Community Research and Development Information Service - CORDIS

Final Report Summary - PILEINSPECT (Integrity Testing of Deep Foundation Piles)

Executive Summary:
An estimated 100,000 kilometres of pre-cast piles are installed every year in Europe: the use of deep driven or cast-in-situ piles as a basis for building foundation is extremely extensive. Problems can occur during pile driving including spalling of concrete at the pile’s head or point and transverse or spiral cracking. Industry figures suggest that as many as 3-5% of piles fail during installation. If a failure is detected, the cost of that pile is increased 4 times due to the remedial work needed to replace it. If a failure is not detected during installation of the new pile, the results can be more catastrophic with costs sometimes exceeding €1 M for a single failure.
There are three million enterprises in the European construction industry (EU27), 95% of which are SMEs with fewer than 20 employees and 93% with fewer than 10 employees. This sector employs 14.9 million people which is 7% of Europe’s total employment and constitutes 29% of industrial employment.
In terms of employee numbers, this sector is the largest in Europe. Estimated investment in the construction industry (EU 27 - 2009) is € 1.2 billion. Competitiveness of the sector has been severely impacted by the global financial crisis and poor performance will impact negatively on the European economic recovery. Within the construction sector there are just under 11,000 companies concerned with deep foundations and piling, 70% of which are SMEs.
Current pile inspection techniques involve dynamic load or sonic integrity testing. These are relatively fast to perform. However, the quality of results depends strongly on the knowledge and skill of the operator. It has been found that small defects that are less than about 0.4m (quarter wavelength) are difficult to detect. Some studies also indicate that defects representing less than 50% of cross sectional area are not detectable via sonic integrity testing. In the PileInspect project, a new technologies and a low cost automatic near real time system were developed for the early detection of incipient damage in pile foundations, both new and existing.
Within the PileInspect, the Consortium has developed ‘best practice’ for inspecting the integrity of piles. A portable shaker which ensures repeatable, tailored excitation spectra, and improves accuracy of estimation of the proposed diagnostic features and the instrumented hammer are employed. Highly innovative signal processing methodologies (based on time- frequency and the higher order spectral techniques formulated for non-stationary signals) have been employed in order to increase the quality of diagnosis and perform automatic defect recognition.
Within the PileInspect, the prototype of the pile diagnosis system with automated decision making was developed and tested in both laboratory and field conditions. The prototype includes widely available commercial off-the-shelf hardware parts and custom made software, implementing the innovative PileInspect diagnosis technologies. For the increased efficiency of the pile damage diagnosis, a fusion of multiple damage detection techniques is used and complemented by the comprehensive novel decision making algorithm, with the final automatic decision made without relation to the operator’s knowledge or his past experience in the pile damage diagnosis.
The comparison with the traditional pile integrity testing approach, based on the low strain sonic integrity testing (PIT) has shown a systematic advantage of the developed PileInspect damage diagnosis technologies over the traditional PIT test by significant margin: 36-58%

Project Context and Objectives:
WP1: Preparatory development and modelling
Work Package 1: Preparatory development and modelling
T1.1 Obtain samples of damaged and undamaged piles
Cranfield and Aarsleff successfully developed a unique comprehensive library (i.e. 15 piles) of damaged and undamaged industrial piles at Aarsleff, Newark, UK using the standard industrial method for pile installation. These piles are of different cross-sections and length. Cranfield and Aarsleff in collaboration with all partners performed successful testing of all piles using the instrumented hammer and two types of shaker: Data Physics shaker and a low cost portable shaker (I-Beam). DFI, BAM and UBrun visited Aarsleff’s site. BAM performed PIT testing of piles at Aarsleff. Some of the measurements were later on repeated by GSP. Aarsleff’s site is available for WP3 and WP4.
BAM made the BAM-TTS (Test site Technical Safety, www.tts.bam.de) available to the project. It comprised more than 20 bored piles in diameters from 60 to 90 cm, length 8.5 to 12 m, various flaws and measurement conditions. The documentation was made available to the consortium. The site was visited by the consortium during the 2nd project meeting. In WP4, UBRUN and BAM performed field test at this site to validate the developed PileInspect prototype and pile diagnosis method using impulse response function.
BAM negotiated successfully to make an additional test site (property of Deltares, Utrecht, NL) available to the project. Cooperation with BMNED was very helpful in achieving this. This site has more than 20 piles with various shapes, lengths and faults available. BAM visited the site and performed preliminary tests with the traditional PIT and the BAM low cost shaker (see WP4) in 2015.
T1.2 Simulate vibration signals from damaged and un-damaged piles
Cranfield in collaboration with all Partners created two types of simulation: dynamic simulation using the bilinear dynamical system and finite element simulation using the linear and nonlinear finite element analysis, including novel methodology of finite element modelling of damaged piles. The software for both simulations was developed. Focus of Cranfield’s simulation and the finite element analysis was on geometries resembling Aarsleff’s test site. A comparison of the simulation results and real test data showed good agreement. Diagnosis results of piles by the simulation data and by the experimental data also showed good agreement.
To produce test data for the new signal processing software as well as to assist the hardware development, BAM carried out numerical simulations. Other than stated in the DoW BAM used a finite integration approach in cylindrical coordinates (CEFIT) to cross validate the simulations of Cranfield and UBrun. The software used was originally developed in the frame of the FP5 project RUFUS.
PileInspect prototype had to be optimized in speed and storage usage as well as to allow custom source inputs (e. g. shaker sweeps). This was done in cooperation with the original author (Dr. Frank Schubert of the Fraunhofer society). BAM participated in a German geophysical conference to acquire knowledge about the latest simulation tools. As promised in the DoW simulations were performed with various impulse shapes and swept sine excitations for a variety of pile shapes, flaw geometries and soil conditions. Cross validation against the 1D finite element solution of UBrun showed good agreement. Focus of the BAM simulations was on geometries resembling the BAM test site. A comparison of the simulation results and real test data showed sufficient agreement. Data became available to all partners and have so far been exchanged mainly with UBrun. Results were reported in D 1.3. The task is successfully finished.
UBrun built a theoretical 1D soil-pile system to simulate vibration signals from damaged and un-damaged piles. The pile-soil system modelled the soil stiffness and radiation damping with a set of springs and dashpots. On the lateral surface of the pile, distributed springs with elastic stiffness and dashpots with damping coefficient were applied to each segment below the ground surface. The soil at the base of the pile was modelled using a spring with elastic constant and a dashpot with damping coefficient. The finite difference method was used to solve the motion equation and simulate the pile response to excitations. Validity of various scenarios was calculated in the cases of different soil properties, pile properties/geometrics. Two types of excitations were considered for numerical simulations: impact excitation by a hammer and sine sweep excitations by a shaker. The simulated travel times of reflection wave at both pile bottom and pile defects matched well with the theoretical values. The simulations and results were used for validation with test data, which were reported in D 2.2.
T1.3 Generate shaker test data for damaged and undamaged piles
BAM set up a new low cost shaker system as an add-on to the high power shaker system used by Cranfield. One reason for this approach was that the Cranfield system was not available for lab or field tests at BAM in the first phase of the project. For details see WP4. The new shaker system was used to generate test data mainly on piles at the BAM test site on undamaged and damaged specimens. Various linear and logarithmic sweep signals were tested, differing in frequency content, durations and amplitude. Various sensor arrangements were used. The data were reported in D1.3 and are available to the consortium.
Cranfield and Aarsleff in collaboration with all partners have successfully generated shaker test data for damaged and undamaged industrial piles at Aarsleff using two types of shaker: Data Physics shaker and a low cost portable shaker (suggested by BAM as an alternative tool) . Two types of excitation were generated by both shakers: the swept sine excitation and impact excitation. Various sensor arrangements and various locations of shaker excitations were used. The data has been reported and are available to the Consortium.
T1.4 Develop and validate the higher order spectra techniques for stationary and non-stationary pile vibrations
Cranfield in collaboration with all Partners successfully implemented and experimentally validated together with Aarsleff by the simulated and the experimental data the novel higher order spectra: the wavelet higher order spectra and the FRF chirp-Fourier higher order spectra for stationary and non-stationary pile vibrations using the swept sine and impact excitations. Results were obtained for different pile vibration modes, different intensities of excitation, different positions of accelerometers and different locations of impact and swept sine excitations from the instrumented hammer and two types of shakers.
BAM contributed to this task by giving feedback to Cranfield’s developments and by developing (in cooperation with UBrun) an additional approach to use the data generated by the shaker systems. While the two approaches proposed by Cranfield and UBrun resulted in criteria to judge whether a pile is damaged or not, traditional PIT also gives an idea on the approximate pile length and (if applicable) flaw location. This information can be gathered from shaker data as well by transforming the receiver data in an appropriate way.
BAM and UBrun tested various ways to perform this transformation. Most promising was to use a deconvolution algorithm, which had to be combined with a regularization term to provide stability. Based on results acquired with simulated data, BAM favours Tikhonov regularization. Results achieved were reported in D1.3. Validation followed in WP3 as well as work to include the findings in the final software (WP4).
UBrun, in cooperation with BAM as well as with all consortium partners, proposed and used impulse response measurement based method for pile damage diagnosis. The PileInspect system replaces instrumented hammer by a portable low-cost shaker, which can excite the piles with repeatable and tailored sine sweeps. The echoes at pile toe and pile defects are visible in the time domain records in the straightforward impact-echo methods. However, these echoes are not visually available in the time domain records of piles subjected to sine sweep excitations.
UBrun and BAM proposed to extract impulse response function from input (shaker excitation) and output measurements (pile response). This technique is similar to Green’s function or receiver-function computations commonly used in seismology for determining source-time functions or for identifying subsurface structures after removing source effects. The impulse response functions were computed by deconvolving the pile response from the sine sweep excitations. Considering the simple spectral ratio representation is unstable when the input spectrum is near to zero, a regularization parameter was used to stabilize the deconvolution. BAM and UBrun followed slightly different approaches here. However, the results were comparable. The reflection wave in the pile, i.e., the echoes from pile toe and pile defects with impedance change, can be identified from the impulse response function of a pile in the time domain. To refine identifying resolution, UBrun and BAM tried varieties of parameters of input sweep signal: including sweep signals types, frequency bandwidth and sweep’s time length. The simulation results were further validated with experimental data. The proposed methodology was reported in D1.1 and validated with pile test data obtained by BAM (described in D2.2).
T1.5 Develop and validate novel amplitude and phase extraction from the stationary and non-stationary higher order spectra
Cranfield implemented and successfully validated together with Aarsleff by the simulated and the experimental data the novel amplitude and phase extraction technique from the non-stationary higher order spectra: i.e. from the wavelet higher order spectra and from the chirp Fourier higher order spectra. The implemented approaches were compared with the standard amplitude approach and essential gains in effectiveness of diagnosis were estimated for various types of excitation.
T1.6 Develop and validate novel anomaly detection for pile damage diagnosis
Cranfield implemented and successfully validated together with Aarsleff by the simulated and the experimental data the novel anomaly detection for pile damage diagnosis. The main advantage of the method is that it can be used in the cases for which training data only from undamaged piles are available. The method was applied for diagnostic features based on the wavelet higher order spectra and on the FRF of the chirp Fourier higher order spectra for the impact and the swept sine excitations respectively. The gain in effectiveness of diagnosis was estimated for various types of excitation.
D 1.1) Report on methodology to process and analyse non-stationary resonance and free vibrations from piles (T1.4 and T1.5): this report was issued by the Consortium, the lead partner is Cranfield
D1.2) Report on anomaly detection method for pile damage diagnosis (T1.6): this report was issued by the Consortium, the lead partner is Cranfield
D 1.3) Report on Experimental shaker test data (T1.3): this report was issued by the Consortium, the lead partner is BAM
D1.4) Report comparing simulation and experimental results (T1.6): this report was issued by the Consortium, the lead partner is Cranfield
Work Package 2 Experimental validation of signal processing by trials
T2.1 Perform trials with piles in undamaged conditions
T2.2 Perform trials with piles with known fault conditions
Cranfield and Aarsleff in collaboration with all Consortium Partners successfully performed trials with industrial piles in undamaged conditions at Aarsleff Newark site, UK. Trials included usage of three exciters: the instrumented hammer, Data Physics shaker and portable low cost shaker. Multiple tri-axial accelerometers were installed in different locations of piles. Excitation was performed in different directions; axial and horizontal. Low and high intensity excitations were employed.
Work on these two tasks was also performed by BAM on piles at the BAM-TTS test site, Germany. The low cost shaker system was used (see WP4). Other than described in the DoW and followed by Cranfield, the shaker was mounted on top of the piles as the BAM pile have a steel casing on the top 1 - 1.5 m. Accelerometers were mounted at various positions on top of the shaker, on top of the pile and the side of the pile using adhesive material or magnetic mounts. Space available on the side was limited as in real applications. Various setups and measurement parameters were tested. Mainly the sweep signal was varied in frequency content, duration and sweep type. When using optimized parameters, a significant difference could be seen even in raw data between an undamaged and a damaged pile while it can’t be determined from raw data which one is the undamaged one.
This task has been successfully completed. Cranfield and Aarsleff in collaboration with all Consortium Partners successfully performed trials with industrial piles in damaged conditions.
Trials included usage of three exciters: the instrumented hammer, Data Physics shaker and portable low cost shaker. Multiple tri-axial accelerometers were installed in different locations of piles. Excitation was performed in different directions; axial and horizontal. Low and high intensity excitations were employed.
T2.3 Process trial data from damaged and undamaged conditions
BAM contributed to this task by giving feedback to Cranfield’s developments and results. In addition BAM used its own test data from T2.1/T2.2 to evaluate the deconvolution approach developed in T1.3 using real field data. Data using various setups and sweep parameters were recorded on an intact pile (11 m length) and on a pile with significant cracks in 3-4 m depth. Deconvolution with Tikhonov regularization was used to transform the data generated by using the shaker sweeps to an estimate of the impulse response, reflecting the approximate pile length and (if applicable) flaw location as in traditional PIT.
At the time of delivering D1.3 (results reported there) BAM had success with the intact pile (pile length was matched within 5%), but not with the cracked pile (only very small flaw signature). Meanwhile (within the reporting period), the deconvolution approach was improved in cooperation with UBrun. Now it is possible to provide a significant flaw signature and an estimate of the flaw position, so far with an accuracy of 10% of the pile length. Results achieved were reported in D1.3. Validation was to follow in WP3 as well as work to include the findings in the final software (WP4).
UBrun used the proposed impulse response method (proposed and reported in D1.1) to process trial data provided by BAM and to identify pile defects of damaged piles. BAM conducted experimental shaker test on two piles at the pile test site in Horstwalde, Berlin. The tested piles were the intact pile P03 and the damaged pile P06. The cylindrical shaped, reinforced concrete piles were 90 cm in diameter and 11 m long. The top one m sticks out of the surface. The failure of P06 (horizontal cracks) was located in a depth of approx. 3.5 m below the pile top.
The finite difference method was used on a numerical 1D pile-soil model to simulate the responses of the piles to both impact excitation and sine sweep excitation. The shaker test data and the comparison between theoretical and experimental results were reported in Deliverable 2.2. Validation of proposed single processing technique on preliminary test data showed promising results: reflections from pile toe and defect were identified from some test data however clear echoes were still not available in some tests.
Cranfield and Aarsleff in collaboration with all Consortium Partners successfully performed processing of trial data from damaged and undamaged industrial piles. The following new techniques were employed: the wavelet higher order spectra for impact excitations from the instrumented hammer and shaker and the FRF based on the chirp Fourier bicoherence for the swept sine excitation from the shaker. The diagnosis results were very successful and showed high effectiveness of diagnosis of damaged and undamaged conditions with the estimate of the total probability of correct diagnosis at 100%. These results were achieved for the first time in worldwide terms.
D2.1 Experimental trial data (T2.1 and T2.2): this report was issued by the Consortium, the lead partner is BAM
D2.2 Report describing results of validation of signal processing techniques by using experimental data (T2.3) this report was issued by the Consortium, the lead partner is BAM
Work Package 3: Development and validation of damage diagnosis technology
T3.1 Develop damage diagnosis technology
Cranfield, BAM and UBrun in collaboration with all Consortium Partners successfully developed and implemented novel diagnosis technologies. The technologies are based on pile excitation using the instrumented hammer and a portable shaker and processing of pile responses using the novel nonlinear nonstationary FRF based on the higher order spectra, novel nonlinear nonstationary higher order spectra, novel approach based on amplitude and phase extraction from the higher order spectra, novel damage diagnosis method based on impulse response function and the novel anomaly decision making for final pile diagnosis.
T3.2 Validate damage diagnosis technology using trials data
Cranfield, BAM and UBrun successfully performed experimental validation of diagnosis technologies using experimental data captured at Aarsleff (UK) and BAM (Germany) with damaged and undamaged industrial piles . The first obtained results showed high effectiveness of pile damage diagnosis technologies
Cranfield has communicated the results of preliminary field test at the Aarsleff site to BAM and all other partners. They have been discussed by the consortium. BAM has performed traditional PIT test on some of these piles (see D 1.3, T3.3, travel of BAM to Newark). Cranfield’s results are successful and very promising with the estimates of the total probability of correct diagnosis at 100%. The low cost shaker system, which was used by BAM, UBrun and Cranfield (see WP4) has been validated at BAM test site. The test results show good identification results on defective piles with precise defect locations.
T3.3. Perform pile diagnosis by the traditional pile integrity test (PIT)
Measurements using the traditional pile integrity tests (PIT) have been performed on piles at the BAM test site in Germany, the Aarsleff test site in the UK and the BMNED test site in the Netherlands. The results were reported to the consortium as well as in D1.3 and D3.2 (for the Aarsleff and BAM test sites) and compared to the preliminary shaker test data.. Tests at the Deltares and BMNED test sites should be followed plus others if appropriate after the project end at an exploitation stage. The task depends on the availability of test sites. Some of these sites are external (not run by the project partners), so delays are possible.
T3.4 Compare damage diagnosis results estimated by the diagnosis technology developed in T3.1 with fault conditions established by traditional NDT techniques
The comparison of data available from Newark for traditional PIT and the newly developed diagnosis technology was performed by BAM and Cranfield with input mainly from GSP. The results were in favour for the new technique (gain in probability of correct diagnosis is 50%) but in so far inconclusive, as the PIT test results were debated and GSP believed that PIT results from Newark could be improved. The comparison results available at that point were reported in D 3.2. It was agreed that the results had to be updated in one of the deliverables in WP5. The improved PIT tests results were later updated and reported by GSP, the comparison was updated by BAM and Cranfield (still in favour of the new technique, but with less distance, i.e. 36%gain instead of 50% gain .
D 3.1 Vibration damage diagnosis algorithms (T3.2): this report was issued by the Consortium, the lead partner is CRANFIELD.
D 3.2 Report on comparison of damage diagnosis results provided by the PileInspect technology and the traditional PIT (T3.4): this report was issued by the Consortium, the lead partner is BAM.
Work Package 4: PileInspect software and hardware development
T4.1 Integrate diagnosis and damage size estimation technologies in open system architecture software platform
BAM set up a low cost vibrator system based on previous developments and components available ready to use on the market. The vibrator is based on a shaker which is used for low frequency sound and vibration generation in cars, home entertainment or cinema applications. Based on existing proprietary control software, a prototype LabVIEW program was developed for source signal generation, instrument control and data acquisition, including data pre-processing and display capabilities. Commercial IEPE accelerometers are used for vibration measurement. Preliminary tests have been performed successfully at the BAM test site. Hardware specifications were made available to the consortium. The source code of the prototype software was made available to UBrun. All parts of the setup are going to be improved in the reminder of the WP and thereafter.
UBrun has upgraded the software platform based on the prototype LabVIEW software developed by BAM by developing toolboxes/modules of signal generating, waveform displaying and data processing; continuously debugging and optimizing the software based on filed test by UBRUN and BAM. The function of signal generating has been developed to create two types of sine sweep signals. Four parameters are required in this function: T: period; FS: sampling frequency; F1: starting frequency; F2: stopping frequency. A pre-ringing module is used to check the pre-ringing effects of sweep signals. These modules were developed by using ‘ActiveX’ as interface of external code for LabVIEW: Firstly, Matlab codes are converted to ‘generic com component’ by ‘deploytool’ of Matlab, and then called by ‘Automation Open Function’ of LabVIEW.
UBrun developed pile diagnosis software based on impulse response measurement with sine sweep excitations. The MATLAB batch processing program was developed to process automatically and in batches of a series of all data collected and saved in a folder. The processed data and figures can be automatically saved. The GUI LabVIEW program was developed to process individually the data set in the data folder. For each data set, the selection of signal section and parameter setting are customized manually for signal processing. A stabilization factor is used for regularization during IRF calculation. The BAM IRF calculation software developed within the project was used for comparison.
Cranfield has successfully developed software of technologies and performed the integration of the diagnosis and damage size estimation technologies into the open system architecture for condition based monitoring (OSA CBM) software platform. In the OSA CBM structure the software platform represents the data manipulation (DM), the state detection (SD) and the health assessment (HA) layers. The software is realised in the MATLAB as a standalone application with its own interface (by agreement with all Partners). The software provides means for the pile damage diagnosis technology training, pile damage diagnosis using the user-defined data and parameters, automated decision making, damage size estimation and reporting. The total amount of code in terms of the MATLAB programming environment exceeds 2000 lines.
T4.2 Select, design and integrate hardware components
Selection strategy for the components of the hardware prototype is based on a balance between the performance and price consideration, including the efforts required for possible software development. Two components of the prototype, which may require software development/change are the data acquisition card and shaker controller. Selection of other components does not affect the software, therefore, general considerations are applied at the selection, such as range of key parameters, price and weight. Some of the hardware units do not have special requirements (i.e. laptop PC, battery, and invertor).
Based on the above selection criteria and strategy, the specifications of the PileInspect prototype are determined as follows:
• Bandwidth for all acquired signals at +/-5% level is from 0.5 Hz to up to 10 kHz (can be adjusted in the data acquisition software).
• Accelerometer and force transducer sensitivities: 100 mV/g (pile accelerometers), 10 mV/g (feedback accelerometer), 4pc/N (instrumented hammer). All sensors are monoaxial.
• Maximum aAcceleration: 50 g pk (pile accelerometers), 500 g pk (feedback accelerometer).
• Maximum force from the instrumented hammer: 60 kN (force sensor limit).
• Shaker specifications: frequency range 20Hz – 800Hz; maximum program force 978 N, weight 1.6 kg.
• Sampling rate per channel: up to 51 kHz (software selectable).
• Data acquisition card parameters: 4 channels, 24 bit resolution.
• Antialiasing filters in each channel with variable cut-off frequency (adjusted automatically according to the selected sampling frequency) and stopband slope of 150 dB/oct.
• Shaker controller, which provides the possibility to generate impact of various shapes and durations; constant and variable frequency (sweep) sine excitation with a flexible control over all parameters.
• The prototype ability to estimate the pile length, diagnose the pile damage, estimate the damage severity and location.
Using different configurations of the PileInspect prototype, four types of integrity testing technologies can be carried out with different hardware configurations, including three technologies are based on the FRF of HOS and HOS and implemented by Cranfield; and one technology is based on IRF and implemented by UBRUN and BAM:
• The pile integrity testing technology based on the FRF of HOS and shaker swept sine excitation (Technology 1).
• The pile integrity testing technology based on the HOS and shaker impact excitation (Technology 2).
• The pile integrity testing technology based on the HOS and hammer impact excitation (Technology 3).
• The pile integrity testing technology based on the IRF (Technology 4).
T4.3 Integrate software and hardware into prototype of the system
The assembled prototype integrated software and hardware platforms which were developed in Tasks T4.1 and T4.2. The integrated PileInspect system is composed of:
• Hardware components including accelerometers, hammer, shaker with amplifier and controller, data acquisition card and power supply.
• Software modules including data acquisition, shaker controlling, and pile diagnosis algorithms based on HOS and IRF methods.
• Data transfer and communications among the hardware components and software modules.
The PileInspect prototype also features the pile excitation control implemented in a closed loop. This feedback loop is implemented with the help of the shaker controller communicating with the control software module. The loop is closed by providing the shaker controller with measurements of the actual pile excitation level with using the shaker accelerometer.
The data acquisition and control software modules provide the data and parameters required for proper operation of the diagnosis software module, where the main signal processing occurs, resulting in the diagnosis decisions on the pile damage.
The comprehensive laboratory tests have been carried out to verify the proper functioning of the integrated PileInspect prototype under laboratory conditions prior to the prototype trials at the construction sites.
The results of the laboratory testing confirmed that the hardware part of the prototype, controlled by the corresponding software, is able to generate the expected excitation signals and produce the excitation force required for the pile diagnosis technologies. The data acquisition hardware under the control of the acquisition software was verified to be able to acquire, display and record all the corresponding signals in a proper format with the adequate precision and with the correct content.
It was also confirmed that the software of the assembled prototype correctly realises all the functions implemented by the developers: correctly reads and writes of files from and to the predefined locations, performs the pile diagnosis correctly and in accordance with the specified parameters using the data from the preliminary performed field trials with confirmed and known damage indicators; displays properly all the messages during the diagnosis procedure and saves all the expected diagnosis results in the prescribed format. Therefore, the results of the laboratory testing confirm that the PileInspect prototype, as the whole system, functions correctly.
D4.1) Prototype open system architecture software platform (T4.1): this report was issued by the Consortium, the lead partner is UBrun
D4.2) Prototype hardware (T4.2): this report was issued by the Consortium, the lead partner is UBrun
D4.3) Assembled prototype (T4.3): this report was issued by the Consortium, the lead partner is UBrun
Work Package 5: Field trials, demonstration and optimisation
T5.1 Install prototype of the system on piles under field conditions
The operation of the hardware prototype assumes fixing the shaker and the accelerometers on the pile’s surfaces. Cranfield performed attachment by screws in order to provide a rigid link between the prototype hardware units (i.e. shaker and accelerometers) and the industrial piles at the pile testing site at Newark (UK) and Kutno (Poland). In order to obtain a good response at high frequencies, a metal anchor with inner thread for accelerometer attachment by threaded studs is used during the pile integrity testing based on HOS and FRF of HOS; for the shaker attachment, plastic plugs are used in addition.
Alternatively, magnetic mounts can be used to attach the accelerometers without drilling. The washers are attached to the polyamide by superglue. Low cost standard bulletin board magnets were used. The sensors are screwed to metal cubes where the corresponding thread of the sensor screw was cut in. The mounting technique proofed fairly stable during the tests at BAM site (Germany) and BMNED site (Netherlands) by BAM and UBrun. After the tests, the washer can easily be removed from the pile by applying a quick hammer impact.
T5.2 Perform a range of tests to validate prototype performance compared to trial results
Field trials were successfully carried out at different sites, on different types of piles in terms of shape at different sites, depths, cross-section areas and site conditions. The PileInspect systems have been tested at the following two sites:
• Kutno site in Poland with precast concrete piles
• Horstwalde site in Germany with cast-in-situ concrete piles
During all tests the conventional pulse echo method (PIT) was used. These tests were performed, interpreted and reported to the consortium by GSP, BAM and Cranfield. PIT tests at Kutno were done as blind tests by Polish experienced pile testing company Metris. These PIT tests allow a comparison of the PIT results to the results obtained using PileInspect technologies. The trial results were also used to evaluate the performance of the developed system and to debug and optimise the final prototype design.
Cranfield in collaboration with Aarsleff Poland developed a comprehensive library of damaged and undamaged industrial piles using the standard industrial pile installation. The site offers a library of piles of the same cross-section. Cranfield has successfully performed the range of tests to validate the prototype performance at Kutno for nine industrial piles. The test results have shown 100% total probability of correct diagnosis when the PileInspect technology is used. The traditional PIT test has been performed on the site too as a blind test. The effectiveness of the PIT test in terms of the total probability of correct diagnosis was estimated as 42%. Based on this comparison, the PileInspect system offers an essential gain of the pile damage diagnosis efficiency, estimated as 58% in terms of the total probability of correct diagnosis.
PileInspect technology results obtained previously in field conditions were also compared with the new PIT test results at Newark obtained by GSP. Based on this comparison, the PileInspect system offers an essential gain of the pile damage diagnosis efficiency, estimated as 36% in terms of the total probability of correct diagnosis.
UBrun and BAM carried out field trials at Horstwalde site (maintained by BAM) in Germany to validated and evaluate the PileInspect prototype system using pile diagnosis based on the impulse response function (IRF) method. PileInspect prototype has been tested on 6 test piles including one undamaged pile and 5 damaged piles. Satisfactory pile diagnosis results were achieved from the system and method: Pile toe and defects are clearly detectable and the depths of pile defects can be identified with high resolution and precision.
T5.3 Debug and optimise prototype design
Cranfield in collaboration with UBrun has actively participated in the debug and optimisation of the prototype design. The sampling rate was reduced to 12.8 kHz thus allowing for reduced amount of storage for recorded data in reduced time for the decision making. The functionality of the prototype has been optimised, providing the end user with access to a wide range of the signal processing parameters. This improves the compatibility of software with alternative options of hardware within the prototype, if the modifications will be introduced into the prototype by the end user.
UBrun optimized the pile diagnosis software by integrated two signal processing algorithms into a user friendly integrated Matlab GUI executive package. To verify the final system design of hardware and software, UBrun and Cranfield did lab tests at UBrun facilities and the system matched well with the specifications.
D5.1 Trial results (T5.2): this report was issued by the Consortium, the lead partner is UBrun.
D5.2 Final System Design (T5.3): this report was issued by the Consortium, the lead partner is Cranfield.
Work Package 6: Technology exploitation and dissemination of results
T6.1: Protection of IPR.
UBrun carried out comprehensive patent searches to update the Intellectual Property landscape and assess the viability of the expected patent applications. D6.7 and D6.8 covered the results of patent and novelty search carried out up to date. If required and viable, patent applications will be prepared and submitted in the following deliverable D6.9. After comprehensive investigation of the searched patents and compared with the standard pile integrity methods, it can be concluded that no patents are related to the Pileinspect methods. There are significant differences between any patented methods and the proposed and implemented technologies as part of the PileInspect project. The consortium believes that the novel technologies proposed and implemented for the project represent a unique solution to the problem.
T6.2: Commercial review
The commercial objectives of the PileInspect system were successfully validated by all the SME-AG in terms of purchase cost, saving testing time and increase in productivity.
T6.3: Knowledge transfer from the RTDs to the SME-AGs
UBrun has drafted the Plan for Use and Dissemination of Foreground (PUDF, D6.1) and Plan for knowledge transfer to SMEs (D6.3). The aim is to maximise exploitation of project results and translate them into commercial benefits for the consortium and Europe as a whole and ensure that the knowledge, research and results of the project is reiterated and passed on to the SME-AGs & SMEs appropriately.
T6.4 Training of the SMEs by the SME-AGs
UBrun established project public website to disseminate the technology developed and consortium events during the project as widely as possible, especially for EU SME-AGs. As reported in D6.4, UBrun has created and is responsible for the maintenances of project website: http://pileinspect-project.eu/. A Secure member area is setup to download deliverables, meeting minutes and other documents for the consortium partners.
UBrun and DFI invited industrial members from UK, Europe and USA to attend the PileInspect workshop on 17th November 2016. UBrun hosted the dissemination event. DFI and GSP introduced current pile integrity testing practice and limitations. UBrun, Cranfield and BAM presented project achievements and prototype demo to the attendances. The PileInspect technologies and developed system show great potentials and excited strong interests from the industrial members.
A number of papers and articles have already been submitted for publication as follows:
• Nowe metody badań w ofercie PILETEST, GEOINŻYNIERIA / ARTYKUŁ PROMOCYJNY, 2014
• International Conference of Condition Monitoring, Oxford, UK, 2015
• MAROVISZ National NDT Conference in Hungary, 2015
• AEND National Conference on NDT in Sevilla, 2015.
• Ertel, Jens-Peter, Ernst Niederleithinger, and Maria Grohmann. "Advances in pile integrity testing.", Near Surface Geophysics (2016).
• Ertel, Jens-Peter, Ernst Niederleithinger, and Maria Grohmann. "Old ideas to improve pile integrity testing revisited and optimized." NDT-CE 2015 - International symposium non-destructive testing in civil engineering (Proceedings) (2015).
• H. Zheng, V. Kappatos, Ernst Niederleithinger, Jens-Peter Ertel, Maria Grohmann, C. Selcuk, T.-H. Gan. "Defect detection in concrete pile using impulse response measurements with sine sweep excitations." NDT-CE 2015 - International symposium non-destructive testing in civil engineering (Proceedings) (2015).
D6.1) Draft Plan for Use and Dissemination of Foreground (PUDF) including future funding strategy. The report has been issued by the consortium. The lead partner for this is DFI.
D6.3) Plan for knowledge transfer to SMEs. The report has been issued by the consortium. The lead partner for this is DFI.
D6.4) Project website. The website has been set up by UBRUN. The lead partner is HANDT.
D6.5) Wikipedia page on the project and results. The page has been created and is updated. The lead partner is HANDT.
D6.6) Video clip. This has been performed by UBRUN. The lead partner for this is HANDT.
D6.7) Patent and novelty search. This has been performed by UBRUN. The lead partner for this is DFI.
D6.8) Draft Report on patents. The report has been issued by the consortium. The lead partner for this is DFI.
Work Package 7: Consortium management
T7.1 Legal and Contractual Management
It was ensured by the Coordinator that legal relations with PO / REA were maintained smoothly and handled efficiently. The contractual requirements have been met without significant deviations.
T7.2 Management of Consortium Agreements
Grant Agreement together with DoW as well as Consortium Agreement were signed in due time. There was no need for change of these documents.
T7.3 Liaison with the REA
Managerial activities were conducted by the Coordinator, mainly with the project’s Technology manager, and in case if it was necessary, with WP leaders and key researchers. In some points Steering Committee meetings had to be called together to ensure timely provision of deliverables.
T7.4 Financial and Administrative Management
During the project financial management it was necessary to deal with an exchange rate loss of the grant. Reason was that the Coordinator’s bank account was led in HUF and the grant arrived in EUR. After first payment and until invoices were collected, the HUF / EUR rate changed dramatically. At the end of a quite long iteration project partners agreed upon the way of managing the loss.
T7.5 Risk management
Project risks were captured and assessed at the beginning of the work in a project risk assessment with enough contingency in place. Project coordinator has ensured that when required these would be correctly implemented.
T7. 6 Progress monitoring
Monitoring project progress was done systematically. After circulating among partners, draft version of reports and documents were sent to Coordinator for final check. Project progress review meetings were organized in approximately every six month. Altogether seven meetings were held as follows:
1. Cambridge – January 2014
2. Berlin – June 2014
3. Madrid – January 2015
4. Warsaw – August 2015
5. Mannheim – March 2016
6. Cranfield – June 2016
7. Budapest – November 2016
T7.7. Internal administration
Project partners made their own administration, meeting minutes were completed and uploaded on the website.
D7.1) Interim Progress Report (T7.6): The Consortium delivered the report with the lead partner was HANDT.
D7.2) Final Progress Report (T7.6): The report has been issued by the consortium. The lead partner was HANDT.
D7.3) Signed Consortium Agreement: The report has been issued by the consortium. The lead partner was HANDT.
D7.4) Exploitation Agreement: The report has been issued by the consortium. The lead partner was HANDT.
Project Results:
1. Work progress and achievement of Cranfield in collaboration with Aarsleff (UK) and Aarsleff (Poland)
(T1.1) Cranfield and Aarsleff successfully developed a unique comprehensive library (i.e. 15 piles) of damaged and undamaged industrial piles at Aarsleff, Newark, UK using the standard industrial pile installation. These piles are of different cross-sections and length. Cranfield and Aarsleff in collaboration with all partners performed successful testing of all piles using the instrumented hammer and two types of shaker: Data Physics shaker and a low cost portable shaker.
(T1.2) Cranfield in collaboration with Partners created two types of simulation: dynamic simulation using the bilinear dynamical system and finite element simulation using linear and nonlinear finite element analysis, including novel methodology of finite element modelling of damaged piles. The software for both simulations was developed. Focus of Cranfield’s simulation was on geometries resembling Aarsleff’s test site. A comparison of the simulation results and real test data showed good agreement. Diagnosis results of piles by the simulation data and by the experimental data also showed good agreement.
(T1.3) Cranfield (in cooperation with Aarsleff) have successfully generated shaker test data for damaged and undamaged piles at Aarsleff using two types of shaker: Data Physics shaker and a low cost portable shaker (suggested by BAM as an alternative to the Data Physics shaker). Two types of excitation were generated by both shakers: the swept sine excitation and impact excitation. Various sensor arrangements and various locations of shaker excitations were used. The data has been reported and are available to the Consortium.
(T1.4) Cranfield implemented and successfully validated by the simulated and the experimental (in cooperation with Aarsleff) data the novel higher order spectra: the wavelet higher order spectra and the FRF of chirp-Fourier higher order spectra for stationary and non-stationary pile vibrations using the swept sine and impact excitations. Results were obtained for different pile vibration modes, different intensities of excitation, different positions of accelerometers and different locations of impact and swept sine excitations from the instrumented hammer and two types of shakers.
(T1.5) Cranfield implemented and successfully validated by the simulated and the experimental (in cooperation with Aarsleff) data the novel amplitude and phase extraction techniques from the stationary and non-stationary higher order spectra: the wavelet higher order spectra and the FRF of the chirp Fourier higher order spectra. The implemented approach was compared with the standard amplitude approach and gain in effectiveness of diagnosis was estimated for various types of excitation.
(T1.6) Cranfield implemented and successfully validated by the simulated and the experimental (in cooperation with Aarsleff) data the novel anomaly detection for pile damage diagnosis. The main advantage of the method is that it can be used in the cases for which training data only from undamaged piles are available. The method was applied for diagnostic features based on the wavelet higher order spectra and on the FRF of chirp Fourier higher order spectra for the impact and the swept sine excitations respectively. The gain in effectiveness of diagnosis was estimated for various types of excitation.
(T2.1, T2.2) Cranfield (in collaboration with Aarsleff) successfully performed trials with piles in undamaged and damaged conditions at Aarsleff Newark site, UK. Trials included usage of three sources of excitation force: the instrumented hammer, Data Physics shaker and portable low cost shaker. Multiple tri-axial accelerometers were installed in different locations of piles. Excitation was performed in different directions; axial and horizontal. Low and high intensity excitations were employed.
(T2.3) Processing of trial data from damaged and undamaged conditions has been successfully performed by Cranfield in collaboration with Aarsleff. The advanced novel signal processing techniques were employed: the wavelet higher order spectra for impact excitations from the instrumented hammer and shaker and the FRF of chirp Fourier bicoherence for the swept sine excitation from the shaker. The results are very successful and show high effectiveness of detection of damaged and undamaged conditions.
These results were achieved for the first time in worldwide terms. All Consortium Partners contributed by providing the feedback.
(T3.1) Cranfield has successfully implemented the novel and important part of the whole PileInspect pile damage diagnosis technology. It consists of a three stage decision making technology for final diagnosis, based on a fusion of anomaly detection results from multiple damage detection technologies using cluster analysis and the k-nearest neighbour technique and comprehensive statistical post-processing implementing sequential majority analysis. The damage detection technologies involve pile excitations using the instrumented hammer and the portable shaker and processing of the corresponding pile responses using the novel non-linear non-stationary higher order spectra and FRF of the higher order spectra and the novel approach based on amplitude and phase extraction from the higher order spectra.
(T3.2) Cranfield and Aarsleff have successfully performed experimental validation of the Cranfield’s pile damage diagnosis technology using experimental data captured at Aarsleff (UK). The results have shown the high effectiveness of the developed pile damage diagnosis technology: the estimated total probability of correct diagnosis is 100% for all tested piles. The results were communicated to the Consortium and discussed by all Partners.
(T3.3) Cranfield was involved in discussion of the preparations to the PIT test (performed at Aarsleff site, Newark, UK first by BAM and later by GSP) and actively participated in the discussion and interpretation of the PIT test results. After extensive discussion the PIT test results in their final form were used for comparison between the pile damage diagnosis by PileInspect technologies and by traditional PIT test.
(T3.4) The comparison between the results of the pile damage diagnosis by Cranfield’s PileInspect technologies and by traditional PIT test was successfully performed by Cranfield. Cranfield proposed the comparison methodology, which was communicated to the Consortium and discussed by all Partners. The technology is based on estimation of the total probability of correct diagnosis based on the diagnosis decisions ‘damaged’-‘undamaged’ specific for the PileInspect technologies and for traditional PIT test. The comparison has shown a significant gain in the total probability of correct diagnosis when the PileInspect technologies are used for the pile damage diagnosis instead of the traditional PIT test: 100% probability of correct diagnosis for the PileInspect technologies and 50% for the first PIT test and 64% for the repeat PIT test. Therefore, the Cranfield’s PileInspect technologies for the pile damage diagnosis provide at least 36% gain in the total probability of correct diagnosis in comparison to the traditional low strain pile integrity testing.
(T4.1) Cranfield has successfully performed the integration of the diagnosis and damage size estimation technologies into the open system architecture for condition based monitoring (OSA CBM) software platform. In the OSA CBM structure the software platform represents the data manipulation (DM), the state detection (SD) and the health assessment (HA) layers. The software is realised in the MATLAB as a standalone application with its own interface (by agreement with all Partners). The software provides means for the pile damage diagnosis technology training, pile damage diagnosis using the user-defined data and parameters, automated decision making, damage size estimation and reporting. The total amount of code in terms of the MATLAB programming environment exceeds 2000 lines.
(T4.2) Cranfield has participated in the selection, design, and integration of the hardware components for the PileInspect prototype. Cranfield has performed the analysis of compatibility of the components (accelerometers and data acquisition card) selected by BAM and UBrun with PileInspect damage diagnosis technologies developed by Cranfield. Cranfield has also performed selection of the instrumented hammer and shaker controller system based on the analysis of their performance parameters and price. Cranfield proposed ways of accelerometers and shaker fixing on the pile surface, suitable for the developed technologies and methodology for the loss minimization during the shaker connection to amplifier.
(T4.3) Cranfield has actively participated in the integration of the software and hardware into the prototype of the system, which were performed in close cooperation with UBrun. The propositions as for optimised integrated software interface, improved performance control loop for the shaker control, laboratory test rig design and the laboratory test program were given by Cranfield. Cranfield has also made a major contribution into the laboratory testing in terms of acquiring, processing and analysing of the test data, and developing the methodology for laboratory testing of the prototype.
(T5.1) The part of the prototype of the PileInspect system realising the pile damage diagnosis technologies was installed at Kutno site of Aarsleff Poland. Cranfield in collaboration with Aarsleff Poland developed a comprehensive library of damaged and undamaged industrial piles using the standard industrial pile installation. The site offers a library of piles of the same cross-section. The installation was performed in accordance to the requirements to the standard industrial pile installation and interconnections of the integrated PileInspect prototype, developed in the Task 4.3.
(T5.2) Cranfield has successfully performed the range of tests to validate the part of prototype performance at Kutno. The test results have shown 100% total probability of correct diagnosis when the PileInspect technology is used. The traditional PIT test has been performed on the site as a blind test by the experienced polish pile testing company. The effectiveness of the PIT test in terms of the total probability of correct diagnosis was estimated as 42%. Based on this comparison, the PileInspect system offers an essential gain of the pile damage diagnosis efficiency, estimated as 58% in terms of the total probability of correct diagnosis.
(T5.3) Cranfield has actively participated in the debug and optimisation of the prototype design. The sampling rate was reduced to 12.8 kHz thus allowing for reduced amount of storage for recorded data in reduced time for the decision making. The functionality of the prototype has been optimised, providing the end user with access to a wide range of the signal processing parameters. This improves the compatibility of software with alternative options of hardware within the prototype, if the modifications will be introduced into the prototype by the end user.

2. Work progress and achievement of BAM
Cranfield and Aarsleff successfully developed a unique comprehensive library (i.e. 15 piles) of damaged and undamaged industrial piles at Aarsleff, Newark, UK. BAM visited Aarsleff’s site and performed PIT testing of piles at Aarsleff. BAM made the BAM-TTS (Test site Technical Safety, www.tts.bam.de) available to the project. It comprised more than 20 bored piles in diameters from 60 to 90 cm, length 8.5 to 12 m, various flaws and measurement conditions. The documentation was made available to the consortium. The site was visited by the consortium during the 2nd project meeting. In WP4, UBRUN and BAM performed field test at this site to validate the developed PileInspect prototype and pile diagnosis method using the impulse response function.
BAM negotiated successfully to make an additional test site (property of Deltares, Utrecht, NL) available to the project. Cooperation with BMNED was very helpful in achieving this. This site has more than 20 piles with various shapes, lengths and faults available. BAM visited the site and performed preliminary tests with the traditional PIT and the BAM low cost shaker (see WP4) in 2015.
To produce test data for the new signal processing software as well as to assist the hardware development, BAM carried out numerical simulations. BAM used a finite integration approach in cylindrical coordinates (CEFIT) to cross validate the simulations of Cranfield and UBrun. The software used was originally developed in the frame of the FP5 project RUFUS.
PileInspect prototype was optimized in speed and storage usage as well as to allow custom source inputs (e. g. shaker sweeps). This was done in cooperation with the original author (Dr. Frank Schubert of the Fraunhofer society). BAM participated in a German geophysical conference to acquire knowledge about the latest simulation tools. As promised in the DoW simulations were performed with various impulse shapes and swept sine excitations for a variety of pile shapes, flaw geometries and soil conditions. Cross validation against the 1D finite element solution of UBrun showed good agreement. Focus of the BAM simulations was on geometries resembling the BAM test site. A comparison of the simulation results and real test data showed sufficient agreement. Data became available to all partners and have so far been exchanged mainly with UBrun.
BAM set up a new low cost shaker system as an add-on to the high power shaker system used by Cranfield. The new system was used to generate test data mainly on piles at the BAM test site on undamaged and damaged specimens. Various linear and logarithmic sweep signals were tested, differing in frequency content, durations and amplitude. Various sensor arrangements were used.
BAM and UBrun tested various ways to perform pile diagnosis. Most promising was to use a deconvolution algorithm, which had to be combined with a regularization term to provide stability. Based on results acquired with simulated data, BAM favours Tikhonov regularization. Results achieved were reported in D1.3. Validation followed in WP3 as well as work to include the findings in the final software (WP4).
BAM set up a low cost vibrator system based on previous developments and components available ready to use on the market. The vibrator is based on a shaker system which is used for low frequency sound and vibration generation in cars, home entertainment or cinema applications. Based on existing proprietary control software a prototype LabVIEW program was developed for source signal generation, instrument control and data acquisition, including data pre-processing and display capabilities. Commercial IEPE accelerometers are used for vibration measurement.
Preliminary tests have been performed successfully at the BAM test site. Hardware specifications were made available to the consortium. The source code of the prototype software was made available to UBrun. UBrun developed pile diagnosis software based on impulse response measurement with sine sweep excitations. BAM IRF calculation software developed within the project was used for comparison.
BAM and UBrun carried out field trials at Horstwalde site (maintained by BAM) in Germany to validated and evaluate the PileInspect prototype system using pile diagnosis based on impulse response function (IRF) method. PileInspect prototype has been tested on 6 test piles including one undamaged pile and 5 damaged piles. Satisfactory pile diagnosis results were achieved from the system and method: Pile toe and defects are clearly detectable and the depths of pile defects can be identified with high resolution and precision.

3. Work progress and achievement of UBrun
To simulate vibration signals from damaged and un-damaged piles, UBrun built a theoretical 1D soil-pile system. The pile-soil system modelled the soil stiffness and radiation damping with a set of springs and dashpots. On the lateral surface of the pile, distributed springs with elastic stiffness and dashpots with damping coefficient were applied to each segment below the ground surface. The soil at the base of the pile was modelled using a spring with elastic constant and a dashpot with damping coefficient.
The finite difference method is used to solve the motion equation and simulate the pile response to excitations. Verities of scenarios were calculated in the cases of different soil properties, pile properties/geometrics. Two types of excitations are considered for numerical simulations: impact excitation by a hammer and sine sweep excitations by a shaker. The simulated travel times of reflection wave at both pile bottom and pile defects match well with the theoretical values. The simulations and results were used for validation with test data, which were reported in D 2.2.
UBrun and BAM proposed and used an impulse response measurement based method for pile damage diagnosis, the details of which were reported in D1.1 and D2.2. The PileInspect system replaces instrumented hammer by a portable low-cost shaker, which can excite the piles with repeatable and tailored sine sweeps. The echoes at pile toe and pile defects are visible in the time domain records in the straightforward impact-echo methods.
However, these echoes are not visually available in the time domain records of piles subjected to sine sweep excitations. We propose to extract impulse response function from input (shaker excitation) and output measurements (pile response). This technique is similar to Green’s function or receiver-function computations commonly used in seismology for determining source-time functions or for identifying subsurface structures after removing source effects.
The impulse response functions were computed by deconvolving the pile response from the sine sweep excitations. The reflection wave in the pile, i.e., the echoes from pile toe and pile defects with impedance change, can be identified from the impulse response function of a pile in the time domain. To refine identifying resolution, UBrun and BAM tried varieties of parameters of input sweep signal: including sweep signals types, frequency bandwidth and sweep’s time length. The simulation results were further validated with experimental data. The proposed methodology was reported in D1.1 and validated with pile test data obtained by BAM (described in D2.2). In PileInspect project, the impulse response function based method was used for pile diagnosis tool to provide quality information of the pile for assessment of pile integrity.
UBrun, together with Cranfield and BAM developed and integrated PileInspect prototype, including hardware and software platforms, for pile diagnosis methodologies based on higher order spectra (HOS) and impulse response function (IRF) methods. UBrun has upgraded the software platform based on the prototype LabVIEW software developed by BAM by developing toolboxes/modules of signal generating, waveform displaying and data processing. The function of signal generating has been developed to create two types of sine sweep signals (Muller type and Farina type). A pre-ringing module is used to check the pre-ringing effects of sweep signals.
UBrun developed pile diagnosis software based on impulse response measurement with sine sweep excitations. The MATLAB batch processing program was developed to process automatically and in batches of a series of all data collected and saved in a folder. The processed data and figures can be automatically saved. The GUI LabVIEW program was developed to process individually the data set in the data folder. For each data set, the selection of signal section and parameter setting are customized manually for signal processing. A stabilization factor is used for regularization during IRF calculation.
UBrun optimized the pile diagnosis software by integrated two signal processing algorithms into a user friendly integrated Matlab GUI executive package. UBrun and Cranfield carried out laboratory tests to verify the proper functioning of the integrated PileInspect prototype under laboratory conditions prior to the prototype trials at the construction sites. The results of the laboratory testing confirmed that the hardware part of the prototype, controlled by the corresponding software, is able to generate the expected excitation signals and produce the excitation force required for the pile diagnosis technologies. The data acquisition hardware under the control of the acquisition software was verified to be able to acquire, display and record all the corresponding signals in a proper format with the adequate precision and with the correct content.
It was also confirmed that the software of the assembled prototype correctly realises all the functions implemented by the developers: correctly reads and writes of files from and to the predefined locations, performs the pile diagnosis correctly and in accordance with the specified parameters using the data from the preliminary performed field trials with confirmed and known damage indicators; displays properly all the messages during the diagnosis procedure and saves all the expected diagnosis results in the prescribed format. Therefore, the results of the laboratory testing confirm that the PileInspect prototype, as the whole system, functions correctly.
UBrun together with BAM carried out field trials at Horstwalde site (maintained by BAM) in Germany to validated and evaluate the PileInspect prototype system using pile diagnosis based on impulse response function (IRF) method. PileInspect prototype has been tested on 6 test piles including one undamaged pile and 5 damaged piles. Satisfactory pile diagnosis results were achieved from the system and method: Pile toe and defects are clearly detectable and the depths of pile defects can be identified with high resolution and precision.
UBrun cooperated with HANDT and DFI to carry out technology exploitations and disseminations:
• UBrun has drafted and finalized the deliverables: Plan for Use and Dissemination of Foreground (PUDF, D6.1), Plan for knowledge transfer to SMEs (D6.3) and Final Plan for Use and Dissemination of Foreground (PUDF) (D6.2). The aim is to maximise exploitation of project results and translate them into commercial benefits for the consortium and Europe as a whole and ensure that the knowledge, research and results of the project is reiterated and passed on to the SME-AGs & SMEs appropriately.
• UBrun setup public website to disseminate the technology developed and consortium events during the project as widely as possible, especially for EU SME-AGs. As reported in D6.4, UBrun has created and is responsible for the maintenances of project website: http://pileinspect-project.eu/. A Secure member area is setup to download deliverables, meeting minutes and other documents for the consortium partners. As reported in D6.5.
• UBrun created video clip to give a brief summary of the project, including an introduction, background, aims, achievements, and practical applications of the PileInspect project. The video clip is uploaded on YouTube to raise awareness of the PileInspect projects to publics and organisations.
• UBrun has investigated a comprehensive study and search of prior art (the body of pre-existing knowledge) relating to the technologies proposed in PileInspect project. D6.7, D6.8 and D6.9 covered the results of patent and novelty search carried out up to date. After comprehensive investigation of the searched patents and compared with the standard pile integrity methods it can be concluded that these patents are all relate to the existing methods. There are significant differences between these patented methods and the proposed technologies in PileInspect project. We believe that novel IRF technology proposed by UBrun for the project represents a unique solution to the problem.
• UBrun and DFI invited industrial members from UK, Europe and USA to attend the PileInspect workshop on 17th November 2016. UBrun hosted the dissemination event. UBrun together with Cranfield and BAM presented project achievements and prototype demo to the attendances. The PileInspect technologies and developed system show great potentials and excited strong interests from the industrial members.
Potential Impact:
UBrun established project public website to disseminate the technology developed and consortium events during the project as widely as possible, especially for EU SME-AGs. As reported in D6.4, UBrun has created and is responsible for the maintenances of project website: http://pileinspect-project.eu/. A Secure member area is setup to download deliverables, meeting minutes and other documents for the consortium partners.
UBrun and DFI invited industrial members from UK, Europe and USA to attend the PileInspect workshop on 17th November 2016. UBrun hosted the dissemination event. DFI and GSP introduced current pile integrity testing practice and limitations. UBrun, Cranfield and BAM presented project achievements and prototype demo to the attendances. The PileInspect technologies and developed system show great potentials and excited strong interests from the industrial members.

A number of papers and articles have already been submitted for publication as follows:
• Nowe metody badań w ofercie PILETEST, GEOINŻYNIERIA / ARTYKUŁ PROMOCYJNY, 2014
• International Conference of Condition Monitoring, Oxford, UK, 2015
• MAROVISZ National NDT Conference in Hungary, 2015
• AEND National Conference on NDT in Sevilla, 2015.
• Ertel, Jens-Peter, Ernst Niederleithinger, and Maria Grohmann. "Advances in pile integrity testing.", Near Surface Geophysics (2016).
• Ertel, Jens-Peter, Ernst Niederleithinger, and Maria Grohmann. "Old ideas to improve pile integrity testing revisited and optimized." NDT-CE 2015 - International symposium non-destructive testing in civil engineering (Proceedings) (2015).
• H. Zheng, V. Kappatos, Ernst Niederleithinger, Jens-Peter Ertel, Maria Grohmann, C. Selcuk, T.-H. Gan. "Defect detection in concrete pile using impulse response measurements with sine sweep excitations." NDT-CE 2015 - International symposium non-destructive testing in civil engineering (Proceedings) (2015).
List of Websites:
http://pileinspect-project.eu/

For more information, Please contact:
Brunel Innovation Centre (BIC)
Granta Park
Great Abington
Cambridge
CB21 6AL
Website: www.brunel.ac.uk/bic
Email: bic@brunel.ac.uk
Tel: +44 (0) 1223899000

Reported by

MAROVISZ HUNGARIAN ASSOCIATION FOR NONDESTRUCTIVE TESTING -MAGYAR RONCSOLASMENTES VIZSGALATI SZOVETSEG TARSADALMI SZERVEZETEK
Hungary
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top