Acoustic Emission and Failure Prediction of Composites
Edouard G. Nesvijski and Paulo J. Sarkis
Federal University of Santa Maria, Santa Maria, RS, Brazil
Corresponding Author Contact:
Email: firstname.lastname@example.org, Web: http://w3.ufsm.br/edouard/
- International Symposium on NDT Contribution to|
the Infrastructure Safety Systems, 1999 NOV 22-26 Torres,
published by UFSM, Santa Maria, RS, Brazil
Dramatic development of new structural materials has followed the way of substituting metals for composites in various industries, such as aviation, shipbuilding, chemical-petroleum, civil engineering etc. It is one of the avenues leading to the safety and durability of infrastructure systems. The acoustic emission (AE) method occupies a strong position among other nondestructive methods of testing of composites. The AE method has a diverse instrumentation that is based on a profound theoretical foundation. At the moment it is at the stage of consolidation and standardization based on a rich practical experience in various practical fields of application. Unfortunately, it is impossible to use directly the background and to transfer technology of the AE testing of metals for analysis of composites. It is connected with different physical natures of cracking and failure in metals and composites. That is why modification of old models along with creation of new ones are necessary for explanation of the AE features of composites. This paper will consider various criteria of material failures based on short-term crack resistance and fatigue. Registration of the AE signatures during short-term and low-cycle load tests allows predicting failure of composites. As an example, some experimental results obtained for cement-based composite materials are analyzed. A scope of practical problems of infrastructure safety, which could be resolved by the AE techniques, is presented.
KEYWORDS: infrastructure, safety, failure criteria, composites, acoustic emission.
Acoustic Emission (AE) is a physical process that is one of many reflections of the materials failure. That is why AE has been used for developing of a wide class of nondestructive testing (NDT) techniques for practical applications. The "Great AE Boom" happened between 70's and 80's due to efforts of an enthusiastic group of researchers from industry and universities. A lot of the AE paper appeared during this time. They displayed a great interest of the NDT community to this area of research with a promise of multiple applications in the future [1,2]. Commercialization of the AE NDT applications was observed in Europe, USA and Japan at the same period of time [3,4]. The NDT market forced to change balance between theoretical research and fast-growing the AE application market [5,6]. On one hand, this process affected negatively evolution of the AE basics and it is possible to notice considerable decrease of the AE publications in early 90's. On the other hand, the commercial capital gave opportunity to develop the high-level AE equipment and software. The practical AE applications demonstrated a number of limitations and restrictions of the existing AE methods and equipment that led to several wrong conclusions about possibilities of AE as an NDT instrumentation. This historical retrospective is made to encourage new researchers to go "back to basics" looking for an opportunity of the next "boom" in the field of the AE development and application in the 21st century.
Pulse loads of hetrogeneous materials
AE may be used for two types testing: short term and long term monitoring depending on the monitoring tasks. For example, investigation of corrosion or fatigue of composites requires long term monitoring and may be used without application of special trial loads. Other testing tasks of investigation of stress, strength or cracking resistance of composites require application of trial loads. These trial loads are of two types quasi-static and dynamic. As a rule AE manifests itself differently depending on the physical or chemical processes taking place in the material. It is also necessary to distinguish the AE effects of quasi-static and dynamic loads in composites. It can be explained by the fact that practically all fracture processes in materials develop in time depending on the reological properties of materials. That is why problem of test loads for composites is one of the principal issues.
General modeling of the problem (CARROL-HOLT MODEL as an example)
Carroll-Holt model may be taken for presentation of the short term test loads. The model is based on a one-cell elementary part, which could be integrated to describe the pore material . In this case relation between load pressure and yield point of the described model material could be presented by the following formula
where P - static pressure; Yo - yield point of material; D = ( rp ) / ( r ) ; rp - density of pore; P - density of material.
Two types of Carroll-Holt pore material models are presented in the Figure 1.
Fig 1: Two types of material models: a- pore material which contains other material inside the pores, b- pore material with vacuum or gas inside the pores: and -are geometric parameters of the models.
These models describe only quality relations in the porous materials and are simplified models, which are far from the real material representation. An attempt is made to develop these models to approach reality. Let us start from the fundamental motion equation
where r - density and sgg, sqq are tensors of stress in a spherical symmetric material cell. The following boundary conditions are set up
|sgg(r = a) = 0|
sgg(r = b) = -P(t)
The initial conditions are also set up
After substitution to (1) taking into account (2), (3) and (4) and routine calculation receive value of the coefficients
where r,a,b,c,ro,ao,bo,ao - geometric parameters and calculated coefficients.
Integrating the elementary cell into the material receive modified formulae for describing the time dependent characteristics of the material 
where PCARR - Carroll's time dependent part of pressure in a porous material.
Here Y - yield point of plasticity of a matrix material with pores; h - viscosity .
After simple integrating receive a new dependency between pressure and porosity of the material depending on dynamic load characteristics. Calculation results for this modified model are shown in the Figure 2.
Fig 2: Pressure in the matrix of the porous material during pore collapse for four modulae of elasticity (40,60,80 and 100 GPa)
As the calculations of the modified Carroll-Holt model show pore collapses lead to off-loading of the material and creation of tension effect corresponding to negative pressure - P(t) in the model. The pore collapses and creation of micro-stresses as a result of the off-loading explain relieve of failure energy in the material and initiation of the AE generation as a certain part of this energy.
Cracking models of Composites
Cracking is one of the principal processes leading to the material failure. That is why it is very important to have special criteria and parameters for description and quantitative estimation of cracking. Until now this approach has been well developed for homogeneous materials. Existing system of cracking criteria could be presented by the following set of parameters [9, 10]:
- KC - maximal values of the coefficient of stress intensity. It characterize value of stress at the moment of the cracking initiation. The coefficient is calculated on the base of effective cross section of the specimen and initial size of cracks.
- K1C - critical coefficient of stress intensity, which is defined as critical point of KC for the moment of tri-axis tension for plain strain.
- dC - maximal size of the crack opening at the deadlock of the crack. It defines plastic properties of the material at the crack pick at the cracking initiation moment.
- gE - energy expended on creating of a unit of the free surface area in the material.
- G1C - intensity of free energy or critical value of J - integral.
The relation between these parameters could be described by the following formulae
where m - Poisson ratio; E - Young module; and s0 - limit of brittle strength of the material. Cracking parameters gE, d1C, K1C and J1C are well studied and provided with standardized testing methods. Their values are calculated and tabulated for the majority of metallic, ceramic, and a number of other homogeneous materials .
Direct translation of this approach to estimation of cracking in composites meets a number of problems connected with structural mechanics of these materials. The main problem is different character of the process in different components of the material and additional variations on the borders between them. Attempts of this approach translation for estimation of cracking in regular composites such as linear reinforced or matrix composites have not given stable values of cracking parameters. Heuristic approach has been used for building up cracking estimation parameters for other types of composites . The task of cracking estimation for composites with stochastic structures is even more complicated. There is neither unique nor integral approach for the majority of composite materials. Absence of fundamentally established base for cracking modeling of composites limits application of AE to their failure estimation.
Special stochastic models of failure can be used for composites with irregular or stochastic structures . The strength of such materials depends on the strength of each component, their mutual interaction and flaws. The flaws may be point defects in each component and/or extended dislocation type defects, which are crossing several composite components. The latter are more dangerous as they carry the most of the stress energy. Nevertheless all these flaws are the sources of AE. In this case AE presents a certain part of the general failure energy and may be looked upon as a "mirror" of the material failure .
AE and failure prediction procedure
Concrete is a typical example of a composite with stochastic structure and can be presented as a matrix-filler composite, where filler is distributed randomly. Matrix is presented as a polycrystalline solid containing micro pores and cracks and due to technological particulars is inhomogeneous mechanical medium. The AE sources in concrete may be generated by different components of the material. In spite of the fact that the sources of the AE signals has different character it is very difficult to tell one of them from another because they all are generated at the same time . That is why the AE integral characteristics are used for analysis of failure. It is possible to name some of them here:
Relative level of stress is based on dynamic identification model, where the input action is considered as trial load for initiation of AE and output action is considered as integral characteristic from a local zone of concrete ;
NN-criterion is based on application of multilevel trial loads and knowledge of the material strength. It is used for failure prediction of concrete and based on the functional . Special approaches could be used for failure estimation of a wide class of composite .
Typical scheme of the trial loads application for the AE testing is given in the Figure 3.
Fig 3: a) the AE intensity response to mechanical compression trial loads in time domain ; b) trial load: t0 - delay time, tfr - head front time, tpuls - pulse time, tstab - stabilization time, tbk - back front time; c) compression test set-up: 1- concrete cylinder specimen, 2- the AE transducer, 3- the AE registration equipment.
Some AE effects in Cement-based materials
Some important AE effects were discovered in the cement-based materials. They could be used for estimation of the main characteristics of these materials, such as strength, cracking resistance, and for failure prediction.
The AE manifestations have been registered in concrete at early ages and immediately after the concrete mix. This important effect permitted to make a rather accurate forecast of concrete strength for the mature material .
Kaiser And Felicity Effects
The Kaiser and Felicity effects have different manifestations in concrete comparing with other materials. They do not exist in clear form and their interaction may give a possibility for developing a new criterion for crack resistance characterization of concrete .
The "Thermo-Acousto-Emission" Effect for Stress and Strength Estimation  may be consider as prospective in-situ method for estimation of concrete propertoes. The TAE effect is a spontaneous AE radiation as a result of propagation of the thermal wave front through material. The thermal waves may be generated by creation of different temperature potentials in various parts of the structure material. The temperature potentials may be created using the following techniques:
- Temperature heating of a local zone (the thermal load technique)
- Temperature cooling of a local zone (the cryogenic load technique)
- Simultaneous heating and cooling of two neighboring zones (thermo-cryogenic load technique)
The TAE effect is characteristic of only composite materials due to the fact that different components of the material have different stress-strain responses to temperature impact. The proposed techniques are based on the physical processes of the thermal (cryogenic) wave propagation through materials. They make a shift in the thermo-dynamic state of the material, create additional concentration of stress on the composite bonds and joints in micro and macro cracks, pores, and other defects, that, in turn, calls the TAE effect.
Experiments with the TAE for Strength Estimation
The TAE method was tried on different composite materials. The presented data were obtained for various types of concrete, which could be considered as multi component composite material. Experiments for study of the TAE effect were carried out using the thermal load technique The experiments took place on concrete specimens imitating column fragments prepared from concrete of various compositions. The specimens had strength from 13 up to 42 MPa. The specimens were lightly reinforced prisms with cross section of 25x25 cm and height 100 cm. Three groups of specimen with different composition were prepared. Three types of filling were used: ceramics, limestone, and gravel.
The equipment setup included a thermo-jet (2), controller (3), recorder (4), acoustic emission equipment (5), and AE transducers (6). The thermo-jet allowed to create local heating using electric energy source. The controller helped to support fixed level of temperature. The data were recorder by data acquisition digital block. Temperature of the specimen surface was also registered during the experiments. The acoustic emission transducer was placed on a fixed place near heating-cooling sources (Fig. 4)
Fig 4: Equipment setup: thermo-jet (2), controller (3), recorder (4), acoustic emission equipment (5), and AE transducers (6).
The special compressor with liquid nitrogen was used for creating of temperature cooling of a local zone. The thermo-jet and compressor were used together for combination of temperature heating and cooling.
The Acoustic Emission equipment
The six-channel acoustic emission equipment "SPRUT-2M" (Russia,1991) was used for the acoustic emission measurements. This portable equipment had possibilities to measure quantitative parameters of the acoustic emission process. It also allowed to calculate source location in 3-D medium.
Fig 5: The six-channel acoustic emission equipment "SPRUT-2M"
Modeling of Heating and Cooling Propagation
Analysis of heating and cooling propagation in the specimen was carried out for estimation of the active zone, where the TAE effect is created. Heisler analysis was used for modeling. The MathCAD 6.0 PLUS software was applied for temperature cooling, and combination of heating and cooling analysis. Heating was carried out during 5, 10, and 15 minutes. In Figure 6 calculation results of thermal waves propagation in concrete slab are shown: heating (first line), heating and cooling (second line), and cooling (third line).
Figure 6 (below). Calculation results of thermal waves propagation in concrete slab are shown: heating (first line), heating and cooling (second line), and cooling (third line).
Cooling Of The Specimen
Some examples of cooling propagation modeling are given in the Figure 6. Cooling was carried out during 5, 10, and 15 minutes.
Combination of Heating and Cooling of the Specimen
Some examples of combination of heating and cooling propagation modeling are given in the Figure 6. Heating and Cooling were carried out during 5, 10, and 15 minutes.
The TAE Heating Technique:
Heating was carried out on the surface of the specimens in the range of 380 up to 510 during time 15-30 min. The acoustic emission pulses of various intensity were registered during the time of the thermal wave propagation into the testing material. Mean of the acoustic emission intensity (average acoustic emission data for each minute) was used as the testing parameter.
The TAE Cooling Technique:
Cooling was created by compressor with liquid nitrogen with approximate temperature 196 degrees C below zero.
Strength Estimation by Heating:
Estimation of strength of local zones of composite structures may be one of the TAE applications. An attempt to compare regression curve of strength versus TAE was made for each group of specimens and for all the specimens together. The latter is shown on the graph. Average of mean of acoustic emission intensity during a fix period of heating (15 minutes) was used for building up the regression curve. Correlation ratio reached 0.85 for all groups of specimens (Figure 7).
AE intensity during the effective period of the AE registration for three groups of specimens: o - 20 MPa, x - 30 MPa, + - 40 MPa.
The specimen loading for creation of different levels of stress:
regression relation of the AE intensity to the relative levels of stress 0 / R .
The TAE for combination of heating and cooling for hard:
The distance between heating and cooling zones was about 20 centimeters. It is possible to conclude from the graph (Figure 8) that combination of cooling and heating gives us more effective representation of the TAE application.
Stress Estimation by the TAE Method:
Some experiments were carried out for estimation of the relative level of compression stress, which is presented by relation of existing stress to ultimate stress level. For this purpose similar specimens were tested by compression loads and the TAE method was added for estimation of the relative level of stress. For hard concrete specimens it is possible to get a very high linear regression. Correlation coefficient was in the range of 0.9 - 0.97. The experiments demonstrated high sensitivity of the TAE method to estimation of stress in real structures (Figure 8).
The "Silence" Effect
The "Silence" effect has a special significance for prediction of failure in concrete using AE. It can be described as a sharp decrease of the AE intensity before failure or fracture of the cement-based materials .
Discussion and Conclusions
Application of AE for estimation of failure of composites represents a complex problem and rise many questions that require solutions. Creation of the basic conception of composite failure is one of these problems. Without this conception it is impossible to explain physical processes that cause AE and, thus, to use AE for testing of composites. Due to great varieties of types of composite materials it is a tough task to find a general model for description of their failure. In this case only partial solutions could be used.
The trial loads method could be used for practical tasks of the AE testing of composites. This method could propose a unique testing procedure to different types of composites. In this case the material is treated as a "black box" dynamic model, the trial load may be considered as an "input" action, AE as an "output" action, and the material itself could be presented as a "transfer function". This approach allows building a flexible model and procedure for description of failure of different types of composites. On the other hand, the AE testing procedure could be unified and standardized for various materials.
Proposal to go "back to basics" is timely for the AE estimation of cracking as criteria worked out for homogeneous materials, for example metals, cannot be applied directly to composites. Simplified model of a composite represented as the Carroll-Holt mechanical model as an example of a basic model. This model explains pore collapses and the AE generation as a result of the process. Despite of considerable difference between the model and the actual material this model is explaining the main effects that could be used for the AE testing. The time parameter is a very important characteristic as for trial loads as for the material response to it.
The Carroll-Holt model is not the only one possible step "back to basics". There are other avenues leading in this direction. But this model allowed to synthesize some heuristic approaches to failures analysis using AE, such as "relative level of stress" and "NN-criterion", which give sufficient results for practical applications.
/DB:Article /SO:NDTISS /AU:Nesvijski_E /AU:Sarkis_P /CN:BR /CT:AE /CT:composite /ED:2000-03
- Green, A.T. Stress Wave Emission and Fracture of Pre-Stressed Concrete Reactor Vessel Materials. Second Inter-American Conference on Materials Technology, ASME, vol. 1, August 1970.
- Dunegan, H.L. and Harris, D.0. Acoustic Emission - A New Nondestructive Testing Tool, Ultrasonics, vol. 7, No. 3, 1969.
- Morton, H.L., Harrington, R.M. and Bjeletich, J.G. Acoustic Emission of Fatigue Crack Growth, Engineering Fracture Mechanics, vol. 5, No. 3, 1973.
- Nogin, S.I., Nesvijski, E.G. The Parametrical Points of the Cracking Process in Concrete under Compression. Journal of Concrete and Reinforced Concrete, No.3,1980, Moscow, Russia.
- Landis, E.N. and Shah, S.P. Recovery of Microcrack Parameters in Mortar Using Quantitative Acoustic Emission. Journal of Nondestractive Evalaluation, 12 (1993), no. 4.
- Shah, S., Ouyang, C., Marikute, S., Yang, W., and Becq-Giradon, E.,(1998). A Method to Predict Shrinkage Cracking of Concrete. ACI Journal, 1998, vol.95, No. 4.
- Carroll M.M., Holt A C. Static and Dynamic Pore-Collapse Relation for Ductile Porous Material. J. Appl. Phis., 1972, vol. 43.
- Nesterenko, V. F. High-Rate Deformation of Heterogeneous Materials . Novosibirsk: Nauka, Russia, 1992.
- Failure Mechanics and Strength of Materials. Editor: Panasuk, V.V. vol. 1, 2, Kiev, Naukova Dumka, 1988.
- Failure Mechanics and Strength of Materials. Editor: Panasuk, V.V. vol. 3, 4, Kiev, Naukova Dumka, 1988.
- Suzuki, H., Takemoto, M. and Ono, K., The Fracture Dynamics in a Dissipative Glass-Fiber/Epoxy Model Composite with AE Source Simulation Analysis. Journal Acoustic Emission, vol. 14, No. 1, 1996.
- Wenger, M.P., Blanas, P., Shuford, R.J., and Das-Gupta, D.K. Acoustic emission signal detection by ceramic/polymer composite piezoelectrets embedded in glass-epoxy laminates. Polymer Engineering Science, vol. 36, no. 24, (December 1996).
- Daniel, I.M. and Hsiao, H. M. Acoustic Emission Monitoring of Fatigue Damage in Metals. Nondestructive Testing and Evalaulation.,1998, vol. 14.
- Prosress in Acoustic Emission IX. Transition in AE for the 21st Century. Proceeding of 14th International Acoustic Emission Conferebce. Big Island, Hawaii, USA August 9-14, 1998.
- Acoustic Emission: Standards and Technology Update. ASTM, STP 1353, West Conshohocken, PA, USA, 1999.
- Ono, K., Huang, Q. and Wu, J-Y. AE/Pattern Recognition. Analysis of the Fracture of Glass Fiber Composites Exposed to Hot-Wet Conditions. Proceedings of the International Symposium on Acoustic Emission from Composite Materials, vol. 5, pp. 197-204, July 1995.
- Nesvijski, E.G. On Application of Acoustic Emission Method for Prognosis of Hard Concrete Properties. Proceedings, Fifth International Symposium on Acoustic Emission of Composites Materials (AECM-5), Sunsvall, Sweden, 1995.
- Nesvijski, E.G. and Nesvijski, T.E. Kaiser and Felicity Effect and Their Application for Evaluation of Concrete by Acoustic Emission. ASNT's Fall Conference, Seattle,WA,USA,1996.
- Nesvijski, E.G. The "Thermo-Acousto-Emission" Effect for Stress and Strength Estimation of Composites. Proceedings: ASNT Spring Conference and Research Symposium, Orlando, Florida, USA, March, 22-26, 1999.
- Nesvijski, E.G. and Nesvijski, T.E. Failure Forecast and the Acoustic Emission "Silence Effect" in Concrete. ASNT's Spring Conference, Houston, TX, 1997.