Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

J. C. Pandey, Manish Raj, Umesh Singhal, N. Bandyopadhyay and O. N. Mohanty

Tata Steel, Jamshedpur - 831 001, India

The integrity of the rolls in the roughing and finishing stands of HSM/CRM is important for the mills productivity and availability, inventory costs, man and machine hour costs and also, the quality of the rolled products. Therefore, the rolls should be allowed to contain only the minimum permissible flaws. Even small surface cracks or bruises, invisible to the naked eye, may affect the product quality. Standardisation of non-destructive evaluation (NDE) techniques is, therefore, essential for roll management. The paper highlights some non-destructive evaluation techniques developed in the R&D Division of Tata Steel, such as ultrasonic and eddy current testing, to meet the above objectives. These techniques have been implemented to assess the shell thickness, shell/core bond quality, fire-cracks and pressure cracks on the double poured work and back-up rolls of the roughing as well as the finishing stands of the HSM. Double poured rolls with indefinite chilled cast iron, high-chrome iron and high speed steel as the shell material are being routinely assessed using ultrasonic and eddy current in the Roll Shop of HSM. Although on-line eddy current inspection is being carried out in the the CRM Roll Shop, ultrasonic inspection using surface waves has also been implemented to measure crack depths less than 0.70 mm. This technique has been applied for the forged steel work rolls and the intermediate rolls of the CRM. These measures have helped to bring down the roll failure rate in the Mill, with consequent savings in down time cost, reduction in roll consumption and also, improvement in the product quality.

TYPES OF ROLLS IN HSM/CRM Roughing Stand of HSM

In the roughing stand of the 4-high Hot Strip Mill, there is normally a fire crack problem on the work roll surfaces because of the high heat impact from the hot slab. Besides the high heat impact, owing to heavy reduction of the slab, the high load and neck strength are also major concerns. Double poured chrome steel (as the shell) grades of rolls with SG iron (as the core) are used as work rolls. For back-up rolls, double poured cast steel rolls are used. The shell is made of high carbon alloyed steel (0.6-0.9 C, 1.5-2.2 Mn, 0.4-0.Si, 1.5-2.0 Cr, 0.2-0.5 Mo, 0.035% S and P max.), whereas the core consists of medium carbon unalloyed steel. In the future, some of the cast back-up rolls will be replaced by forged and hardened alloyed steel rolls.

Finishing Stand of HSM

The finishing stand of the Hot Strip Mill consists of six stands. Earlier, the first three stands (F1, F2 and F3) used double poured high-chrome iron rolls with highchrome iron as the shell and SG iron as the core material, whereas the last three stands (F4, F5 and F6) used double poured indefinite chilled cast iron with indefinite chilled cast iron as the shell and SG iron as the core material. Double poured high speed steel rolls are under trial in stand F2.

Cold Rolling Mill

In the 6 high TCM (tandem cold mill) and the 4 high SPM (skin pass mill) of the CRM, the work as well as the intermediate rolls are made of differentially hardened forged 3% Cr steel. The back-up rolls of the SPM are made of double poured alloyed cast iron with indefinite chilled cast iron as the shell and SG iron as the core material, whereas the back-up rolls of the TCM are made of differentially hardened forged 3% Cr or 5% Cr steel.

Key Words
Ultrasonic testing, Eddy current testing, Shell/core disbond, Surface cracks, Rolls, Non-destructive evaluation.

FACTORS DAMAGING THE ROLLS INTRODUCTION

The integrity of rolls used in a Hot Strip Mill (HSM)/Cold Rolling Mill (CRM) affects the mills performance in terms of productivity and availability, inventory costs, man and machine hour costs and also, the quality of the rolled products. Immediately after commissioning of the Hot Strip Mill at Tata Steel, owing to frequent failures of the rolls, the mill suffered in terms of mills productivity and availability. A need was felt to check the integrity of these rolls at the Roll Shop of HSM by the standardisation of NDT techniques (ultrasonic and eddy current).

Shell/Core Disbond in Double Poured Rolls

These rolls are centrifugally cast in which, two dissimilar metals (shell and core) are bonded with each other. The quality of this metallurgical bond depends on foundry variables like molten metal temperature, shell and core metal weights poured, spun cycle, quality of the flux applied on the hot shell, etc. In case of a poor bond or disbond area on the shell/core interface, when the roll diameter approaches the scrapping diameter, under the influence of internal stresses, the disbond size may become critical, be the source of crack propagation and lead to roll failure due to spalling.

Tata Search, 2003

338

J. C. Pandey et al. Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

Thermal Stresses
During each rotation, the work rolls of the early finishing stands or roughing stands experience different temperatures at different locations on the roll surface, which lead to thermal shocks and hence, generation of fire cracks. Fire cracks are generally found on the rolls of the roughing and the early finishing stands where the temperature of the rolled slab is high. During roll rotation, the maximum temperature of the roll surface which is in contact with the hot slab, has been reported as 613o C against the minimum roll surface temperature of 70o C, after passing through the water cooling zone in the early finishing stands (F1 - F3). It has been reported that the usual depth from the surface of the rolls at which the temperature drops to 320o C (where there is 15 % drop in the hardness when compared with the room temperature hardness) is 0.20 - 0.72. mm(1). This results in thermal stresses leading to fire cracks in the rolls. If such cracks join the poor quality shell/core bond interface, a part of the shell material can come out from the roll. Under the influence of stresses, the size of these cracks may become critical leading to roll failure.

EARLIER WORK
In the beginning of 1980s, when the primary Rolling Mills (1 and 2) and Sheet Mills were in operation in Tata Steel, attempts were made by the R&D Division to use the available NDT techniques in order to reduce the roll failures, particularly cog rolls of the roughing stands of the primary rolling mills and the clear chilled rolls of the finishing stands of the Sheet Mill owing to deep fire cracks. Thermography(5) was used to understand the effectiveness of spray cooling in the different passes of the cog rolls during mills operation and the AC potential drop method was used to measure the depths of the fire cracks in the cog rolls. These techniques were found useful in finding the reason for deep fire cracks in a particular pass where the cooling was insufficient. The cooling system was modified and the failures owing to deep fire cracks were reduced. The development of fire cracks in clear chilled rolls of the Sheet Mill was a common phenomenon and some of the deeper cracks (particularly transverse cracks) led to the failure of the rolls in two pieces. The AC potential drop method was used to measure the depths of these fire cracks and the rolls with crack depths more than 40 mm were taken out from the mill to reduce their failures. When the Sheet Mill was about to close down (1995-1996), the failure of clear chilled cast iron rolls of the finishing stands became alarming. Deeper chill depths in these rolls was suspected to be one of the reasons for the failures. An ultrasonic method was developed at R&D in order to non-destructively measure the chill depth in these rolls(6,7). Based on these measurements, recommendations were made to reduce the chill depth to a maximum of 7 mm. Using a 6 MHz twin crystal probe, the ability to detect back-scattered signals from the start of mottle (a globular aggregate of graphite, ferrite and cementite) was found to be the basis of these measurements. Fig. 1 shows some typical chilled depth profiles in two clear chilled cast iron rolls of the Sheet Mill.

Abnormal Mill Behaviour

Besides the damage mechanisms discussed above, abnormal mill behaviour during cobble or power trip, etc. can also damage the rolls. Bruise/soft spots are induced on the roll body surface when the temperature of that surface area during mill service exceeds the tempering temperature used by the roll manufacturer to obtain the desired hardness. A residual stress within the bruise is created as the retempered martensite contracts from the adjacent roll material. A stress crack (fine crack) is initiated which further propagates circumferentially generating a fatigue path leading to brittle fracture i.e. spalling. When there is a cobble, high loads (exceeding the normal loads), adhering metal weld on the surface, etc. localised contact stresses are developed, which lead to crack initiation and propagation leading to spalling.

10.00 O|C - 3483

INSPECTION PROCEDURES ELSEWHERE AND AT TATA STEEL

Hetchman Paul et al.(2) reported measurement of surface crack depths in rolls using ultrasonic surface wave probes having different frequencies. Takada Hazime et al.(3) reported the use of an on-line ultrasonic inspection system for the measurement of surface crack depths in high speed steel rolls in which, a broad band surface wave probe was used. On-line automatic inspection systems (EMAT ultrasonic and eddy current) for checking soundness of the back-up rolls at roll shops have been reported by K. Berner et al(4). Most of the advanced mills in Europe, USA and Japan use on-line NDT (ultrasonic and eddy current) systems and the acceptance norms are based on the calibration blocks with artificial defects. In the Hot Strip Mill of Tata Steel before 1996, the rolls were not inspected non-destructively and the acceptance of the rolls was based on the suppliers inspection report. In spite of the suppliers acceptance of the rolls, frequent failures of the rolls during 1996-1997 (mostly due to poor shell/core bond) led to the reinspection of the rolls at the Roll Shop at regular intervals. Hence, standardisation of nondestructive testing of rolls was taken up.
Chill depth, mm

8.00

6.00

O|C - 3957

4.00

2.00

0.00 1 3 5 7 9 11 13 15 17 19 21 23 25 Positions

Fig. 1 : Typical chilled depth profiles in two clear chilled cast iron rolls of the finishing stand of the Sheet Mill

PERFORMANCE OF ROLLS IN HOT STRIP MILL

In the years 1993 - 1997 after commissioning of the Hot Strip Mill, a sharp increase in the failure of HSM work rolls was a matter of great concern. The roll failure doubled

Tata Search, 2003

J. C. Pandey et al. Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

339

in the year 1997, compared with 1996 as is obvious from Fig. 2.

16 14 12 No. of Rolls failed 10 8 6 4 2 0 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 1999-2000 2000-01 2001-02 Year After implementation of ultrasonic and eddy current inspection techniques developed by R & D.

axis is the time scale and the y-axis represents the amplitude of the ultrasonic signal. The acoustic impedance (Z) of a material is given by the product of the material density () and velocity (v) of the ultrasonic wave in that material : Z=.v .....(1) If subscripts 1 and 2 denote the parameters for material and ultrasonic waves for the shell and core materials of centrifugally cast double poured rolls respectively, then for the shell, Z1 = 1 . v1 and for core, Z2 = 2 . v2 The reflection co-efficient, R is given by : R = (Z2 Z1)2 / (Z2 + Z1)2 .....(4) .....(2) .....(3)

Fig. 2 : Failure of double poured rolls at the Hot Strip Mill between 1993 to 2002

The failure analysis of these rolls revealed poor shell/core bonds, deep surface breaking cracks, and in some cases, abrupt variation in the shell thickness (eccentric shell in static cast double poured rolls). In centrifugally cast double poured rolls, failure beacuse of shell/core disbond was a common phenomenon (in some cases, the whole shell was found to be stripped out from the core). Fig. 3 shows the failure of a double poured centrifugally cast indefinite chilled cast iron roll because of poor shell/core bond quality at the Hot Strip Mill.

Experimental
From a broken roll of the HSM, having an indefinite chilled cast iron (ICCI) shell and a SG iron core, a sample piece of 100 mm thickness was cut for preparing a standard. Using the through-transmission technique, the ultrasonic velocity was measured at different locations from the roll surface as tabulated in Table 1 and plotted in Fig. 4. Table 1 : Ultrasonic velocity measured at different depths from the roll surface
Measuring points 1 2 3 4 5 6 7 8 9 Distance from roll surface, mm 10 20 30 40 50 60 70 80 90 Ultrasonic velocity, m/s 5263 5172 5169 5124 5210 5603 5641 5630 5656 Remarks ICCI shell ICCI shell ICCI shell ICCI shell Interface SG iron core SG iron core SG iron core SG iron core

Ultrasonic longitudinal velocity, m/s

Fig. 3 : A double poured indefinite chilled cast iron roll of the finishing stand of the Hot Strip Mill showing poor shell/core bond

5700 5600 5500 5400 5300 5200 5100 5000 0 10 20 30 40 50 60 70 80 90 100 Distance from the roll surface, mm

Hence, a need was felt to develop and implement NDT techniques to ensure the supply of only good quality rolls to the mills which would subsequently reduce incidents of roll failures. NDT techniques were used to detect and measure shell thickness, shell/core bond quality(6,7), to check the existence of graphite nodules in SG iron core of these double poured rolls(8), and surface cracks on the high speed steel rolls of the HSM.

MEASUREMENT OF SHELL THICKNESS AND SHELL/CORE BOND QUALITY Theoretical Aspects

When an ultrasonic wave passes through a medium, because of the resistance offered by the medium, the ultrasonic wave gets attenuated. Depending upon the characteristics of the medium, the extent of attenuation varies. The resistance is characterised by acoustic impedance and ultrasonic waves get reflected from the point where there is a change in the acoustic impedance. Since the shell/core interface provides substantial change in the impedence, some parts of ultrasonic waves are reflected and displayed as a rectified signal on a calibrated CRO screen of the ultrasonic instrument, where the x-

Fig. 4 : Variation of ultrasonic longitudinal velocity the indefinite chilled double poured cast iron rolls of the Hot Strip Mill

At 50 mm below the roll surface, a change in the measured ultrasonic velocity was observed, which indicated the interface of the shell and core. Fig. 5 shows a macro-etched sample of a broken piece from a failed roll. The white portion shows the shell thickness, which was also confirmed by the shell thickness measured by the ultrasonic testing. Fig. 6 shows photomicrographs of indefinite chilled cast iron shell, shell/core interface and the SG iron core. From these experimental results, if the average velocities in the ICCI shell and the SG iron core are 5200 m/s and

Tata Search, 2003

340

J. C. Pandey et al. Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

Table 2 : Ultrasonically measured shell thickness in different rolls of the HSM

Roll No. Roll diameter (d), mm 994. 84 674. 00 702. 85 695. 48 685. 53 Ultrasonically measured shell thickness (X), mm 50. 00 44. 00 47. 00 50. 50 50. 00 Scrap diameter of roll (ds), mm 950 660 660 660 660 Useful shell thickness in scrapped roll (Xs) = 1/2 [ds-(d-2X)], mm 28 37 26 32 37

GP 56760 01-DG-03052 01-SC-325 01-SC-325 01-SH-0008

selected from a scrapped good roll which had given satisfactory life in service. Investigations were carried out for :
5

ultrasonic evaluation of all failed rolls, particularly due to poor shell/core bond leading to spalling ultrasonic evaluation of new in-coming rolls ultrasonic evaluation of the rolls in service.

Fig. 5 : Macro-etched sample of a broken piece from a failed roll

5 5

The test results are shown in Fig. 7.

Ref. scrap roll Eccentric Shell Disbond Good bond

100

(a)

(b)

(c)
Interfacial echo ht., %

80

Fig. 6 : Photo-micrographs of (a) indefinite chilled cast iron shell, (b) shell/core interface and (c) SG iron core (magnification X 100)

60

5600 m/s respectively, and considering the density of iron as 7.8 g/cm3, the reflection co-efficient (R) = 0.00137 for proper shell/core bonding. In case of shell/core disbond (lack of metal to transmit the ultrasound), the value of R will increase, depending on the size of the disbond. In case of complete disbond, the value of R will be 0.99996, assuming air as the medium in the disbond. Although in conventional ultrasonic flaw detectors it is not possible to measure the absolute values of the reflection co-efficient from the shell/core interface, it can be indicated by the amplitude (after amplification) of the reflected ultrasonic pulsed signals.

40

20

0 10

20

30

40

50

60

70

Shell thickn

Fig. 7 : Shell/core interfacial echo heights for different used and failed, as well as new indefinite chilled cast iron rolls of the HSM.

Measurements on Rolls
Table 2 shows some typical ultrasonically measured shell thicknesses in double poured rolls of the HSM. The results indicated shell thickness values between 44 - 50.5 mm. This indicates that even if the rolls approach the scraproll diameter, there will be sufficient useful shell thickness of 26 - 37 mm. The variation in shell thickness of about 11 mm needs to be further reduced in order to economically manufacture and supply these rolls to the customers.

Using a 4 MHz normal probe, ultrasonic instrument at 40 dB attenuator setting, it was observed that good bond rolls showed shell/core interfacial echo heights less than 25%.

ESTIMATION OF GRAPHITE NODULES IN SG IRON CORES Theoretical Aspects

In solids, the ultrasonic longitudinal velocity (Vl) can be expressed in the following manner : E (1- ) E . (1 + )(1 2 ) .....(5)

Assessment of Shell/Core Bond Quality

The shell/core interfacial echo in ultrasonic testing is an indicator of the bond quality. A good bond roll will show a lower interfacial echo, whereas a poor bond roll will show a higher interfacial echo. In case the disbond area is more than the probe crystal diameter, multiple reflection echoes may appear on the oscilloscope screen of the ultrasonic flaw detector. Since it is difficult to cut a sample from a new roll, a reference roll was

V1 =

where, E = Youngs modulus of elasticity = Density = Poissons ratio

Tata Search, 2003

J. C. Pandey et al. Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

341

It is seen from equation (5) that the ultrasonic longitudinal velocity (Vl) is dependent on Youngs modulus of elasticity (E), Density () and Poissons ratio (). In cast iron, the graphite phase affects the values of E and hence, affects the measured values of the ultrasonic longitudinal velocity (Vl) in these materials. A transformation from flaky to spherical form in SG iron increases the value of the ultrasonic longitudinal velocity. Longitudinal ultrasonic velocities measured by different authors E. P. Papadakis(9), R. J. Klink(10), H. E. Henderson(11) and A. G. Fuller(12) were found to be 5690, 5720, 5740 and 5560 m/s respectively. The velocity difference between ferritic and pearlitic SG iron was found to be 10 m/s, which commercial ultrasonic equipment cannot differentiate. Since 100 m/s velocity difference was found between cast and heat treated SG iron castings, there was no problem in using the commercial equipment.

Fig. 9 : A photograph of a band on the barrel surface containing deep fire cracks

Theoretical Aspects
Ultrasonic surface waves are generated by wave mode conversion at the boundary of two media. When a longitudinal wave traveling at speed VL1 in medium I arrives at the boundary medium I and II as shown in Fig. 10, it is reflected(13).

Measurements on Rolls
Total transit time (T) required for an ultrasonic wave to travel through the barrel diameter is : T= 2x D 2x + vs Vc .....(6)

where, x = shell thickness, mm Vs = ultrasonic velocity in the shell T = transit time required for the ultrasonic wave to travel through the barrel diameter, s The measurements on indefinite chilled cast iron rolls with SG iron cores are shown in Fig. 8.
5780
Ultrasonic longitudinal velocity, m/s

5770 5760 5750 5740 5730 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 No. of rolls tested

Fig.10 : Wave mode conversion of ultrasound at a boundary of medium I and II

Fig. 8 : Typical ultrasonic longitudinal velocity distribution in SG iron cores of indefinite chilled cast iron rolls of the HSM

The results show that the ultrasonic longitudinal velocity values lie within 5699 5782 indicating the presence of graphite nodules in the SG iron cores.

The refracted longitudinal wave in medium II deviates from its original path and travels with a velocity VL2. Due to mode conversion, another component of the wave is the shear wave which is reflected in medium I and travels with velocity VS1. Refracted shear wave in medium II travels with velocity VS2. Longitudinal wave velocities are always greater than the transverse wave velocities. Therefore, their relative angular positions are different. If, 1 = Angle of incidence of a longitudinal wave in medium I and also the angle of reflection for the same wave in medium I 2 = Angle of refraction for the transmitted longitudinal wave in medium II 1 = Angle of reflection for the shear wave in medium I 2 = Angle of refraction for the transmitted shear wave in medium II then, using Snells law, Sin 2/Sin 1 = VL2/VL1 Sin 1/Sin 1 = VS1/VS2 .....(7) .....(8)

ULTRASONIC EVALUATION OF SURFACE CRACKS IN ROLLS

Surface cracks in the HSM rolls occur either due to manufacturing defects and/or due to thermal and mechanical stresses. Fig. 9 shows a photograph of such cracks developed during their service in the mills. After roll dressing, if these cracks are missed during visual inspection, they will open-up during operation in the mill and leave marks on the strips causing rejection. If the depth of these cracks approaches the scrap diameter zone, they would make the roll unsuitable for use.

Tata Search, 2003

342

J. C. Pandey et al. Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

There is a second critical angle when the shear wave is also totally internally reflected and no sound wave penetrates into the specimen which is 56.60 for steel using a plexiglass wedge. Surface (Rayleigh) waves are induced at solid/liquid or solid/gas interfaces. When the probe is at the second critical angle setting, the surface wave profile is approximately a shear wave at its critical angle. Depending on the type of particle vibration and medium used, the surface waves have been classified as follows :
5

Echo amplitude, % FSH

There is a critical angle of incidence for the longitudinal wave where 2 = 90o and the longitudinal wave does not penetrate in the medium II across the boundary. This is the principle of generating angular shear waves within a specimen where the ultrasonic transducer generating longitudinal wave is set at an angle greater than the first critical angle. Plexiglass is often used on a steel specimen and the critical angle is 26.7o.

110 100 90 80 70 60 50 40 30 20 10 0 0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6

Depth of steps of calibration blocks, mm

Rayleigh waves in which, the particles vibrate in an elliptical orbit and the medium can be any surface where the major axis is perpendicular to the surface and the minor axis is parallel to the direction of wave propagation. Plate waves (Love) in which, the particles vibrate parallel to the plane layer and perpendicular to the wave direction. The medium is the surface layer bonded to solid. Plate waves (Lamb) in which, the components of vibrations are perpendicular to the surface. Stanely (Leaky Rayleigh waves) in which, the waves are guided along the interface. The medium can be the boundary of dissimilar solids. Sezawa in which, the particle vibrates in antisymmetric mode and the medium is layered solids.

Fig. 11 : Variation in echo amplitude (% FSH) in the calibration blocks with steps of varying depths, using 2 MHz, surface wave probe at 49 dB attenuator setting

the minimum detectable beam path as 10 mm, for the commercial flaw detectors, the minimum detectable crack depths using different beam angles are shown in Table 3.

Ultrasonic M Crack A O B probc N MN = Specimen surface AB = Surface breaking crack OB = Ultrasonic beam path for angle beam probe

Usually the depth of penetration of the surfac (Rayleigh) waves is one wavelength below the roll surface. Surface cracks with depths less than one wave length of ultrasound can be measured if the amplitude of the reflected signals from these cracks is properly calibrated. To calibrate the ultrasonic equipment, calibration blocks with various steps 0.55 to 3.54 mm were fabricated from mild steel. Fig. 11 shows the response of the ultrasonic waves towards the various steps of calibration blocks in terms of % FSH (full scale height) on the oscilloscope screen. The graph shows that there is a gradual increase in the echo amplitude with an increase in the depth of the step up to 1/2 wavelength of the ultrasound (= 0.7 mm), beyond which, there is a decrease in the echo amplitude with increase in the step-depth. Further, beyond one wave length, there is no systematic response of ultrasonic waves, which can be used for crack depth measurement. However, the crack depths up to 0.7 mm can be measured using this calibration curve. By lowering the frequency of the surface wave probe, the depth of the ultrasonic surface wave penetration can be increased and higher crack depths can be measured. For deeper cracks, more than 1.4 mm, 80o, 70o, 60o, 45o probes can be tried. As shown in Fig. 12, crack depth, AB = OB cos a, where a = probe angle and OB = beam path. Considering

Fig. 12 : Schematic diagram showing measurement of crack depth using an angle beam probe

Table 3 : Maximum measurable crack depth using surface wave and minimum measurable crack depth using angle probes
Surface wave probes Probe frequency, MHz 4.0 2.0 1.0 0.5 Max. crack depth measured, mm 0.35 0.70 1.4 2.8 Angle beam probe at 2 MHz frequency Probe angle, o Minimum detectable crack depth, mm 1.74 3.42 5.00 7.07

80 70 60 45

Crack Depth Measurements on Rolls

Figs. 13 (a) and (b) show a schematic diagram for measurement of the surface crack depth of a typically oriented crack using the ultrasonic NDT method. During one mill interruption, ultrasonic evaluation revealed that the high chrome steel rolls of the roughing stand had developed crack depths in the range 13-35 mm, whereas the indefinite chilled cast iron rolls of the finishing stand F4 and F6 had developed fire cracks in the range 2-14 mm.

Tata Search, 2003

J. C. Pandey et al. Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

343

Detector

Inspection coil

Direction of magnetic fields

Fig. 13 (a) : Crack depth measurement using ultrasonic angle o o beam (70 /80 ) probe

Discontinuity

Co n spe duct cim ing en

Eddy current

Fig. 14 : Schematic diagram of eddy current testing

For AC : V= iZ
Fig. 13 (b) : Crack depth measurement using ultrasonic normal beam probe

.....(10) .....(11)

I = Io Sin t

where, = angular frequency = 2f (f is the frequency)

DETECTION AND MEASUREMENT OF SURFACE CRACKS IN ROLLS BY EDDY CURRENT Theoretical Aspects
Eddy currents are the induced electric currents which are generated in a conducting material owing to changes in the magnetic field. An AC source is applied to an inspection excitation coil so that the magnetic lines of force penetrate into the specimen conducting surface providing a good magnetic coupling as shown in Fig.14. Eddy currents circulate on the specimen surface and are modified by the presence of discontinuities.(i.e. surface cracks). Eddy currents flow parallel to the plane of the windings of the coil. The detector instrument notes the changes with time variation of voltage and current in the specimen coil(13). The depth of penetration of the eddy current () in metres, within a material is given by : = 50 f r where, = Electrical resistivity in Ohm-cm f = Frequency of AC in Hertz r = Magnetic permeability of the material The depth of penetration () increases with an increase in electrical resistivity ( ), decrease in magnetic permeability (r) and frequency (f) of the AC current of the material. In rolls, the material microstructure influences the magnetic permeability and resistivity and hence, the depth of eddy current penetration. An eddy current testing circuit is equivalent to a resistance R and inductance L. .....(9)

Then, V L.di/dt = iR

.....(12)

Since, the inductance opposes the applied voltage, the induced voltage : VL = -L.di/dt = -L..io. cos t = - XL.io. cos t where, XL = L .....(13)

Current i and induced voltage VL are always 90o out of phase, V - XL. io. cos t = io.R. t .....(14) Therefore, V= io.(R. Sin t + XL. cos t) If i=V/Z where, Z is the impedance Z = R. Sin wt + XL. cos t The magnitude of Z IZI = (R2 + XL2) 1/2 = [R2 + (L) 2]
1/2

.....(15) .....(16)

.....(17)

.....(18)

and phase angle between Z and R = tan-1 (XL/R) = tan-1 (L/R)

.....(19)

The inspection coil circuit measures the resistive component R which is fed to the X-plate and the reactance component XL is fed to the Y-plate of the oscilloscope.

Comrison between Eddy Current and Ultrasonic Testing

A detailed comparison between eddy current and ultrasonic testing is shown in Table 4.

Tata Search, 2003

344

J. C. Pandey et al. Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

Table 4 : Comparison between ultrasonic and eddy current testing

Ultrasonic testing
5

Eddy current testing It is based on the electrical and magnetic properties of the specimen to be tested. Eddy current generation is confined to its limited depth from the surface. Hence not suitable for checking internal soundness. Eddy current is ideal non-destructive method for measuring the path length of the shallow surface cracks i.e. fire cracks and pressure cracks in the rolls of HSM/CRM. The probe (excitation coil) does not remain in contact with the specimen during testing. Hence, the inspection is fast and suitable for on-line inspection. On-line eddy current system is being used at the Roll Shop of CRM of Tata Steel.

It is based on the acoustic properties of the specimen to be tested. Internal soundness can be checked i.e. suitable for inspection of shell/core interfacial flaws and flaws within the shell of a double poured rolls of HSM/CRM. Conventional flaw detectors are not suitable for measuring depths of shallow surface cracks due to their inherent dead zone problems. Conventional manual ultrasonic testing is slow in speed due to application of couplant on the surface. However, EMAT systems do not require any couplant and are non-contact with the specimen hence are comparable to eddy current in speed and are being used for online inspection of rolls. It is possible to measure depths of the surface cracks (normal as well as oriented to the roll surface) in the rolls using transverse surface wave, transverse angle beam and longitudinal wave normal probes. In ultrasonic testing beam penetration is not the problem like eddy current and the crack depths which cannot be measured by eddy current (due to penetration problem) can be measured by ultrasonic testing. Effective radial depth has been reported(14) to be : 0-1.27 mm for surface wave 12.7-mm bore for straight beam

Fig. 15 : Microstructure of special high speed steel calibration block (magnification X 400) showing complex carbide of Fe, W, V and Cr in temper martensitic matrix (Etchant : 3% Nital)

8 7

X-axis deflection, units

Crack path lengths can be measured by using calibrated eddy current instrument, but crack depths of the oriented cracks cannot be measured. Depth of penetration of eddy current in ferrous metals (i.e. cast iron, steel, etc.) is lower than that in non-ferrous metals and austenitic steels. Hence, only shallow cracks can be detected and measured in cast iron/steel rolls. Effective radial depth has been reported (14) to be : 0-0.076 mm

6 5 4 3 2 1 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Notch depth, mm

Fig. 16 (a) : Calibration curve and empirical equation to measure crack depth in high speed steel at 64 kHz frequency in eddy current testing

Surface micro-cracks <0.152 mm and macro-cracks >0.152 mm both can be detected using surface wave probe(14). Using pitch catch dual probe and straight beam probe, subsurface flaws can be detected(14). It is not possible to detect work hardened and bruise/soft spots on the rolls(14).

Surface micro-cracks can be missed while macro-cracks can be detected (14). There are chances of missing subsurface flaws(14). Work hardened and bruise/ soft spots can be detected(14).

14 12 10 8 6 4 2 0 0 0.05 0.1 0.15 0.2 0.25

Experimental
For detection and measurement of surface cracks in high speed steel rolls, a calibration block made of special high speed steel with composition similar to that of the high speed steel roll material was made. Fig. 15 shows the photo-micrograph of such a block. Figs. 16 (a) and (b) show the calibration curves using this calibration block with EDM notches of depths 0.05, 0.10, 0.15, 0.20, 0.25, 0.30 and 0.35 mm. Fig. 16 (a) shows the calibration curves for 64 kHz frequency whereas, Fig. 16 (b) shows it for 128 kHz. Good correlation was obtained between the depth of the EDM notches and eddy current response (X-axis deflection, in units). These calibration curves were used for measuring the crack depth in high speed steel rolls.

X-axis deflection, units

Notch depth, mm

Fig. 16 (b) : Calibration curve and empirical equation to measure crack depth in high speed steel at 128 kHz frequency in eddy current testing

Measurements on High Speed Steel Rolls

The eddy current test results for high speed steel rolls are shown in Table 5.

Tata Search, 2003

J. C. Pandey et al. Non-destructive Evaluation of the Integrity of HSM and CRM Rolls

345

Table 5 : Eddy current test results of high speed steel rolls of HSM
Roll No. 1457 1458 1441 Roll diameter, mm 714.18 714.13 714.43 Maximum crack depth, mm 0.10 0.10 > 0.15 Remarks Crack was eliminated by grinding. Crack was eliminated by grinding. Eddy current did not inform the depth of crack. Ultrasonic 70o angle probe could measure the crack depth 3.4 mm which was eliminated by turning and grinding.

ACKNOWLEDGEMENT
The authors are thankful to the Management of Tata Steel for permitting to publish this paper. Thanks are also due to the staff of the Roll Shop of the Hot Strip Mill, and M/s GIIB, Jamshedpur for their cooperation in conducting the experiments, and implementation of different NDT techniques.

REFERENCES
1. J. H. Ryu, O. Kwon, P. J. Lee and Y M Kim, Evaluation of the Finishing Roll Surface Deterioration at Hot Strip Mill, ISIJ International, Vol. 32, 1992, pp. 1221 - 1223. 2. Paul Hetchman, John Ballani and Roger Zorn, Surface Crack Depth Classification through Improved Ultrasonic Techniques, 36th Mechanical Working and Steel Processing, Iron & Steel Society, Vol. 32, 1994, pp. 21-24. 3. Takada Hazime et al., Development of Roll Surface Technique by Use of Broad Bandwidth Surface Wave, 15th World Conf. on Non-destructive Testing, Rome, 2000. 4. K. Berner et. al., Automatic Inspection of Backup and Work Rolls with Ultrasonics and Eddy Currents, Proc. Int. Seminar on Roll Tech 2000, Jamshedpur, April 11-12, 1997, pp. 80-85. 5. B. N. Panda, B. S. Bhatia, J. C. Pandey and Amit Chatterjee, Plant Condition Monitoring by Thermography, Transaction of the Indian Institute of Metals, Vol. 57 No.1, 1984. 6. J. C. Pandey, G. Kaur. A. S. Prasad and O. N. Mohanty, Quality Assessment of Rolls : Use of Ultrasonic Back - scatter Technique, Proc. Int. Conf. on Recent Adv. in Metallurgical Processes, II Sc., Bangalore, July, 1997. 7. J. C. Pandey, Manish Raj and B. N. Panda, Evaluation of Integrity of Double Poured Rolls of Hot Strip Mill by Ultrasonic Technique, Proc. Int. Seminar on Roll Tech - 2000, Jamshedpur, April 11-12, 1997. 8. J. C. Pandey, Manish Raj, T. C. Janghel and B. N. Panda, Ultrasonic Velocity Technique for Quality Assurance of SG Iron Core in Centrifugal Cast D P Rolls of Hot Strip Mill, Proceeding, NDT in Steel and Allied Industries 98, Jamshedpur, April 2-3, 1998. 9. E. P. Papadakis, Metal Trans. Vol. 2, 1971, pp. 575578. 10. R. J. Klink, Personal Communication to E. P. Papadakis, 1975, Ford Motor Co. Dearborn Mich. 11. H. E. Henderson, Iron Worker, Vol. 40, No 3, 1976, pp. 15-19. 12. A. G. Fuller, Trans. AFS, Vol. 85, 1977, pp. 509-526. 13. Lois Cartz, Non-Destructive Testing, 1st Ed. ASM International, 1995, p. 94,175,182,183. 14. George A. Ott, Roll Tech 2000, April 11-12, 1997, p. 55.

For high speed steel roll numbers 1457 and 1458, the maximum crack depth measured was 0.10 mm, which was easily eliminated by grinding. However, the crack depth in roll no. 1441 was found to be > 0.15 mm, which could not be measured by eddy current. Using a 70o angle probe in ultrasonic testing, it could be measured and it was found to be 3.4 mm which was removed by roll turning and grinding.

CONCLUSIONS
Non-destructive evaluation (NDE) plays an important role in detecting and measuring the flaws in rolls to improve their performance. Efforts were made to develop and implement ultrasonic and eddy current techniques at the Roll Shop of the Hot Strip Mill. The implementation resulted in reduction of roll failures. Comparison of roll failure data for the period 1993-1997 (4 years) with that for the period 1997-2001 (4 years) after implementation of NDT techniques in the Roll Shop of HSM shows that roll failure was reduced from 31 rolls to 20 rolls and thus, 11 rolls were saved from failure. This has led to the following advantages :
5

Mills down time cost due to failure of roll in the mill is approximately Rs. 1 million. Hence, there is a saving in the mill down time cost of Rs. 11 million for the period of 4 years. Assuming 50 % life achieved by the failed rolls and cost of each roll being Rs. 0.8 million, there is a saving in the roll material of Rs. 4.4 million for the period of 4 years. Besides savings in mill down time cost and roll materials cost, other benefits (not quantified) were reduction in strip rejection due to roll failure, savings in man and machine hour costs, etc. Shell thickness and the shell/core bond quality in centrifugally cast double poured rolls of the finishing stand of the Hot Strip Mill by ultrasonic method using a normal beam probe. Surface cracks in high speed steel rolls up to 0.7 mm depth using ultrasonic surface wave probes and 1.74 to 5.00 mm using 80o, 70o and 60o angle probes were seasured. For angularly oriented cracks, normal probes were also been used. Surface cracks were found using eddy current in high speed steel rolls under trial in finishing stand F2. Calibrations have been made for crack depths up to 0.25 mm.

The NDT techniques developed were used to assess :

5

Tata Search, 2003