencies can be based on attenuation properties described above.
Optimising Transducer Parameters
Another factor that needs to be taken into consideration is the propagation distance. This is highly affected by the attenuation in the material and the frequency selected. To cope with this beam modelling was used together with information about the geometrical constraints. Beam field modelling is conducted to assure that the beams generated by the array in the housing are performing in a desired way. In order to do this, the SimulUS software was used. This software can be used to minimise side lobes and to make sure the desired propagation distance is achieved. Steering at higher angles will always push the limits of the transducer and side lobes will be generated. However, these need to be minimised in order to get more energy to the desired inspection region, see Figure 39 for examples of high angle steering for both the 4MHz and the 2.25MHz transducers.
An important factor when developing inspection techniques and optimising the hardware, i.e. transducers and wedges, to be used, is the geometrical constraints. The developed inspection technique will tell where the transducer will be placed and, hence, give indication on the physical footprint of the transducer. For the butt weld, the inspections will be carried out with the transducer placed on the side of the weld. The area next to the weld is smooth and does not have any direct limitations in size, which help the transducer parameter optimisation. However, for the fillet welds, the optimal position of the transducer is on the weld cap. This limits the maximum size of the transducer, which needs to be taken into consideration. Several linear 1D transducers with frequencies between 1-5MHz have been used in the project. The parameters for the transducers are presented in Table 3. The number of transducers for each joint type has been kept at a minimum to make the system as flexible as possible. The optimum would have been one transducer for each joint size, but the wide range of sizes and type of defects to detect covered in this project did not allow that. The transducer parameters have then been optimized in terms of resolution and coverage and the optimum solution is presented in Table 3.
Transducer Specifications
The two transducers TJ1 and BW1 have been optimised to inspect the weld joints identified in WP1 to detect the types of flaws inserted into the welded samples made during WP1. The transducer BW1 is available at TWI and has been used in the project to cover the BW joints. The transducer is shown in figure 40. The transducer TJ1 was not available from TWI stock and, following the principles outlined above, a specification for the required transducer was developed. This specification was sent to the relevant suppliers of phased array transducers listed below for quotes for manufacture:
• Sonaxis
• Vermon
• GE
• Zetec
• Imasonic
• Olympus
Quotes were received from three of the suppliers. However, the lead times to supply these bespoke transducers was not acceptable within the project timescales and so an in stock transducer with acceptable parameters was identified and procured. This transducer is shown in figure 41. During the course of the inspection procedure development it was identified by the project steering committee that it would be beneficial to detect an additional type of defect to those already detailed in WP1. This resulted in two additional techniques, one of which could be accomplished with the use of transducer TJ1, but the other technique required a transducer with different parameters to those already detailed. Fortunately another transducer already available at TWI could fulfil those parameters and is detailed in table 3 as TJ2 and shown in figure 42.
Prototype Ultrasonic Probe Wedges
In order to transmit ultrasound from a phased array ultrasonic transducer into the test material, the transducer needs to be coupled with the material using one or more mediums which are acoustically compatible with the material under examination. This will facilitate the transmission of ultrasound and reduce the amount of energy which is reflected at the interface. If the component material surface is flat and smooth and the required angle of the ultrasonic beam is approximately normal to the surface then the probe need only be coupled by the use of a suitable fluid, such as Ultragel, to eliminate air at the interface. However, if a beam of any significant angle from the normal to the surface is required then it is highly advantageous to introduce the sound from the transducer at a pre-defined angle by the use of a wedge made from a material with suitable acoustic properties. Traditionally, for metallic material inspection, the wedge material is often a type of plastic called Rexolite. However, this is not always suitable for HDPE. If the component material surface is uneven then the probe can often be coupled at any angle by using a water path between the probe and the material. A simple way to achieve a water path is to submerge the component and transducer in a bath of water, but this is not always practicable. For site work a new type of wedge has been developed consisting of a wedge with a cavity to hold water for the water path, see figure 43. This may be either open faced, with the water in direct contact with the material surface and a flexible skirt to maintain a seal, see figure 44, or alternatively the wedge could be fitted with a flexible membrane to contain the water path within the wedge.
The butt welds produced have a very similar profile to a conventional steel double V butt weld. Initially an open faced water wedge was proposed, using a technique similar to that developed for HDPE pipe inspection. However, it was found that a standard Rexolite wedge, as used for steel weld inspection, produced very good results. The wedge allowed the transducer to produce compressional beams in the sample material of between 30 and 40 degrees and is shown with the probe in figure 40. Because the plate material is flat and smooth the fact that the Rexolite is inflexible presents no problems in coupling. The technique used to inspect the fillet welds for production flaws is to introduce ultrasound into the weld cap to examine the weld fusion faces. This technique is novel to weld inspection as it is not normally employed in conventional metal weld inspection. Although the HDPE weld surface is usually fairly smooth, it is not very even, in some cases being convex and in others concave. Furthermore, it is usually desirable when applying the inspection technique to introduce the UT beams at a mean angle approaching the normal to the fusion face, which is often at an angle of approximately 35 degrees to the weld cap surface. In order to determine the best design for a water wedge the optimum position of the required probe, BW1, in relation to the fillet welds for both T-joint and pipe insert was modelled and is shown in figures 45 and 46. To accomplish the correct probe angle and coupling a membrane water wedge was designed to accommodate the probe, BW1, and to fit the fillet weld geometry. The wedge was manufactured by a plastic prototyping technique and the final product is shown in figure 47. Although this wedge was designed to closely fit the probe BW1, it was found that it was just slightly too large to successfully couple with the smaller profile of the fillet welds in the 15mm thick sheet and the pipe in sheet welds. To overcome this shortfall a more compact membrane wedge was designed. This wedge was machined from aluminium and is shown in figure 48.
During the course of the project an additional type of defect was identified as being desirable to detect. This is in-service cracking at the toes of the internal fillet welds at the base of the tank. To detect these, two different techniques were developed, one for the weld toe in the tank wall and one for the weld toe in the tank base. To inspect the weld toe in the tank wall the technique involves introducing the sound beam at an angle into the tank wall. The method of construction of the tank wall results in a smooth but uneven surface. To accommodate this an angled membrane water wedge was employed as shown in Figure 49. This wedge was found from existing TWI stock. The technique to examine the base plate weld toe involves direct contact of the PAUT transducer onto the edge of the base plate and therefore no wedge is required.
Inspection Procedures
This section contains a description of the work carried out to develop the basic ultrasonic inspection procedures and techniques to inspect the weld geometries specified earlier. Work has been conducted to understand the geometry of the joints to be inspected. This is important for the inspection configuration to achieve appropriate coverage of the fusion zones, and also for the interpretation of the data from the scanned welded joint samples. Test specimens for each joint, tank material and joint thickness were created for complete development of the inspection techniques. The inspection results of these specimens have been interpreted to evaluate the performance of the inspection techniques.
Ultrasonic probes
Several linear 1D probes with frequencies between 1-5MHz were used in the project, depending on the joint type and the pipe size. The butt welded joints have good accessibility for the phased array probes and the physical size of the probe is not a significant problem. This means that the element size, pitch and gap can be optimised for the best resolution and penetration without considering any physical constraints. a big fusion area and normal 0-degree beams are to be used for the inspection. The T-joints will have restricted accessibility for a phased array probe. The base plate will be cut close to the fillet weld and the tanks usually have an irregular surface, not optimal for a phased array probe. The only remaining location for a probe is then on the weld itself. The probes then need to be physically smaller. The frequency dependent attenuation has been described earlier. The results indicated that if a longer propagation distance is required, a lower frequency of the probe should be used. In the T-joints, due to the long propagation distance, a low frequency probe needs to be used. However, when decreasing the frequency of a probe, usually the physical size of the elements increases. Therefore, the number of elements needs to be reduced. The pipe outlet joints have similar welds as the ones in the T-joints. The same ultrasonic probe will be used for these welds. The parameters for the probes desired in the project are presented in Table 3. The number of probes for each joint type has been kept to one each to make the system as flexible as possible. The probe parameters have then been optimized in terms of resolution, coverage and physical size and the appropriate solution is presented in Table 3.
Wedge
Water wedges have been employed for the inspection of all different joints in PE storage tanks. The advantages with a water wedge are the low attenuation; good acoustic matching with the PE material and the velocity ratio enabling the steering of angled beams to the fusion zone. The main challenges with a water wedge are possible air bubbles and to keep the water between the elements and the material. When designing and manufacturing the wedges, the aim is to make them as physically small as possible, but still allowing for the desired range of angles to be transmitted into the material. The possibility to use membrane water wedges is investigated. A membrane water wedge will have two major advantages; the water can be maintained in the wedge without any loss; the wedge surface is flexible and can be adapted for the surface profile.
Instrument
The variety of ultrasonic techniques used for the inspection in this project sets high demands on the ultrasonic instrument. The instrument that was used to inspect the welded samples in the project was the Dynaray from Zetec with their Ultravision 3 software. This is a very powerful instrument with 256x256 channels, high voltage and good signal-to-noise ratio. All ultrasonic techniques in this project can be implemented with this instrument. However, when integrating the system in the field trials, the Omniscan MX II PA by Olympus with its Tomoview software was employed. This is a portable phased array system that enables site inspections.
Inspection Techniques
Due to the different joint geometries described above, different inspection solutions are required. This includes phased array probe, coupling solution (wedge) and focal law settings in the phased array controller.
Butt welds
The inspection technique used for double V butt welds is shown in Figure 50. A standard sectorial scan is used, where the beam angle is swept between lower angles up to higher angles to ensure appropriate coverage of the fusion zones. By inspecting the weld from both sides, the sector pulse-echo configuration is believed to give a complete coverage of the weld. However, due to the complexity of the structure, caution must be taken when interpreting the scans. Direct pulse-echoes will be received together with signals form a skip on the back wall, which could be hard to interpret. Figure 50 shows the inspection from one side, but to get full coverage the welds need to be inspected from both sides. The lower fusion face is covered by the direct angles and the upper fusion face is cover through a skip on the back wall. A sector scan, using all the elements in the array to create an aperture, sweeping the beam from the lower angle to the higher angle is used. The step between each beam could be 0.5° to increase the resolution. The transmitted beams are focussed at the inner surface distance and Dynamic Depth Focussing (DDF) is used when receiving the beams.
The standoffs for the sector pulse-echo technique are depending on the plate thickness and the weld cap. The ultrasonic probe is positioned with an offset from the weld centreline due to the weld cap. The geometry of the weld cap and the weld will help defining the focal law settings to use. When the thickness of the plate increases the weld cap increases as well and the probe have to be positioned further away from the weld centreline. Therefore, different focal law settings are needed for different thicknesses of plastic plates. The focal law settings for the three different plate thickness used in this project are given in Table 4.
T-joints and pipe outlets
Due to the geometry of T-joints, different inspection techniques are required. Several different solutions, including the probe placed on the base, the tank and finally the weld itself, have been considered to be able to access all possible areas where a defect could be present, see Figures 51 and 52.
Development Results
A 20mm thick plate with aluminium discs inserted was used for the development of the inspection technique for the butt welds. Phased array scans using the setups described above were conducted on the plate. Figure 53 shows a typical S-scan on one location on the plate, with the weld overlay and some descriptive text. Figures 54-56 show three images of different features. In all three figures: A-scan (top left), S-scan (top left), and B-scan (bottom) are shown. The A-scan displays what the cursor in the S-scan indicates at the position where the cursor in the B-scan is located. The B-scan shows the data from the cursor in the S-scan all across the weld. Figure 57 shows the B-scan with all the detected aluminium discs highlighted. Inspection techniques have been developed for the inspection of welds in PE storage tanks. The developed inspection technique for double-V butt welds has proved capable of detecting all inserted aluminium discs in the test piece.
PAUT Detection Limits
Table 5 details the results of the inspections carried out on the Butt weld and T-joint Weld samples. The Pipe-in-sheet samples were not inspected due to the inability to successfully position the probe on the fillet weld. Further probe wedge development is required to achieve an inspection of this complex geometry. There were four types of weld flaw produced:
• Welds with no flaws.
• Welds with discrete flaws of varying sizes ranging from 2 to 25mm diameter.
• Welds with 25mm diameter flaws at varying distances from the root ranging from 2 to 10mm
• Welds with a cold weld element.
Figure 58 shows a typical scan of a butt welded sample with no flaw; Figure 59 shows a typical scan of a butt welded sample with discrete flaws and Figure 60 shows a typical scan of a butt welded sample with a cold weld flaw.
Figure 61 shows a typical scan of a T-joint welded sample with no flaw; Figure 62 shows a typical scan of a T-joint welded sample with discrete flaws; Figure 63 shows a typical scan of a T-joint welded sample with 25mm flaws and Figure 64 shows a typical scan of a T-joint welded sample with a cold weld flaw.
It can be seen from Table 5 that for all of the T-joint welded samples which were produced as fault free no flaws were detected. However, of the butt weld samples which were produced as fault free, four out of six (the thicker plates) contained 100% lack of root fusion in the centre of the double V joint. In addition, sample 42 contained 35mm of no weld and sample 82 contained 60mm of lack of side wall fusion.
For all of the butt welds produced with discrete flaws, all of the flaws were detected, some as small as 2mm. For all of the T-joint welds produced with discrete flaws, some smaller flaws were not detected, as few as 50% of the 2mm flaws. The percentage of discrete flaws detected is illustrated in Figure 65. Only the T-joint samples were produced with 25mm flaws at varying distances from the root, and unfortunately, the 40mm thick sample was produced without any of these in error. Therefore only two samples with this type of flaw produced any results, the 15mm thick sample and the 25mm thick sample. All of the flaws were detected in the 25mm thick sample, but for the 15mm thick sample the detection rate reduced from 100% for the discs at 2mm from the root to 25% for the discs at 10mm from the root. Of the samples produced with cold welds the current PAUT technique was unable to detect any signs that the welds were cold. However the inspection revealed that all of the double V butt welds contained up to 100% lack of root fusion.
The PAUT techniques developed were able to detect most of the artificial flaws inserted into the welded samples with the following results:
• The fault free samples were successfully inspected and some unintentional flaws were detected.
• The samples containing discrete flaws were inspected with a detection rate of 50% for the 2mm size flaw in the T-joint welds rising to 100% for the 25mm flaw.
• The detection rate for the 25mm size flaws at varying distances from the root was from 25% for the 10mm distance to 100% for the 2mm distance.
• Evidence of cold welds was not detected, however up to 100% lack of root fusion was detected.
WP3 – Development of Flaw Acceptance Criteria
WP4 – Integration and Validation of Ultrasonic Inspection System
A scanning system has been designed to encompass the full range of weld geometries specified in WP1. Currently available scanners which may be suitable for tanks are designed for metallic tanks and are frequently held in place with magnets, which will not be applicable for plastics tanks. It was the aim that a single scanner should encompass all the sizes of tank and the base fillet weld, pipe outlet fillet weld and base butt weld configurations. To extend the use of the manipulator and minimise costs it has been designed in a modular form with simple changes for different configurations. A complex part of the manipulator has been the probe holder, which is required to hold the probe in a very precise position around the fillet welds or butt welds. A system has been designed and manufactured in a modular format, with the capability of accommodating the different types of joints. The scanner system is integrated with an encoder for positional information.
Aim
The aim was to develop a 3-axis portable scanner that can be put in place by a single operator. It should be able to scan the weld at the bottom of the tank and also the weld around a nozzle pipe. The scanner should be:
• As lightweight as possible.
• As low a profile in z direction as possible to be able to operate in constrained spaces.
• Able to measure the force applied by the probe to enable the scanner to apply a constant contact force with a surface.
• Detect obstacles at the end of axis 2 to prevent collision with the ground and the nozzle pipe.
Assuming a minimum tank diameter of 2m, then a 600mm long scanner with three axis should be able to inspect the weld in segments of at least 500mm length at each pass. The three joints of the scanner should be prismatic (3 linear slides actuated by motor/gearbox combinations). The scanner will be attached to the tank wall by four or more suction cups actuated mechanically using levers. If required, an air compressor with venturi valves can actuate the suction cups for long periods. To inspect the weld at the bottom of a tank, attach the scanner above the weld. Axis 2 moves the probe over the weld and is then locked. Axis 3 brings the probe into contact with the weld with the desired contact force. Axis one then deploys the probe along the weld. Axis 3 makes any adaptation to the curvature by maintaining a constant contact force. To inspect a nozzle weld, the scanner is placed vertically on the tank wall. Axis 2 and 3 is actuated to bring the probe in contact with the weld at a starting point. Axis 3 is locked.(assuming that there will be little change required in the depth when the probe is moved around a half circumference of the pipe weld). Axes 1 and 2 then execute a curved trajectory to follow the weld line. The proximity sensor at the end of axis 2 can be used to guide the following of the trajectory.
In concept the scanner was envisaged to consist of 3 linear slides, 3 motor/gearbox combinations, 3 servo amplifiers and controllers, one supervisory controller.
• The motor/gearboxes on board the scanner should be splash proof.
• The servo amps and controllers will be in a separate splash proof rack quite close to the tank. Initially the system will not be designed for intrinsic safety operation in an explosive environment. The splash proof control rack should be wheeled for easy transportation to a test site.
• The Phased array probe should be spring loaded with a dual ball bearing, dual gimbal arrangement to keep the probe normal to the surface on slightly contoured surfaces.
• The control system should be easy to setup with programmed trajectories e.g. operator specifies points A and B by placing the probe on two points. The controller automatically performs a scan between A and B by maintaining the desired contact force and distance from the ground or pipe.
A review of commercially available NDT scanners with a similar concept was undertaken to aid in the design process. Taking into account the design considerations in above and using some of the ideas generated from the review of commercially available NDT scanners, several concept designs were produced as shown in figures 66, 67 & 68. After careful consideration and discussions with the interested SME, Innotech, and the project steering committee it was decided to take forward the concept shown in Figure 68. This concept was further refined into two options shown in figures 69 & 70. Both of these options will satisfy the project requirements and it was the option shown in figure 70 which was the one taken forward for assembly. However, this option can easily be modified to satisfy the option shown in figure 69. The scanner has been designed around commercially available components. Figure 71 shows the fully assembled scanner.
Validation Work
Table 2 details the joints which were produced for PAUT testing by the PolyTank system. All of the Double V butt welds and all of the T-joint fillet welds were inspected, however, due to problems with the membrane water wedge development; it was not possible to inspect the Pipe in sheet fillet welds. This shortcoming was accepted by the consortium as these were considered to be of a lower priority to the other welds. Two techniques were used to inspect the samples, an angle beam PAUT sectorial scan to inspect the butt welds and a zero degree PAUT sectorial scan to inspect the fillet welds. For each weld, the phased array system electronically focuses a compression wave acoustic beam at user specified depth within the material, which is determined by a series of “focal laws”. The focused beam is then electronically swept through a range of angles to fully cover the inspection volume. An image of the configuration for the technique used on butt fusion joints is shown in Figure 72. Butt fusion welds were scanned from both sides of the weld and both top and bottom of the plate. The datum point was the plate edge and linear scan measurements were taken from this point. Figure 73 shows the probe positions and scan direction. Table 6 gives the relevant probe and focal law parameters for the BF joints. The array controller software display screen was set up to show the A-scan, a B-scan and an S-scan (sectorial scan). Figure 74 shows an example for the BF joint inspection. An image of the configuration for the technique used on the T-joint fillet weld is shown in Figure 75. The T-joint fillet welds were scanned from the weld cap with the probe orientated to produce an angle between the central beam and the respective fusion face approaching 90 degrees. The datum point was the end of the welded section of each sample and linear scan measurements were taken from this point. Figure 76 shows the probe position. Table 7 gives the relevant probe and focal law parameters for the T joints. The array controller software display screen was set up to show the A-scan, a B-scan and an S-scan (sectorial scan). Figure 77 shows an example for the T-joint inspection.
Results
Table 5 details the results of the inspections carried out on the Butt weld and T-joint Weld samples. The Pipe-in-sheet samples were not inspected due to the inability to successfully position the probe on the fillet weld. Further probe wedge development is required to achieve an inspection of this complex geometry. There were four types of weld flaw produced:
• Welds with no flaws.
• Welds with discrete flaws of varying sizes ranging from 2 to 25mm diameter.
• Welds with 25mm diameter flaws at varying distances from the root ranging from 2 to 10mm
• Welds with a cold weld element.
The PAUT system was assembled and successfully tested in the laboratory. The Butt welded sheets and fillet welded T-joint samples were inspected and the results presented above. The pipe-in-sheet samples were not inspected due to limitations of the membrane water wedge.
WP5 – Field trials using validated Ultrasonic Inspection System
A finalised inspection procedure has been produced and was used during the field trials. The field trials were undertaken in three parts, firstly initial trials at Chemresist to demonstrate the capabilities of the techniques developed in the laboratory, then trials at Univar on live tanks to demonstrate the in-service flaw detection capabilities and finally blind trials at Chemresist to demonstrate the in-service flaw detection capabilities.
Initial Trials
The initial trials were carried out at the premises of Chemresist in Dewsbury, UK on the 24/25 June 2014. The objective of the initial trials was to demonstrate the capabilities of three inspection techniques which had been developed in the laboratory in a site environment. The three techniques were:
• T-joint fillet weld inspection for lack of fusion.
• T-joint fillet weld tank wall toe inspection for cracking.
• T-joint fillet weld tank base toe inspection for cracking.
The PolyTank probe fitted to the specially designed membrane water wedge was used to inspect the sample T-joint fillet welds cut from a sample tank provided by HSL identified as ‘Base-T1’. Figure 78 shows an image of the scan being carried out and Figure 79 shows a snap shot of the resultant scan. The resultant scans from this inspection gave indications from the nearside fillet weld root and the opposite fillet weld toes. There were no lack of fusion flaws present to be detected and the technique was unable to distinguish between a signal from the remote weld toes and the crack which was present. It was concluded from this that this technique is capable of inspecting the nearside fillet weld for fusion defects but not able to detect in-service cracking in the remote fillet weld toes.
A 4MHz probe fitted with an angle beam membrane water wedge was used to inspect the tank sample Base-T1 for cracking in the inner (remote) fillet weld to tank wall toe. An image of this scan being carried out is shown in Figure 80 and the resultant scan is shown in Figure 81. The resultant scans from this demonstration have shown that the technique using the 4MHz probe fitted with an angle beam membrane water wedge was able to successfully detect cracking in the fillet weld toe on the tank wall.
The PolyTank probe fitted to the specially designed compact membrane water wedge was used to inspect the sample T-joint fillet welds cut from a sample tank provided by HSL identified as ‘Base-T1’ to detect cracking in the remote fillet weld toe in the tank base. An image of this scan being carried out is shown in Figure 82 and the resultant scan is shown in Figure 83. This inspection technique provided a possible indication of the cracking in the remote fillet weld toe in the tank base plate, however the indication could not be positively confirmed and therefore further technique development is required.
The 4MHz probe fitted with an angle beam membrane water wedge was used to inspect the full tank sample T2 for cracking in the inner (remote) fillet weld to tank wall toe. An image of this scan being carried out is shown in Figure 84. There was no resultant scan as there was a significant inner and outer weld toe mismatch which created interpretation difficulties.
In-service Trials
The in-service trials were carried out at the premises of Univar in Wellingborough, UK. The objective of these trials was to apply the in-service crack detection techniques demonstrated in the initial trials to a small number of fully functional in-service tanks in a live tank farm. The tanks inspected are detailed in Table 8.
Tank WL23
This was the first tank identified for a trial inspection. Initial examination revealed that the construction of this tank was at variance with those inspected to date. Instead of the tank wall resting on the base in an inverted T form, the base was a sloping disc inserted into the tubular tank wall. Figure 85 illustrates the different construction methods. Due to the construction of the tank the only inspection technique which could be applied was the T-joint fillet weld tank wall toe inspection. Before this could be applied it was necessary to locate the exact position of the internal fillet welds. This was achieved by the use of a zero degree probe to plot the absence of the tank wall signal. Where there was no response from the inner side of the tank wall it was presumed that this was the position of the fillet weld. During this investigation the wall thickness was determined to be approximately 25mm thick. Figure 86 shows the weld position plotted onto the outer surface of the tank wall. The sample length of weld plotted out as shown in Figure 86 was scanned using the 4MHz probe fitted with an angle beam membrane water wedge and an example of the resultant scan is shown in Figure 87. No crack indications were detected at the weld toe location however several irregular indications were detected mid-way through the wall thickness as shown in Figure 88. The manufacturer (Chemresist) was able to confirm that these were due to the manufacturing process causing a certain amount of inter-wrap lack of fusion which is acceptable.
Tank WL 2R
The construction of this tank is the same as WL 23 and so the weld position was plotted in the same way. During this investigation the wall thickness was determined to be approximately 35mm thick. Figure 89 shows the weld position plotted onto the outer surface of the tank wall. The sample length of weld plotted out as shown in Figure 89 was scanned using the 4MHz probe fitted with an angle beam membrane water wedge and an example of the resultant scan is shown in Figure 90. No crack indications were detected at the weld toe location however several irregular indications were detected mid-way through the wall thickness as shown in Figure 91. The manufacturer (Chemresist) was able to confirm that these were due to the manufacturing process causing a certain amount of inter-wrap lack of fusion which is acceptable.
Tank WL 22
The construction of this tank was as expected as shown in the left side diagram in Figure 85. A zero degree UT probe was used to confirm the wall thickness as approximately 17mm. The tank is shown in Figure 92. This tank was scanned using both the 4MHz probe fitted with an angle beam membrane water wedge to inspect the tank wall and the PolyTank 2.25MHz probe without a wedge to inspect the tank base plate. An example scan of the tank wall is shown in Figure 93, no defects were found. The scan of the tank base plate was not recorded, however a small indication at a range of 62mm was noted, which is beyond the expected range for the base plate fillet weld toe (45mm) and therefore unlikely to be due to cracking. Univar took note of this indication.
Blind trials
The blind trials were carried out at the premises of Chemresist in Dewsbury, UK. The objective of the blind trials was to demonstrate the capabilities of two inspection techniques in a site environment to successfully detect discrete artificially induced cracking in the tank wall to base plate fillet welds. The two techniques were:
• T-joint fillet weld tank wall toe inspection for cracking.
• T-joint fillet weld tank base toe inspection for cracking.
To simulate discrete cracks in the inner fillet weld toes (tank wall and base) slots of approximately 3mm deep and between 30mm and 120mm in length were cut into the weld toe as shown in Figure 94. The defects were introduced into two different areas, each containing 3 defects in the wall weld toe and 2 defects in the base weld toe. Area 1 is shown in Figure 95 and Area 2 is shown in Figure 96. Each area was marked out for scanning as shown in Figures 97 and 98. Table 9 shows the results of the inspections and figures 99 to 107 show the indication for each defect. All the artificial defects were detected.
The series of field trials successfully demonstrated the capabilities of the Polytank PAUT system to detect in-service cracking in the toes of the internal fillet welds. The techniques to detect production fusion flaws were not demonstrated during the field trials; however, the system validation inspections have successfully demonstrated these.
Potential Impact:
The expected impacts that the PolyTank project will have on the EU can be divided into three categories:
- Economic
- Societal
- Environmental
Economic impacts
The implementation of the results of the PolyTank project will result in reduced capital cost of many tanks and tank systems by facilitating greater use of thermoplastics as opposed to steel and other metals. It will also reduce repair costs by enabling the identification of defective joints before entry into service. Maintenance costs will be reduced by facilitating greater use of thermoplastics, which generally have a greater resistance to the stored chemicals and consequently longer operating lives than currently used materials. Losses of product from tank failure will be avoided as well as costs of emergency repairs and consequential losses from tank leaks. There will be a reduced risk of leakage of environmentally damaging fluids and consequent costly repairs and clean-up costs and reduced risk of catastrophic failure resulting in loss of life and consequential economic costs particularly for hydrocarbon and acid carrying tanks.
Societal impacts
There are advantages to the community in the use of thermoplastic tanks over the conventional metal storage tanks. Many corrosive chemicals (e.g. hydrofluoric acid) are more safely stored in thermoplastic tanks, often in areas of high population density. Chemicals of this type are not only safer to store in thermoplastic tanks but it is also cheaper and requires less maintenance. The societal value is therefore in getting a safer and cheaper storage of toxic and corrosive chemicals when compared to storage in steel tanks. Also, because the developed PolyTank systems and techniques will be applied from the outside of the tank, there will not be a need for man entry into confined spaces for pre-cleaning or testing of tanks that can often hold hydrocarbons or toxic chemicals. It is anticipated that the project will create new EU employment opportunities (through new knowledge generated from PolyTank), which will be passed on to the EU SMEs operating in the thermoplastic tank manufacture, installation, repair/maintenance and inspection markets. In addition, funding from the FP7 will propel EU SMEs to the forefront of in-service monitoring of thermoplastic tanks and other similar components, thereby enabling other industries to benefit from increased product reliability (e.g. tanks used for liquids transportation).
The results of this project will lead to an improvement in the level of skills and knowledge for European SMEs in NDT, thermoplastic welded products (including storage tanks) and materials technology. The introduction of ultrasonic NDT technology in the thermoplastic storage tank industry will represent a step change. This will necessitate training of operators of the ultrasonic NDT equipment and procedures. Currently, there is a shortage of NDT inspectors in the EU and member states are investing heavily in this area. The results of PolyTank will contribute to this effort.
Environmental impacts
The use of ultrasonic NDT on welded joints (circumferential and helix welds and nozzle welds) in thermoplastic tanks as a quality assurance method will lessen the risk of failure and increase confidence in the use of these materials. Due to the demonstrable evidence of thermoplastic tank weld integrity that will result from the PolyTank project, this will lead to the more widespread application of thermoplastic storage tanks within the process industry sector, and will allow thermoplastic tanks to be used for more stringent service applications in the future. Failures in storage tanks can range from the environmentally unacceptable to the catastrophic (Stonehill J, Bainbridge H and Heyes PF: “Specification and inspection of thermoplastic storage tanks‟. HSL Report HSL/2006/21, 2002). The referenced HSL document describes failure and likely defect mechanisms that include in no small part welding defects, and the report recommends that thermoplastic tanks should be inspected (at this point in time by vessel entry) within the first 2 years of operation, this being a clear indicator of the concern of the HSL report as to the integrity of this type of tank for service. Furthermore the HSL report states that there is considerable scope for further examination and consideration of methods of inspecting and monitoring thermoplastic tanks. Leaks of hydrocarbons, chemicals and processed liquids from storage tanks can be potentially harmful to the environment, local workers and residents and is therefore totally unacceptable. Leaks of this type must be rendered as extremely unlikely events and therefore the integrity of storage tank welds throughout their life must be carefully monitored and understood against realistic acceptance criteria that will be developed as part of the project. The direct impact of the project on working and living conditions is limited, but will accrue from reduced disruption to industry, making the workplace safer by reducing hydrocarbon toxic chemical spillages. Minor leakage from water storage tanks for example has been accepted in the past, but is increasingly penalised as water demand rises and concerns grow for the amount of water abstracted from rivers and boreholes.
Main dissemination activities
Project website
The project website, http://www.polytank.eu/(odnośnik otworzy się w nowym oknie) was launched on 26 March 2013 and contains the following pages:
Public facing area
• Home – Welcome
• About the project
o Project background
o Project concept
o Strategic objectives
o Work packages
• Project Partners
o CHEM RESIST
o TAB
o ACUTECH
o INNOTECUK
o TWI
o HESSEL
o HSL
o UNIVAR
• Publications
• News
• Videos
• Contact us
Partners’ area (Consortium only)
• Project notices
• Project documents
• Project contacts
• My account
• Sign out
Admin only menu
• Logs: User access
• Logs: Database usage
• Add new item
The website also contains the project logo, which is shown in Figure 108.
Project flyer, video and banner
The PolyTank flyer was produced in 2013 (Figures 109 and 110) and is available as a download from the project website (http://www.polytank.eu/publications.jsp(odnośnik otworzy się w nowym oknie)). The text for the flyer is available in English, Dutch, French and German. A video has been produced outlining the project and introducing the project partners. This is hosted on YouTube and is linked from the project website http://www.polytank.eu/videos.jsp(odnośnik otworzy się w nowym oknie). A final project video has been produced, including footage to illustrate the welding procedures, inspection procedures and the equipment design. A freestanding banner has also been produced (Figure111).
Conferences
The PolyTank project has been presented at the following conferences and events:
• 53rd Annual Conference of the British Institute for Non-Destructive Testing, 9-11 September 2014, Manchester, UK.
• European Conference on Non Destructive Testing. 6-10 October 2014, Prague, Czech Republic.
Articles
The following papers and articles about the PolyTank project have been published:
• http://www.sensorsportal.com/HTML/DIGEST/january_2013/PolyTank_TAB.htm(odnośnik otworzy się w nowym oknie) Sensorsportal e-digest newsletter January 2013
• http://www.twi.co.uk/news-events/news/2013-04-polytank-project-to-develop-safety-critical-approach-for-testing-welded-joints-in-thermoplastic-storage-tanks/(odnośnik otworzy się w nowym oknie) Published on TWI website April 2013
• http://www.mundoplast.com/noticia/proyecto-polytank-sobre-testeo-juntas-tanques-plastico/70151(odnośnik otworzy się w nowym oknie) Published in ETD Prensa Profesional (MUNDOPLAST)
• http://www.europeanplasticsnews.com/subscriber/headlines2.html?cat=1&id=2791(odnośnik otworzy się w nowym oknie) European plastics news April 2013
• http://www.sensorsportal.com/HTML/PolyTank.htm(odnośnik otworzy się w nowym oknie) Web page
• http://inspectioneering.com/tag/storage+tanks(odnośnik otworzy się w nowym oknie) Inspectioneering
• Facebook May 2013
• Google +
• Welding and cutting. Technical Journal. Issue 03, 2013
Sensors & Transducers e-Digest, Vol. 148, Issue 1, January 2012: Products News http://www.sensorsportal.com/HTML/DIGEST/january_2013/PolyTank_TAB.htm(odnośnik otworzy się w nowym oknie)
• http://www.polytank.eu/publications/Flyer_PolyTank_v3.pdf(odnośnik otworzy się w nowym oknie) Flyer
• Introduction to PolyTank video http://www.polytank.eu/videos.jsp(odnośnik otworzy się w nowym oknie)
• Flyer Distributed at SENSORCOMM' 2013 conference, 25-29 August 2013, Barcelona, Spain.
• TWI connect magazine. Issue 184 – May/June 2013. http://www.twi.co.uk/news-events/connect/connect-may-june-2013/(odnośnik otworzy się w nowym oknie)
• Cordis Wire https://cordis.europa.eu/wire/index.cfm?fuseaction=main.viewrelease
• Silobreaker http://news.silobreaker.com/polytank-project-to-develop-safetycritical-approach-for-testing-welded-joints-in-thermoplastic-storage-tanks-5_2267122038646767662(odnośnik otworzy się w nowym oknie)
• Flyer Circulated to ten UK tank manufacturers
• March 2014 website updates include: addition of new partner info, NIKK, changes to existing partner info, changes relating to EU guidance on logos and funding statement
• 48 companies contacted using the project flyer
Dissemination seminars
• “Vorstellung des Forschungsprojektes Polytank zur zerstörungs-freien Überprüfung von Schweißnähten mittels Ultarschall” (Presentation of the Polytank research project for non destructive inspection of welding seams by ultrasonic testing) Thermoplastics for the chemical processing industry Seminar organised by Rochling. 6 June 2013, Haren, Germany.
• Kunststoffe im Anlagenbau – Plastics in Process Plant. TÜV Süd Seminar. 25-26 February 2014. Munich, Germany.
• NDT of Welds in Thermoplastic Tanks - NDT von Schweißnähten in Thermoplastbehältern. DVS AG W4.3b 9 April 2014, Minden, Germany
• New developments in NDT of Plastics and Composites. TWI Technical Group Meeting. 23 May 2014, Chesterfield, UK
Exploitation of results
The main exploitable products from the PolyTank project are:
• Phased array probes
o Ultrasonic low frequency probes for testing of high attenuation thermoplastic welds.
• Ultrasonic instrument
o Ultrasonic remote site data acquisition/display instrument and techniques for volumetric inspection of thermoplastic welds.
• In-service inspection of PE tanks
o Flaw acceptance criteria.
• Prototype inspection system
o Automated manipulator scanning system.
List of Websites:
http://www.polytank.eu/(odnośnik otworzy się w nowym oknie)
