RICERCA ITALIANA     

Research program

Integrated Methods and Algorithms for NonDestructive Evaluations of architectural heritage

University Co-ordinator

Università di PISA - SISTEMI ELETTRICI E AUTOMAZIONE - PISA(PI)

Research Unit Leader

Antonino MUSOLINO

Description

The aim of the research project of this unit is to identify and characterize from a theoretical and experimental point of view a new class of non destructive (ND) test based on guided ultrasonic waves for the inspection of ancient historical buildings of architectural and artistic importance.

Non destructive ultrasonic based techniques have reached a satisfactory usage level in several industrial applications. In the inspection of Cultural Heritage these techniques are currently under study and development in order to obtain a better fitting to the peculiarities of the structures that have to be investigated.
Two are the main interests of he researchers in the conservation and recovery of artistic heritage. The first regards the knowledge of the preservation state and of the structural integrity of historical buildings that can be affected by hidden structural damages occurred in the course of the centuries as consequences of earthquakes or war events. The other regards the characterization and the identification of the materials that constitute the structures, as is not uncommon that artistically valuable frescos are hidden under the surface plaster.
The commercial devices commonly used are of difficult tuning with the specific structure that is under inspection; their user interfaces are not flexible and do not allow the interpretation of the measured data. Furthermore the conventional non destructive ultrasonic based techniques are able to inspect only a narrow zone (though deep) near the point where the excitation is applied and as a consequence a great number of tests have to be performed to inspect the whole structure.
In order to overcome these problems this research unit will investigate the use of an inspection technique based on the ultrasonic guided waves.
Non destructive techniques based on ultrasonic waves are commonly used in the remote inspection of both tubular and planar metallic structures. At present it is used in the framework of a research activity turned to detect and identify defects in metallic pipes. This inspection method that is classified as a long distance technique could be advantageously applied in the characterization of ancient architectural buildings as it allows the inspection of large portions of the structure at great distance from the application of the exciting pulse thus reducing the time needed to complete the investigation.
Moreover this technique could allow the inspection of regions that are not directly accessible, as the elastic ultrasonic waves propagate for long distance and deep inside the structure and can be generated and detected by transducers and sensors that can be positioned at distance one from the other.

This research unit has recently acquired a non destructive test equipment based on the propagation of elastic ultrasonic waves that is currently used for the inspection of metallic structures. This instrument will be adapted for the inspection of stone and masonry buildings in order to examine their structural integrity and to characterize their composition.
The instrumentation is based on a Magnetostrictive Sensor; it is able to generate the propagation of elastic waves in guiding tubular or planar structures and to detect the echoes caused by the presence of defects due to corrosion, fractures or discontinuities in the medium where the waves propagate. The waves that are generated by the transducer are of longitudinal and torsional kind in cylindrical structures and transverse and Lamb kind (symmetric and asymmetric) in planar ones. This technique allows the inspection of three dimensional regions of the order of tenth of meters from the point where the sensor is applied. When used with masonry a reduction of the volume that can be inspected is expected because of the physical constants that govern the propagation in the medium; nevertheless the dimension of the regions that can be inspected should be of the order of the meters.
The instrument is able to operate by using an interface localized with the region under inspections and has a good sensitivity until frequency of the order of hundreds of kilohertz.
A set of transducers and of sensors that are able to operate in the range of the Megahertz will be acquired. They will be driven by external sources and for this purpose an arbitrary function waveform generator will be acquired too. The frequency characterization of the medium where the elastic waves propagate will be performed by the use of a spectrum analyzer.

A widespread use of the proposed methodology would be possible if it was able to give a three-dimensional image of the inspected portion of the building. Furthermore, its efficiency would related to some critical aspects whose analysis will be addressed in the research activity: the choice of the type of optimal technique in terms of the kind of elastic perturbation, its frequency, the location of the transmitter and of the receiver and so on; the setting of efficient numerical models for a better understanding of the more important physical phenomena and the choice of the signal processing methodology for the elaboration and interpretation of the transmitted and received data from the investigated structure.
In particular, this last aspect has a fundamental importance for the possibility of extraction of the information on the investigated structure (geometry, nature of the materials, etc.).

This research Unit, besides the verification of a possible use of a new guided waves ultrasonic technique, will try to analyze some types of signals processing methods in order to identify which of them is suitable for a better extraction of the necessary information for a complete evaluation of the inspected structures. Such signals processing can be performed by the use of different methods. In particular, this Research Unit will be interested in the use of the Wavelet Transform and of the Hilbert-Huang Transform.

The Wavelet transform is a powerful tool for transient signal analysis. Its zooming property makes the resolution of high and low frequency more accurate than other signal processing method. In particular the multi-mode coexistence phenomenon can be framed in the wavelet transform results.
In the Hilbert-Huang transform, the key role is played by the so called "empirical mode decomposition" (EMD), a method that allows a decomposition of any complicated data set into a finite and often small number of "intrinsic mode functions" (IMF) that admit well-behaved Hilbert transform. Utilizing Hilbert transform on those obtained IMFs, it is possible to get a full energy–frequency–time distribution of the signal, designated as the Hilbert–Huang spectrum. This method is adaptive, and, since the decomposition is based on the local characteristic time scale of the data, it is applicable to nonlinear and non-stationary processes. One of the advantages of HHT is that its most computation consuming activity, the EMD operation, does not involve the convolution or other time-consuming operations, therefore the HHT can deal with signals constituted by a large number of samples. Additionally, the Hilbert–Huang spectrum does not involve the concept of the frequency or time resolution but only uses the instantaneous frequency that is able to give a sharp identification of the signal structure.

Activity program
The potentialities of the physical principle will be investigated, both from the numerical point of view and from an experimental one, by the use of numerical codes and by means of several tests on different laboratory specimens and eventually on existing building structures of architectural interest.

Phase 1): The applicability of the ND guided waves ultrasonic technique will be theoretically verified for the inspection of architectural structures in concrete, in stones masonry or tiles; computer codes will be used for studying the elastic waves propagation in non homogeneous media in order to determine among the many parameters that govern the propagation those that have the greatest influence on the phenomenon. The CAPA software recently acquired by this Unit and currently used for a research activity concerning the inspetion of portion of pipes will be used. The frequencies that better characterize the different structure components will be investigated. The numerical models will be used to build a database useful for the classification of the most common defects present in the structures and for the identification of their compositions and constructive typologies.

Phase 2): The data obtained by the simulations performed in the phase 1, and in particular those contents in the signals detected by the sensors, will be processed to reconstruct the characteristics of the investigated structures. For such aim, the most suitable technique of data analysis will be identified for a better extraction of the necessary information for a complete evaluation of the inspected structures. This is a crucial phase of the project proposed by this Unit, since nowadays the researchers' preference for the sonic tests is owed to the simplest interpretation of the experimental data (although of longer execution because of the limited ray of action of such methodology). The techniques based on Wavelet and Hilbert-Huang transforms will be used, compared and integrated among them to get an interpretation as clear and complete as possible on the state and on the constitution of the architectural structures.

Phase 3): Data obtained by the other research units by using methods that differ from the one here proposed often contains information that are complementary to those obtained by using guided ultrasonic waves. A matching of these results will be performed in order to obtain a more precise description of the masonry under testing. This activity will be tightly co-ordinated with the other research units; data fusion and soft computing techniques will be implemented in order to obtain procedures that will be used to perform the training of neural networks.

Phase 4): An experimental activity will be performed in order to validate the results obtained by using analytical or numerical models. Samples of different kind of masonry will be realized and the experimental data obtained by the proposed technique will be compared with the results of the numerical simulation in order to asses the capability of the method to detect the presence of defects and to reconstruct the inner structure of the sample.