A COMPREHENSIVE REVIEW OF ACOUSTO ULTRASONIC-ECHO (AU-E) TECHNIQUE FOR

FURNACE REFRACTORY LINING ASSESSMENT

*A. Sadri1, W. L. Ying1, P. Gebski1, P. Szyplinski1, T. Goff2 and B. van Beek2

1Hatch Ltd.

2800 Speakman Drive

Mississauga, Canada L5K 2R7

(Corresponding author: asadri@hatch.ca)

2Lonmin Platinum

Private Bag X508

Marikana 0284

South Africa

ABSTRACT

Monitoring the refractory lining integrity is essential to assure the safety of the operators and to determine

the remaining campaign life of a furnace. Early detection of refractory lining degradation is essential to avoid sudden

furnace failures and prevent unscheduled shutdowns. In mid-1990s a stress wave propagation technique was

developed and since then is has been utilized by Hatch Non-Destructive Testing (NDT) Group to evaluate furnace

linings. For the last 17 years the AU-E technique evolved in terms of hardware, software and signal acquisition and

interpretation. Most importantly, based on more than a decade of commercial applications, an extensive database of

various failure modes, refractory deterioration trends and material properties has been created. This knowledge base

is an integral part of AU-E which progressed from a NDT technique towards a comprehensive approach for furnace

condition monitoring. All of the improvements introduced over the last several years have greatly improved the

accuracy of the measurements. The data analysis process has also become more time efficient making the critical

results available to the furnace operators in a more timely manner. Today the AU-E technology is used on a variety

of process vessels in non-ferrous, glass, chemical and iron and steel industries.

In this paper we are reviewing the available techniques for furnace refractory lining measurement and

discussing the pros and cons of each method. The need for continuous improvement and development of new

techniques for refractory monitoring is also emphasized. Finally, we are presenting a number of case studies and

comparisons between the physical verifications and the AU-E measurements.

KEYWORDS

Non-Destructive Testing (NDT), process vessels, integrity monitoring, maintenance, refractory, acoustic,

ultrasonic, Acousto Ultrasonic-Echo (AU-E), impact-echo, furnace, smelters

INTRODUCTION

Refractory linings in metallurgical furnaces undergo deterioration and wear over time. The deterioration is

caused mainly by mechanical and thermal stress cracks and chemical attack resulting in loss of heat transfer capability,

shrinkage and change of thermal conductivity for the refractories. Deterioration starts immediately after the start up

and progress continuously, however, we have examples of spalling refractory that appears to break abruptly in chunks.

The degree and mechanism of deterioration depends on many different factors over the huge range of metallurgical

furnaces but certainly varies by zones within the same vessel. Failure of the lining is dangerous and could affect the

structural integrity of the furnace. Rebuilding after a sudden failure is expensive in terms of lost production, man

hours, and materials.

Non-destructive Testing (NDT) and monitoring of the refractory lining leads to better safety, longer campaign

life, production optimization, scheduled maintenance and reduced costs of production. Accurate thickness

measurement and monitoring of the refractory lining in operating furnaces is possible using acousto ultrasonic-echo

(AU-E) technique. This technique uses stress waves of both acoustic and ultrasonic ranges in order to penetrate into

various layers of refractory lining. The lining thickness and integrity can be determined by the patterns of reflected

waves.

The AU-E technology was invented in 1998 to help a client to monitor the formation of excessive build-up

in their flash furnace hearth. For the last 17 years we have inspected and monitored a large t number of furnaces and

process vessels around the world. The AU-E technology has gone through an impressive evolution and transformation

since it was innovated mainly because of the need for more accurate, repeatable, rapid and reliable results to assist in

better interpretation and faster delivery of the results.

REFRACTORY

Refractory materials are designed to be physically and chemically stable at high temperatures (above 500oC)

[1] and are used in linings for furnaces, kilns, ladles, reactors, converters and other types of high temperature vessels.

Depending on the operating environment, refractories need to be resistant to thermal shock, be chemically inert, and/or

have specific ranges of thermal conductivity and/or coefficient of thermal expansion. The particular process dictates

the specific design of the refractories. For example, blast furnaces hearths are lined with carbon and graphite

refractories because of their high heat conductivity but the stack is lined with silica carbide based refractories to resist

the abrasion due to movement of coke and pellets in that area. Refractory can corrode or wear thin, it can be

impregnated by slag/metal, or it can change chemically in other ways.

In Figure 1 we are demonstrating six different stages of refractory wear in a furnace. Certainly, there are

many other mechanisms that cause refractory deterioration, such as hydration, oxidation and carbonization which

involves reaction of refractory with unwanted elements such as water and oxygen, but the six stages of wear in Figure

1 are universal throughout the smelting industry. In an operating furnace, refractories are non-uniform and contain

metal elements, cracks and chemical alterations. A reliable refractory measuring and monitoring technique should be

able to detect and distinguish between different stages of metal impregnation, penetration, joints, cracks and build-

ups.

EXISTING REFRACTORY THICKNESS MEASURMENT METHODS IN OPERATING PROCESS

VESSELS AND FURNACES

Refractory Thickness Estimates Based on Thermo-modelling

To determine the remaining refractory thickness and monitor deterioration and lining wear in a furnace

numerous mathematical models have been developed. The specialized models have been developed for each industry,

such as blast furnaces and pyrometallurgical furnaces, and the operational components, such as “elephant foot” and

“mushroom effects” in blast furnaces and “electrode positions” and “freeze line” for electric arc furnaces [2 and 3]. In

modern furnaces, the walls and the hearth have embedded thermocouples to monitor the temperatures of the furnace

walls. By using the temperature readings of the thermocouple the heat fluxes in the refractory lining can be calculated.

These heat flux calculations are used in the computational model(s), based on heat flow and energy conservation to

determine the remaining refractory and build-up thicknesses. There are various mathematical model approaches and

numerous types of thermocouple that are used to make refractory wear calculations. The models are as good as the

accuracy of the assumptions and coefficients. Also the models heavily relay on the quantity of the installed

thermocouples and their distribution in the furnace structure.

Figure 1 - Different stages of refractory deterioration in an operating furnace

Refractory thickness monitoring by use of thermocouples and mathematical and thermal models is essential

and part of the operator’s routine task. However, due to limitations in the effective range of the thermocouples and

multivariable assumptions for the mathematical models, it is important to use other means of “non-intrusive”

techniques to determine the refractory wear profile in operating furnaces. In addition, refractory wear monitoring by

mathematical modeling techniques are not able to distinguish between various defects that are demonstrated in Figure

1. The following sections briefly describe the more direct, nondestructive and non-intrusive refractory measuring and

monitoring techniques.

Radioactive Tracers

Radioactive tracers have been used for determining refractory thickness in converters, glass furnaces and

blast furnaces since the 1960s. All of the published literature is about the use of tracers in blast furnaces [4, 5, 6, 7,

and 8] and there is no record of using radioactive tracers in other types of process vessels. In this method radioactive

isotope tracers with short half-life are introduced to the blast furnace at the top of the furnace. As the tracer moves

downward with the burden, the receiver or isotope’s tracer counter counts the radiation that passes through the build-

up, refractory and the shell. In an ideal theoretical situation, where quality of the refractory is the same throughout

the furnace and shell thickness is uniform, the increase in radiation count indicates a thinner and more worn refractory

lining. However, blast furnace linings are far from uniform and several types of refractories with different densities

and material properties are utilized in the lining. In addition the presence of staves and plates for cooling purposes

adds to the complexity of the lining. Therefore the fluctuation in radiation counts do not necessarily indicate changes

in refractory thickness. The presence and thickness of accretion, impregnation of molten metal/slag in the refractory

and quality variation of refractory have an effect on the radiation counts.

A literature review on the use of radioactive tracers shows the use of 40La, 192Ir and 60Co isotope for lining

thickness measurements 4, 5, 6, 7, and 8]. In addition to radiation exposure another issue with radioactive tracers is

their highly toxic nature that may cause health hazards to the operators.

Infrared (IR) Thermography

The infrared (IR) thermography is the most common NDT technique used for determining hotspots and

troubled areas. In the recent years a large variety of systems have been available commercially with a broad range of

image resolutions and prices. The IR camera provides a thermal image showing the temperature of the object on a

color scale where the darker colors are cooler and the lighter areas are hot spots. Temperature differences on the

surface, may be related to the subsurface defects, which can be located from the thermal images.

Furnaces and converters have a steel shell which oxidizes over time. The steel shell rusting will affect the

surface emissivity of the vessel which results in inaccuracies in the IR thermography measurements [8]. In addition,

gaps, cracks and looseness of the refractory bonding will affect the results. Nearby heat sources and reflections from

nearby objects also affect the IR thermography data [10].

For the above mentioned reasons IR thermography is mainly used for rough estimates and the thermal images

must be carefully collected under the same conditions for comparison purposes. In 2008, Hatch has developed a

methodology to use IR thermography to measure the refractory thickness in single layered hollow cylindrical vessels

such as converters and reactors [11]. This method uses a thermal model of the converter and the thermographic image

of the converter’s outer surface to calculate the refractory thicknesses at the inspected spots.

Electromagnetic (EM) / Microwave Techniques

Microwave and radar systems have successfully been used for determining refractory thickness prior to their

installation. The change in dielectric constant or the change in ability of the medium to transfer electricity causes the

EM waves to reflect back from an interface [12]. This property enables the EM waves to be used for measuring

refractory lining thicknesses. However, most of the process vessels have external steel shells and EM waves cannot

pass through the steel shell and into the lining. In isolated cases where part of the furnace lining is exposed and

antennas can be placed directly against the refractory wall or hearth the data can be collected. Never the less, presence

of support and grillage beams or any metallic object in close vicinity of the radar antenna will cause signal phase

distraction and which may result in false signal interpretations. If metal penetration between brick layers is present

then radar will not be able to measure beyond this interface.

The commercial short-pulse radars use the 500 MHz- 1GHz range of the FM spectrum. The useful frequency

range for measuring refractory lining is between 500 MHz to 1 GHz hence the commercially available radar systems

are within the desired frequencies. The frequency is critical when dealing with different type of materials and

structures. Ground Probing Radar (GPR) is commonly used in non-destructive testing of concrete and masonry

structures. For these standard applications it produces multiple results, including thickness of the structure, location

of voids or inclusions, and location of rebar in reinforced structures.

Hatch NDT Group has utilized the GPR to measure refractory thickness in electric furnaces. We used high

frequency radars to be able to measure the brick wall and roof thickness in several furnaces. The measurements were

done from outside where there was no shell and the refractory lining was exposed. Although this approach was

successful, it requires special care as it cannot be applied at any furnace.

In addition to impulse radar systems such as GPR, Frequency Modulated Continuous-Wave Radar (FM-CW),

which uses a triangular waveform to modulate the frequency of an oscillator, was also used in laboratory condition

for detection of refractory wear in glass furnaces with limited success [13].

The EM signals are not very sensitive to refractory/air interfaces therefore voids and gaps cannot be detected

in the lining. The air gaps between the refractories cause signal attenuation and reduce the signal-to-noise ratio. The

refractory/metal interface will cause EM waveform reflections, hence metal impregnation into the brick, metal

penetration in between the lining courses and presence of accretion or build-up will affect the accuracies. In addition,

the dielectric constant of the refractories changes at high temperature which requires a correction factor if EM

techniques are going to be used for refractory thickness measurements.

In metal processing vessels, the GPR and microwave systems cannot be used for postmortem study of the

furnace lining condition because of heavy presence of metals inside the furnace and within the matrix of the refractory

lining, which causes signal attenuation and distortion.

Ultrasonic

Low frequency “Ultrasonic Pulse Velocity” (UPV) systems such as PUNDIT and V-Meter have been used

to determine refractory quality since Mid-80s. Using the UPV systems, a refractory with known thickness is placed

between a transmitting and receiving transducer which are at opposite ends of the refractory sample. A low frequency

“Ultrasonic Through Transmission” (UTT) system with frequencies between 20 to 200 kHz is used to determine time

of flight between the transmitter and receiver. Generally, high ultrasonic wave speed indicate low porosity, high

density and high modulus of elasticity. “Ultrasonic Pulse-Echo” (UPE) systems are required to measure refractory

thickness on an operating furnace. In UPE systems the transmitting and receiving transducers are placed on the same

surface. The distance between the transmitting and receiving transducers should be minimal. Parker et al. [14]

demonstrated that the refractory/molten metal interface creates a distinguished reflection surface because the acoustic

impedance in liquid is typically 13-15% less than for a solid refractory or accretion which means that at least 10% of

the waves are reflecting from the wall. In their paper they demonstrated that a large amount of ultrasonic energy

attenuates within the solid refractory because of temperature and porosity. A decrease in the amplitude of reflected

ultrasonic signals was noted as the sample was heated to higher temperatures. Temperature also affects the travel path

of an ultrasonic waveform in solid [15]. As the temperature increases the longitudinal or P-wave path widens which

results in higher attenuation and a change in the reflection angle.

For measuring refractory thickness wear in furnaces conventional ultrasound is not an option. A conventional

ultrasonic system uses single frequency transducers with fixed and narrow bandwidths. An ultrasonic system with a

“sweep frequency” pulse generator and broad band transducers was developed by Hatch NDT Group to determine

stave thicknesses in blast furnaces. This Low Frequency Pulse Ultrasonic (LFPU) is a pulse-echo system that is

capable of measuring stave thicknesses with a 2 mm precession. The multi-frequency ultrasonic pulses can travel into

different layers of blast furnace wall and measure accretion and refractory at the tip of the stave. Even though the

LFPU was proven to be accurate for blast furnace stave thickness measurements [16], the system has not been used

on thicker furnace refractories. The authors believe that the LFPU could be successfully implemented for refractory

wall thickness measurements of less than 300 mm thick. In the near future LFPU will be tried on furnace sidewalls

and the results will be published by Hatch NDT Group.

In mid-1990s a group of researchers from Magnitogorsk State Technical University in Russia made attempts

to utilize the “ultrasonic tomography” technique to identify blast furnace hearth lining layers. Ultrasonic tomography

is widely used in medical profession and refers to imaging by sections or sectioning, through the use of penetrating

ultrasonic wave. In their approach Magnitogorsk which formed a Techno-Consulting in Mid-2000s, is using

“Ultrasonic Mirror-Shadow Method”, to transmit and receive ultrasonic waves across sections of a blast furnace (see

Figure 1). In these measurements ultrasonic sensors above 20 kHz are used to transmit and receive the signals. The

receiving transducers are placed at 0, 20, 40, 60 degrees angles from the transmitting sensor across the blast furnace.

The measurements are repeated in three to four different locations around the hearth. Based on wave speed differences

due to material changes for each layer, the captured waves are used to create an “image-shadow”. The so called

“image shadow” is developed based on mathematical modeling and correlations between the arriving ultrasonic waves

with blast furnace layers [17] (Figure 2).

Figure 2 - Left; claimed transmission and receiving of the ultrasonic signals across the blast furnace, right; software

for result generation [17]

Based on the published literature [17], the ultrasonic signals are going through the shell, ramming, thick

refractory lining, molten metal, hot and porous “Deadman” in the center of the hearth, again the molten metal, thick

refractory lining, ramming and finally shell then are detected and collected on the opposite side (Figure 1, a central

pathway). The high frequency of ultrasonic wave (above 20 kHz), high rate of signal attenuation because of

temperature and material porosity, complex signal pathway from high to low and low to high acoustic impedances

makes are part of the challenges involved in conducting ultrasonic tomography on an operating furnace.

The only accurate method of measuring refractory thickness is the “direct” measurement, while the furnace

is out of commission and accessible. , such as using measuring tapes, drilled core measurements and use of lasers,

which is not possible while the furnace is in operation. All other techniques are using “indirect” approaches and their

capabilities and limitations must be taken into consideration once the data is used for evaluations and comparisons.

The review of the above mentioned traditional and new techniques for measuring and monitoring refractory lining in

operating furnaces and process vessels lead us to believe that there were no reliable, accurate, repeatable, and non-

intrusive measurement technique available and we need to develop a practical system to fulfill the above requirements.

Therefore we developed the Acousto Ultrasonic-Echo (AU-E) technique.

ACOUSTO ULTRASONIC-ECHO (AU-E)

The Acousto Ultrasonic-Echo (AU-E) technique was developed in late 1990s based on the same principles

governing the impact-echo (IE) technique [18, 19, and 20] with additional modifications, to account for the effects of

extreme temperature on the wave propagation, multi-layer lining with a different acoustic impedance for each layer

and furnace shape and dimensional effects. AU-E is a patented technology that has been commercially used for

refractory thickness and quality evaluation for more than sixteen years.

AU-E is a stress wave propagation technique that uses time and frequency data analysis to determine

refractory thickness, and detect anomalies such as cracks, gaps or metal penetration within the refractory linings.

During the measurement, a mechanical impact on the surface of the structure (via a hammer or a mechanical impactor)

generates a stress pulse, which propagates into the refractory layers. The wave is partially reflected by the change in

material properties of each layers, but it also propagates through the solid refractory layers all the way up to a

brick/brick or brick/gas or brick/molten metal interface.

The compressive waves (or the P-waves) are received by a sensor / receiver and the signals are analyzed for

quality and thickness assessment of the refractory. In the AU-E technique instead of the actual P-wave speed the

apparent wave speed is used in the calculations which is an average wave speed in a three dimensional geometric

space. The speed of the P-wave in each layer is affected by the density, thermal gradients, shape factor and elastic

properties of brick.

The AU-E technique uses correction factors to account for the speed in each layer. The thermal correction

factor, α is calculated based on refractory dynamic Young's modulus of elasticity under service conditions compared

to the dynamic Young’s modulus at room temperature (Equation 1 and Figure 3). If the dynamic modulus of elasticity

over a temperature gradient is given as Ed (T), the temperature correction factor, α can be calculated as:

Figure 3 – Effect of increase in temperature on dynamic Young’s modulus for refractories

o

T

T

d

E

dTTE

2

1

 

1

 

(1)

Where Eo is the dynamic modulus of elasticity at room temperature and in this calculation we are assuming

that the Young’s modulus of elasticity of the refractory changes within a uniform gradient between the hot and cold

face as a result of temperature (see Figure 3).

The shape factor β, accounts for reduction in wave velocity due to geometry of the structures the waves are

travelling in. The apparent wave speed offset is due to excitation of structure’s natural frequencies by the impact

force. The shape factor value depends on the cross section dimensions of the testing area. For example the β factor for

the cross section of a column is 0.87 [18]. For most furnaces where lateral dimensions are at least six times the the

thickness of the lining, β is 0.96.

The governing equation for the AU-E technique is indicated by Equation 2.

T = (αβVp) / (2 fp) (2)

Where fp is the P-wave frequency, Vp is the propagation speed of P-wave in the material, T is the thickness

or depth of the reflecting surface, α is the temperature correction factor and β is the shape factor.

For a multilayered section such as furnace hearth, the thickness of the final refractory layer (Tn) is calculated

based on the following equation:

1

1

21

2

n

i iiiP

innnp

nV

T

f

VT

(3)

Where ƒ is the resonance frequency for the thickness of the nth layer. Equation 3 can be used to determine

the refractory lining thickness on the hot face, considering the P-wave speed (Vp)i, the thermal correction αi, the shape

correction factor βi and the thickness Ti of the layers prior to the inner most layer are known. Equation 3 assumes that

the stress wave generated by a controlled impact source contains sufficient energy to reach the inner most layer of the

lining and resonates back and forth between the two faces to create a desirable or defined P-wave thickness frequency.

In addition to understanding the mechanisum of the stress wave measurments, an important factor for a

successful AU-E inspection is the utilization of the right tools to implement the inspection procedures. A broadband

vertical displacement transducers of desirable frequency range was designed with the ability to function at high

temperatures and in wet environments. For pulse generation, impactors with specific spherical tip diameter, capable

of generating a specific range of frequencies were selected. The data acquisition system is military grade, water and

dust resistant sturdy system which can stand high and low temperatures between -50°C to 90°C.

Calibration

Prior to the field data collection the apparent P-wave speed for each brick type is determined representative

brick samples at room temperature. The wave speed measurements must be carried out on all the materials that the

wave is passing through. The α factor can be calculated experimentally, by heating brick samples and measuring the

wave speeds at the desired temperatures or it can be calculated by the brick elastic and thermal properties. For β

factors the value can be calculated once the dimensions of the testing area are available. After the calibration is done

a mathematical model is created to help the specialist about their field data collection hardware and software settings.

Thickness Measurements and Refractory Wear

The Data collected in the time domain and it is extremely complex, containing numerous frequencies and

multiple reflections, diffractions, refractions from body and surface waves (see Figure 4).

Figure 4 - Typical time domain signal from a furnace wall

In the frequency domain the results are better defined, but still there are many different elements that can

cause misinterpretation. Figure 5 shows three typical resulting frequency spectrum including a signal collected on a

furnace wall with no cracks or impregnation; a small and filled crack across a brick, and an opened crack across a

single brick in a furnace wall, from top to bottom perpendicularly. Note that lower frequency reflections are associated

with greater distances to the interface reflecting the signal

Refractory wear is usually thought of as the reduction of refractory thickness over time caused by the thermal

and mechanical stresses in the furnace. When using the AU-E technique there are a few other refractory conditions

that will appear as detected remaining refractory thickness; namely, metal impregnated refractory and accretion or

build-up (see Figure 6).

Figure 5 - Sidewall refractory top to bottom: without impregnation or build-up; refractory thickness with small

closed crack; refractory thickness with large opened crack

When refractory is impregnated by metal (see Figure 1 Stage 3), the impregnated portion has a significant

reduction in elastic property compare to good refractory. AU-E signals will be reflected from the impregnation

boundary, which can be mistaken for the remaining refractory thickness (see Figure 7).

Figure 6 - Sidewall refractory lining top to bottom: echoes from good refractory, good refractory and impregnated

refractory; good refractory, impregnated refractory and build-up

Figure 7 - Hearth refractory lining, top: without refractory impregnation, bottom; with metal impregnation

When accretion or build-up is formed on the hearth, the build-up/molten material interface can be

misinterpreted as the remaining refractory thickness, while the build-up/refractory interface can be misinterpreted as

a crack. Proper interpretation of AU-E signals can be aided by a good understanding of the furnace process and

operating conditions. Like other tasks that requires judgment, this takes the experience of an AU-E specialist.

Detection of Anomalies

Here anomaly is a clear signal reflection from within the refractory lining but we are not certain of the source.

Anomalies could be cracks, voids, metal penetration, hydration or a combination of these features. When there are

discontinuities, cracks, or interfaces of different materials (see Figure 1), the signals tend to have multiple reflections

and at higher frequencies compared to the low frequency reflection for the full refractory thickness. The signal

reflection frequencies are used to determine the position of the anomaly.

When significant gaps/cracks are present in the bricks, impact signals may not transmit through the entire

thickness of the brick. They will be attenuated by the gaps and thus, signals seem to reflect from a thinner region

closer to the cold face. As such, the actual remaining refractory could be thicker than detected by the signals.

Detection of Refractory Chemical Changes Such as Hydration, Carbonization, and Oxidation

The speed of the P-wave is dramatically lower in hydrated/carbonized or oxidized refractory as compared to

non-affected refractory; therefore, greater than expected refractory thickness is measured in areas of

hydrated/carbonized or oxidized refractory. However it is important to understand that hydration, carbonization or

oxidation happens as a result of a chemical reaction between the refractory and oxygen or water at certain

temperatures. The severity of the chemical changes to the bricks depends on the volume of the external agent, water

in the case of hydration, time of exposure, temperature and how favorable the conditions are for the chemical change.

The effect of the P-wave signals depends on the severity of material changes within the lining. If the area of the

chemical change is small in relation to the length and geometry of the lining, the AU-E signals may not be affected so

the change in refractory quality may not be detected. On the other hand, if the chemical effects are extensive in

geometry and material properties in relation to the lining, AU-E signals will readily detect the defected area.

Unfortunately at this time we don’t have enough field data to readily identify the different stages of chemically

defected refractory using AU-E signals. An illustration of the effect of hydration on the reported refractory thickness

is shown in Figure 8.

Figure 8 - Illustration of the Effect of Hydration

Metal Penetration

Another possible explanation for “thicker than normal” lining sections is the presence of metal penetration

into the refractory. AU-E cannot identify metal penetration within the hearth lining layers when the penetration is

“smaller than signals half wavelength” which is less than 50 mm. When metal penetration is too small for AU-E to

detect as a separate interface the additional metal thickness will be reported as part of the lining resulting in a “thicker

than normal” reading (see Figure 9).

Figure 9 - Metal penetration

When using the AU-E technique in areas with both metal penetration and hydration/carbonization or

oxidation, it is possible that they could be mistaken for each other. Metal penetration is normally the cause of “thicker

than normal” readings near the center line of the furnace.

The Accuracy of AU-E Measurements

In a previous article we have discussed the causes of AU-E errors in great details [29]. In general based on

numerous verifications, the accuracy of the AU-E measurements are between 4-7% of actual thicknesses or anomaly

positions. Accretion or build-up thickness measurements are within 15% of physical measurements due to the greater

uncertainty in the wave speed in accretion.

We have written and presented numerous papers on the various aspects and applications of AU-E inspections,

namely principals [21. 22, 23 and 24], case studies for blast furnaces [25, 26, 27 and 28], non-ferrous furnaces and

vessels [29, 30, 31 and 32], preventing and predicting maintenance [33 and 34].

Improvements in AU-E Results and Reflections on Case Studies

Since the beginning of AU-E invention in 1998 many different improvements have been introduced. Some

of them can be credited to the overall rapid development and increased affordability of computers, digitizers and

software development tools. As a result, the data processing speed increased and the results can be presented in a

customized way to clearly and quickly report on the issues detected in the lining. Another source of improvements in

terms of data interpretation has been achieved by the increase in our experience and knowledge about furnaces and in

part thanks to better cooperation with furnace operators and other technical personnel at smelters with our staff as the

value of AU-E became more apparent to operating companies. Obtaining feedback from the post mortem analysis of

the linings and analyzing it in conjunction with the AU-E results proved to add great value to the understanding of the

refractory deterioration process. All these improvements have been applied to the AU-E evaluation and their

importance is reflected in the following case studies.

CASE STUDY 1: ELECTRIC ARC FURNACE

Hatch had measured the refractory thickness of Furnace 1 at Lonmin operations in Marikana, South Africa,

for a long time while the furnace was in operating during their previous furnace campaign. The first AU-E survey was

done in August 2011, followed by four more rounds of measurements. Each of the five AU-E surveys included

thickness measurements at 103 locations around the furnaces, at three levels between the plate coolers. Each time the

remaining thicknesses were plotted for all the locations at individual levels. These results are shown in millimeters in

Figure 10. The design thickness of 650 mm is also shown for reference. The wall was expected to remain stable unless

it is worn back to 60%-70% of the depth of the plate coolers (this is indicated in the drawing below by the red dashed

line). The last survey was completed in March 2013, approximately one month before the furnace shutdown for a

rebuild. These latest NDT results including the brick thicknesses of the furnace wall computed based on the AU-E

measurements were compared to the actual thickness measured during the demolition.

In summary, the final measurements done shortly before the shutdown revealed the following thicknesses:

Level 1: average thickness 453mm, minimum thickness 410mm;

Level 2: average thickness 467mm, minimum thickness 440mm;

Level 3: average thickness 468mm, minimum thickness 440mm.

The above results were later compared to the actual thickness measurements taken at 23 locations (out of 103

location previously used for AU-E surveys) during the furnace demolition. The lining wear pattern was relatively

uniform around the furnace (see Figure 11) which was in agreement with the results of the AU-E survey. As a result

it was concluded that the sample size (i.e. the number of actual measurements) is representative for the entire furnace

and can be used for the purpose of validating the AU-E results.

Whenever the actual remaining brick size was measured it was compared against the AU-E result from March

2013. Such comparison is presented in Table 1, together with the difference obtained between AU-E and the actual

measurement. Overall, the accuracy of the NDT measurements proved to be very high, with the average discrepancy

between AU-E and actual size being 4%. Given the average thickness of the remaining lining equal to 480mm (at the

tested locations) that error translates to 20mm. Such a level of accuracy is sufficient to accept AU-E technique as a

highly reliable technique for lining monitoring. Only one place where the discrepancy between the AU-E result and

the actual thickness was high is location 3a. The AU-E indicated 420mm, while the actual thickness was 510mm. The

AU-E data were verified for this location and it still resulted in the same thickness of 420mm. The visual inspection

of this particular area revealed a void at this location at 400mm from the cold side, which could have been detected

by AU-E and mistakenly interpreted as a full thickness. It is however unclear whether this void was present during

the furnace operations or it appeared during the demolition as a result of removing the coolers and breaking down the

wall. Another possible explanation is that the false reading was caused by an internal inclusion in the particular brick.

Figure 10 - Thickness measurement results

Figure 11 - Wear pattern around the furnace wall

Table 1 - AU-E results vs. the actual brick size

Sample pictures taken at one of the locations where the AU-E results were validated against the actual brick

sizes are presented in Figure 12

Figure 12 - Manual brick thickness measurements (Location 5b – Level 1)

After validating the AU-E measurements Lonmin acknowledged the accuracy of this technique and the

proposed approach. With a high level of confidence the AU-E measurements are currently being implemented in

regular time intervals for two Lonmin furnaces. The refractory wear trend and rate, as shown in Figure 13 for the

previous campaign life, constitute an important piece of information allowing the furnace operators proper

maintenance scheduling and implementing process modifications based on the ongoing brick thickness results.

Figure 13 - Refractory wear trend (left) and rate (right)

CASE STUDY 2: BLAST FURNACE

The monitoring of a blast furnace has been continued since 2013. The campaign life of the blast furnace was

coming to an end and the client wants to safely extend the current campaign life. The first AU-E inspection was

performed in July 2013. During the inspection, the area of interest was divided into grids for measurements. The

resulting remaining refractory, estimated thickness of accretion and location of cracks or gaps were plotted in sections.

Figure 14 illustrates the typical partial cross-section of the blast furnace hearth. The brown line indicates the remaining

refractory profile detected in the previous AU-E survey, while the blue line indicates the results of the most recent

survey. Any anomalies such as cracks, gaps or openings within the remaining refractory were indicated as red squares.

The location of any potential cracks is important for designing mitigation measures or monitoring plans, because

cracks that coalesce and propagate may cause the spalling of refractory at the hot face, which will result in a sudden

reduction of refractory. Layers of accretion were also detected (indicated in green triangles in Figure 14). The same

measurement locations were inspected during each subsequent inspection so that the results can be compared and wear

trends can be determined (Figure 15).

Figure 14 - Typical profile of the remaining refractory, changes between two consecutive AU-E inspections, and the

extent of possible accretion and layering of accretion in the lower hearth of the blast furnace

At that time of the first inspection, the hearth refractory was already half worn (Figure 15). Strong AU-E

signals were reflected at the interface between the refractory bricks. This suggested loose ramming material or gaps

opened at the brick interface. Under unfavorable furnace conditions, the remaining refractory at the hot face of the

aligned loose ramming may spall off and cause a sudden and significant deterioration of the hearth lining. These areas

were continuous monitored by thermocouples to detect any thermal spikes due to spalling of refractory. In addition,

titania (TiO2) was recommended at 5 kg per ton of hot metal, to create a protective layer at the hot face of the lining.

Figure 15 - The trend of refractory deterioration of a blast furnace hearth

AU-E was adopted as part of the routine hearth maintenance program for this blast furnace. A follow-up

(second) inspection was performed four months after the first inspection. With reference to Figure 15, the results

indicated that the average remaining refractory thickness has further reduced from 50% to 48%. The third inspection

was carried out eight months after the second inspection. The remaining refractory thickness was significantly reduced

to 41%. This is likely related to the spalling of refractory at regions with loose brick interfaces. The furnace operation

and processes were reviewed and it was speculated that the furnace was operated at high temperature and production

rate which created unfavorable conditions for the formation of accretion. The recommendations were refined based

on the findings. The addition of titania was then targeted at specific regions and the titania charging rate was further

increased to 10-12 kg per ton of hot metal. To allow proper formation of accretion, the furnace production was

controlled at intervals of slow and normal runs, as the formation of protective accretion layer was expected to be

favorable during the slow run periods. The fourth inspection was performed seven months after the third inspection.

The average remaining refractory thickness has been stabilized at 41% (Figure 15), which indicated that the preventive

measures of additional titania and throttling their through put was effective.

Annual inspections are performed for furnaces at their early stage of campaign or with good refractory

condition; however, for furnaces which are approaching the end of their campaign lives, more frequent inspections

will be required. This can be quarterly or half-yearly inspections to provide an accurate deterioration trend, and to

monitor the effectiveness of the mitigation measures. The schematic diagram of the stages of refractory wear and

mitigation measures are illustrated in Figure 16.

Figure 16 - Schematic Diagram of the Stages of Refractory Wear and Corresponding Mitigation Measures

Monitored by AU-E Technique

DISCUSSIONS AND CONCLUSIONS

Smelting furnaces are the key equipment in a complex system for the production of metals. All the furnaces

are lined by a range of refractory bricks or castable to contain the molten metal and protect the integrity of the crucible.

One of the main reasons for termination of a furnace’s campaign life is refractory wear and loss of vessel’s capability

to operate safely. From mid 1960s, due to advances in technology and equipment developments various tools have

been developed to help furnace operators to determine the status of their refractory lining. In late 1990’s the only

reliable and non-destructive measuring technique based on stress wave propagation principals was developed. The

AU-E system, like all dynamic technologies has gone through extensive evolution and improvements. Besides the

extensive evolution of our software and hardware, huge improvements have been achieved on signal recognition,

interpretation and general correlations with furnace conditions. Our aim is to effectively use AU-E monitoring for

providing information on decision making for predicting and preventing maintenance, relining and campaign life

extensions. We believe that the AU-E has reached the maturity as a newly developed technique to be utilized as part

of the furnace campaign decision making tools.

REFERENCES

[1] ASTM C71-12, Standard Terminology Relating to Refractories , ASTM International, West Conshohocken,

PA, 2012, www.astm.org.

[2] Preuer, A., Winter, J., & Hiebler, H. (1992). Computation of the erosion in the hearth of a blast furnace, Steel

Research, 63, 147–151.

[2] Dong, H. & Shaojun, C. (2013, June). Mathematical model for fast computation of erosion profile in

submerged arc furnace with freeze lining, Proceedings of the thirteenth International Ferroalloys Congress

Efficient technologies in ferroalloy industry, INFACON XIII, 799-810, Almaty, Kazakhstan.

[3] Borbas, J. J., Padfield, R. C. & Moscker, E., (1967) Density determination of refractories by measurement of

gamma-radiation absorption. Journal of the American Ceramic Society, 50(8), 421-424.

[4] Staicu, L. & Radu, R. (1983). The use of the (γ, n) reaction to measure wear in blast furnaces and other

industrial furnaces: comparison with present methods. Journal of Physics D: Applied Physics. 16(12), 2541-

2546.

[5] Salgado, J., Oliveira, C., Moutinho, A., & Silvério, C. (1988). Control of refractory lining wear by using

radioisotopesInternational Journal of Radiation Applications and Instrumentation. Part A. Applied Radiation

and Isotopes. 39(12), 1265–1267.

[6] Prasad, A.S.; Sinha, P.; Qamrul, M.; Chatterjee, A.; & Chakravarty, P.K. (1979, July-October). Some

experience with radioisotopes in the study of the wear of blast furnace linings. Tata Iron and Steel Co. Ltd.,

Jamshedpur, India. Research and Development Division. TISCO; v. 26(3-4), 81-87 ISSN 0496-6546;

Worldcat.

[7] Perko, J.F. & Spirko, E.J. (1979, June) Blast furnace refractory lining wear status using radioactive sources.

Repub Steel Corp, Cleveland, Ohio. Ind. Heat; v. 42(6), 74-78.

[8] Bolf. N. (2004). Application of infrared thermography in chemical engineering. Journal of Chemists and

Chemical Engineers, 53(12), 549-555.

[9] Maldague, X. (2002). Introduction to NDT by active infrared thermography. Materials Evaluation. 6(9),

1060-1073.

[10] Gebski, P., Shameli, E., & Sadri, A., (2009, January). Non-destructive testing (NDT) techniques for

determination of refractory deterioration in smelting furnace; Molten 2009, Proceedings of the VIII

International Conference on Molten Slags, Fluxes and Salts, Editors: Mario Sanchez, Roberto Parra, Gabriel

Riveros and Carlos Diaz, Santiago, Chile.

[11] Ulitzka, S., Hadrich, H., & Gebhardth, F. (1995). Non-Destructive testing of ceramically bonded refractories

– Application of Ultrasond and Microwaves. Glastech. Ber. Glass Sci Technology. 68 (2), 51-58.

[12] Varghese, B., DeConick, C., Cartee, G., Zoughi, R., Velez, M., & Moore, R. (2004, July). High‐temperature

monitoring of refractory wall recession using frequency‐modulated continuous‐wave (FM‐CW) radar

techniques. American Institute of Physics Conference. AIP Conf. Proc. 760, 1714, Golden, Colorado, USA.

[13] Parker, R. L., Meanning, J.R., & Peterson N.C. (1985.). Application of pulse-echo ultrasonics to locate the

solid/liquid interface during solidification and meltring of steel and other metals. Journal of Applied Physics,

58 (11), 4150-4164.

[14] Sadri, A., Gebski, P., Ghorbani, H., McGarrie, G., & De Vries, T. (2009,). Monitoring Deterioration of

Waffle Cooler Thickness at Polokwane Smelter, JOM: A publication of the Minerals, Metals & Materials

Society, 61(10), 69-73.

[15] Sadri, A & Ying, W. L., (2015, May) Monitoring of Stave and Castable Refractory Wear in Blast Furnaces.

Proceedings of AISTech 2015, Cleveland, USA.

[16] Retrieved on January 21, 2015 from Techno-Consulting website: http://www.techno-

consulting.ru/en/about/index.html.

[17] Sansalone, M.J., & Streett W.B. (1997). Impact-Echo: Nondestructive Evaluation of Concrete and Masonry,

Bullbrier Press, Ithaca, NY. p. 339.

[18] Sansalone, M.J. & Carino, N. J. (1986). Impact-Echo: A Method for Flaw Detection in Concrete Using

Transient Stress Waves, National Bureau of Standards, USA, NBSIR 86–3452, NTIS PNB#87-104444/AS,

Springfield, Virginia, USA. p. 222.

[19] Carino, N. J., & Sansalone, M. J., (1988, August), Impact-Echo: A new method for inspecting construction

materials. Proceedings of Nondestructive Testing and Evaluation of Materials for Construction, University

of Illinois, IL., USA.

[20] Sadri, A. (2005, August). Non-destructive determination of refractory and build-up thickness in operating

furnaces using an acousto ultrasonic reflection technique. Proceedings of the Materials Degradation:

Innovation, Inspection, Control and Rehabilitation Symposium, COM2005. Calgary, Alberta, Canada.

[21] Sadri, A. & Walters, G. (2005, August). Determination of refractory and castable quality in operating

industrial furnaces, using a stress wave reflection technique. Proceedings of the Materials Degradation:

Innovation, Inspection, Control and Rehabilitation Symposium, COM2005. Calgary, Alberta, Canada.

[22] Sadri, A., Lachemi, M., & Walters, G. (2005, June). Determination of refractory lining thickness and quality

in operating industrial furnaces using a stress wave reflection technique. Proceedings of First CSCE Specialty

Conference on Infrastructure Technologies, Management and Policy. Toronto, Ontario, Canada.

[23] Sadri, A. (2008). An introduction to stress wave non-destructive testing and evaluation (NDT&E) of

metallurgical furnaces and refractory condition monitoring. CINDE Journal, 29(2), 7-11.

[24] Sadri, A. & Timmer, R. (2006, May) Blast furnace non-destructive testing (NDT) for defect detection and

refractory thickness measurements, AISTech 2006, Proceedings for Iron and Steel Technology Conference,

Vol. II, Cleveland, Ohio, USA.

[25] Sadri, A., Gordon, I., & Rampersad, A. (2007, May). Acousto ultrasonic-echo (AU-E): A non-destructive

testing technique for blast furnace hearth refractory condition monitoring., AdMet 2007, Proceedings of

International Conference Advances in Metallurgical Processes and Materials, Vol. 2. 77-85, Dnipropetrovsk,

Ukraine.

[26] Sadri, A., Marinelli, P., Doro E., Gebski, P., & Rampersad, A. (2009, May). Comparing the accuracy of

acousto ultrasonic-echo (NDT), finite element analysis (FEA), and drilling when obtaining a blast refractory

lining wear profile; AISTech 2009, Proceedings of Iron and Steel Technology Conference, St. Louis,

Missouri, USA.

[27] Sadri, A. & Gebski, P. (2009, October). Non-destructive testing (NDT) and inspection of the blast furnace

refractory lining by stress wave propagation technique. Proceedings of the 5th International Congress on the

Science and Technology of Ironmaking (ICSTI’09); 951-955, Shanghai, China.

[28] Sadri, A., Gebski, P., Mirkhani, K. & Ying, W. L., (2011, October) Monitoring refractory lining in operating

furnaces by acousto ultrasonic-echo technique; Proceedings of Conference of Metallurgists, Com 2011;

Montreal, Quebec, Canada.

[29] Sadri, A., Gebski, P. & Shameli, E. (2010, June). Refractory wear and lining profile determination in

operating electric furnaces using stress wave non-destructive testing (NDT). Proceedings of the Twelfth

International Ferroalloys Congress, INFACON XII. 881-890. Helsinki, Finland.

[30] Sadri, A., Ying, W.L., & Gebski, P. (2013, December) Application of specialized non-destructive testing

(NDT) for operating copper process vessels. Proceedings of Eight International Copper Conference 2013,

Instituto de Ingenieros de Minas de Chile (IIMCh). 813-833Santiago, Chile.

[31] Sadri, A., Gebski, P., Rojas V. P, Ibanez S. R. & Diaz. C. W. (2008, September) Inspection and evaluation

of refractory lining for slag cleaning furnace and Teniente reactor at Fundición Hernán Videla Lira Empresa

Nacional De Minería (ENAMI). Proceedings of MAPLA 2008, International Conference in Mine Plant

Maintenance, Edited by Eduardo Nilo, Raul Stegmaier and Patrico Guzman, September 3-5, 2008, Santiago,

Chile.

[32] Sadri, A., Ying, W.L., & Dempsie, B. (2013, September), Preventing costly smelting furnace and process

vessel run-outs (both in English and Spanish versions) , Proceedings of MAPLA 2013, 10th International

conference in Mining Plant Maintenance, Edited by Sergio Armstrong, Peter Knights, José Bolaños, and

Edmundo Mares, Santiago, Chile.

[33] Sadri, A, Mirkhani K., & Ying, W.L., (2012, September) Utilization of innovative NDT and condition

monitoring techniques for improving maintenance strategies of process vessels, Proceedings of MAPLA

2012, 9th International conference in Mining Plant Maintenance, Edited by Victor Babarovich, Alvaro Endo,

Rodrigo Pascual, and Raul Stegmaier, Santiago, Chile.