07380 - Corrosion Sensors for Detecting Graphitization of ...matergenics.com/wp-content/uploads/2018/04/07380... · CORROSION SENSORS FOR DETECTING GRAPHITIZATION OF CAST IRON IN

CORROSION SENSORS FOR DETECTING GRAPHITIZATION OF CAST IRON IN WATER MAINS

Mehrooz Zamanzadeh, Gordon C. Kirkwood, Sam Scheinman, and George T. Bayer Matco Associates, Inc.

4640 Campbells Run Road Pittsburgh, Pennsylvania, 15205

ABSTRACT This paper describes a new method for the non-destructive detection of graphitization corrosion in gray iron pipe, based on magnetic flux density measurements. Graphitization corrosion is unique to gray cast iron, which comprises much of the municipal water distribution infrastructure in the United States. Graphitic corrosion is particularly insidious in that graphitized pipe may appear perfectly sound upon visual inspection, despite being embrittled and prone to premature failure under load. Keywords: sensor, graphitization, water main, gray cast iron, water main failure, corrosion detection.

INTRODUCTION Graphitic corrosion, or graphitization, occurs when the metallic constituents of gray iron are selectively removed or converted into corrosion products. This process leaves behind the graphite matrix of the gray iron, in the shape of the original casting. While pipes undergoing graphitization may appear sound and may conduct water adequately, the metallic portion of the pipe wall may, in places, be significantly thinner than the apparent thickness of the wall. Graphitized regions of pipe wall will be brittle and subject to failure under load as the result of temperature variation, heavy traffic, or shock. Water main failures are very expensive for municipalities. Not only do they incur expenses in terms of repair, flooding damage, and loss of revenue to affected businesses, but water main failures can potentially interrupt the operation of vital services including medical care and fire fighting operations. Currently, millions of dollars are spent annually by industry and municipalities on the repair of failed gray iron pipe. The rate of failure will only increase in the future as the existing gray iron infrastructure continues to age. Therefore, it is important to develop a sensing technique that will allow for the non-destructive detection of graphitization before failures occur. This will enable repair of graphitized pipe to be undertaken before failure, and so minimize the expense incurred due to corrosion in the water distribution infrastructure.

1

BACKGROUND

Gray Iron Metallurgy Krause1 has discussed the metallurgy of gray iron in detail. The most important elements in gray cast iron, aside from iron, are carbon and silicon. The silicon content affects the carbon distribution in the metal. Unlike the carbon in ductile iron and steel, which is disbursed as graphite spheroids and pearlite, respectively, the carbon in gray iron is present in flake form. These flakes form in the eutectic cell boundaries during cooling of the cast metal. As a result, the graphite flakes form a continuous matrix throughout the gray iron. Increasing the silicon content decreases the amount of carbon present in the eutectic, causing more carbon to take the form of pearlite and less to be present in the graphite matrix, in flake form. This lowering of the flake graphite content of the iron results in increased tensile strength. Graphitic Corrosion Graphitic corrosion is one example of the dealloying of a metal. During dealloying, one component of an alloy is selectively dissolved, leaving other components behind. In the case of gray iron, the preferential attack on iron results from graphite’s highly noble, or corrosion resistant, position in the galvanic activity series. The relative position of two metals in the galvanic activity series determines which will most readily participate in electrochemical reactions, such as corrosion. The galvanizing of iron with zinc inhibits corrosion for this reason: iron is more noble in the activity series than zinc. Therefore the zinc plating layer is preferentially attacked, greatly extending the service life of the iron substrate. Similarly, graphite is far more noble than iron. Given the right conditions of soil composition and moisture, the graphite matrix within the gray iron can act as the cathode in an electrochemical reaction, and the iron will be undergo anodic attack. And, just as zinc galvanization protects iron substrate, so the graphite matrix of gray iron survives while the iron is dissolved away. Figure 1 presents a full cross section photograph of a graphitized pipe. Figure 2 presents a close-up photograph of the cross section of a graphitized pipe. Magnetic Properties of Materials Magnetic permeability is that property of greatest interest to the present investigation of a method for detecting graphitic corrosion. The permeability is denoted by the Greek character µ and is a measure of how easily magnetic fields will penetrate a material. More often, the “relative permeability” of a material, µr, is used.2 This is the permeability of the material relative to that of free space, in dimensionless units. It can be thought of as roughly analogous to the idea of electrical conductance, applied to magnetic fields instead of electric current. Since the conductance is the reciprocal of resistance, 1/R, the greater the conductance, the easier it is for currents to flow in a material. Analogously, magnetic fields more readily pass through materials having higher values of µr. The analogy isn’t exact, however. All (non-superconducting) materials possess a conductance between zero and one. That is to say that, inserting an additional length of any material into a circuit can only serve to reduce the total current flowing, as the new material will introduce more resistance. The magnetic permeability is not so constrained. It can take on values ranging from less than zero to much greater than one. Materials possessing µr less than zero actually oppose applied magnetic fields. As it happens, graphite is such a material. On the other end of the spectrum, some materials respond to applied magnetic fields by becoming strongly magnetically polarized in the direction of the applied field. This is the ferromagnetic behavior exhibited by iron. Ferromagnetic materials possess µr values greater then 1. Due to the contribution of

2

the material’s own magnetization, the field strength near the material is actually increased, above that of the applied field in free space.

A CASE HISTORY OF GRAPHITIZATION-INDUCED WATER MAIN FAILURE

Analysis was undertaken on a sample of gray cast iron pipe, from a water main which had fractured due to graphitization corrosion.3

Chemical analysis of the sample indicated the presence of carbon (3.11%), manganese (0.39%), phosphorus (0.39%), sulfur (0.073%), and silicon (1.59%) in quantities meeting the chemical requirements for gray cast iron pipe. The Brinell hardness of the pipe was 83.5HB. The hardness near the fracture measured 82HB. The pipe met the Talbot test (modulus of rupture and secant modulus of elasticity) requirements for gray cast iron. Visual inspection revealed primary and secondary cracks on the outside surface of the pipe. The crack had initiated at the outer surface and propagated inward, resulting in a longitudinal fracture. The inside of the pipe was coated with cement. This coating was continuous and strongly adherent. No evidence of corrosion was present on the internal pipe surface. Micrographs reveal a carbon distribution consistent with ASTM standards for gray iron, with graphite, ferrite, and pearlite visible. The graphite was of ASTM A247 type B. The soil above and below the failed pipe had a measured resistivity of 1100 to 2300 ohm-cm, which falls into the category of “corrosive.” The municipality involved experienced an increasing time-rate of failures. This increase in failure rate is typical of a piping system undergoing corrosion. Were the failures due to brittleness alone, no increase in the time rate of failure (decrease in mean time between failures) would be observed. In light of the evidence above, it is probable that the cracking and failure resulted from graphitization corrosion and wall thinning which occurred after the pipe was installed. The longitudinal fractures then occurred when the pipe was subjected to a crushing load at locations weakened by graphitization. Visual observation, optical microscopy, and EDS (Energy Dispersive X-Ray Spectroscopy for analysis of chemical composition) microanalysis showed several features of interest:

1. Localized corrosion was visible on the exterior surface of the pipe. 2. Examination of the internal surface showed no signs of either localized or uniform corrosion. 3. Metallographic examination of the cross section showed the characteristic appearance of localized

graphitization with 25% penetration of the pipe wall. The cost of repairing damages caused by broken water mains (and subsequent flooding damage) may become an important item in many municipal budgets. This could be minimized by the evaluation of pipe and the detection of localized corrosion or graphitization before actual failures occur. A typical program for evaluation will include:

1. Identification of microstructure (gray iron, ductile iron, or other). 2. Identification of corrosion mechanisms. 3. Determination of the extent of internal and external corrosion (maximum and minimum wall thickness). 4. Finally, analysis of data and determination of preventative action. Monitoring should be continued for a

few years beyond the application of corrosion control measures. These are the concerns that motivated the current work in the development of sensor techniques for the detection of graphitization corrosion in gray iron.

3

MATERIAL PROPERTIES POTENTIALLY USEFUL FOR ASSESSING GRAPHITIZATION Several properties of iron could, potentially, be used detect graphitization. These include magnetic permeability, ductility, electrical resistivity, and the acoustic properties of sound velocity and attenuation. Of these, only magnetic permeability is suitable for non-destructive analysis and amenable to an implementation sufficiently robust to be used in the field, to detect graphitic corrosion in situ, in exposed buried pipe. Ductility can be ruled out immediately because its measurement, by nature, involves destructive techniques. Point-to-point measurements of electrical resistivity yield poor localization of corroded regions, as the current path through an inhomogeneous pipe wall will not necessarily follow a straight line. And, as ultrasonic measurement of acoustic properties requires a very clean interface between the probe and pipe for purposes of acoustic transmission and impedance matching, it is poorly suited for use with exposed buried pipe which is often wet and dirty. Also, neither electrical resistivity nor acoustic methods are suitable for use with a wide variety of pipe coatings, as both methods require surface contact with bare, clean metal. Magnetic permeability can deliver highly localized measurements and is vastly more tolerant than electrical or acoustic methods with respect to surface preparation. Therefore, a magnetic measurement method was selected for use in the instrument described below.

THEORY OF OPERATION Iron and graphite have very different magnetic properties. Iron is, of course, ferromagnetic. This means that in the presence of an external magnetic field, the magnetic domains in the metal tend to align with the applied field. This results in high magnetization (magnetic polarization) of the metal and high permeability to applied fields. The relative permeability, µr, of pure, non-heat-treated iron is generally around 200 and can be higher. µr is a measure of the magnetic permeability of a material relative to that of free space, in dimensionless units. Annealed iron can attain relative permeability as high as 5,000. On the other hand, graphite is diamagnetic, weakly opposing applied fields. This results in a relative magnetic permeability not just much lower than that of iron but actually less then zero, namely -0.6. Due to the selective removal of iron, graphitized areas of gray iron will exhibit reduced magnetic permeability. This change in permeability is the basis of the sensing techniques described below. Two methods have been employed by this group to take advantage of this change in permeability: magnetic force sensing and magnetic flux density measurement. Magnetic Force Sensing: In the magnetic force sensing scheme, an external field is applied by a permanent magnet, at the end of a rigid beam. The hand held sensing apparatus is designed so that, when placed against a pipe, the magnet is in close proximity to the test surface. Measuring the deflection of the beam (by, for example, a strain gauge) gives an indication of the force of attraction between the magnet and the metal. The decreased permeability in graphitized regions will result in lower flux density through; and measurably lower magnetic force over; graphitized regions as compared with intact regions. Note that such a force-based device will be sensitive to the orientation in which it is used, as the weight of the magnet itself will deflect the beam more or less, depending on the orientation. Electronic compensation for this effect can be provided, for example, by incorporating low-cost MEMS accelerometers into the sensing device. Magnetic Flux Density Measurement This second method gives a more direct measure of permeability by way of magnetic flux density measurement and is not sensitive to the orientation of the sensor. In this method, the hand held sensing device

4

incorporates a magnet to apply the external field and a Hall effect sensor to measure flux density. The device is arranged so as to hold the magnet several centimeters above the test surface, when the device is placed in contact with the pipe. The Hall effect sensor is positioned so as to be very close to the metal surface. This arrangement puts an air gap of several centimeters between the sensor and the physical pole of the magnet. The field of the magnet in free space diverges (and diminishes in strength) with increasing distance from the pole. However, bringing the magnet near a concentration of iron will concentrate flux at the iron surface and the field strength will not fall off with distance as it does in free space. Thus, the measured flux density will be greater at an iron surface than in free space. Due to the difference in permeability between graphitized and intact gray iron, a lower flux density will be observed at the surface of a graphitized area than at an intact iron surface. Further characterization of the distribution of graphitized metal could be accomplished with an array of Hall effect sensors at the metal surface. With several such sensors laid out in a pattern, for example a grid or concentric rings, the gradient of magnetic field strength could be measured across the surface. As there would, consistently, be a gradient of increasing field strength from damaged to intact regions, such an array could aid in the mapping of graphitized areas.

WORK TO DATE A proof of concept has been achieved using a novel graphitization calibration standard. This standard, shown in Figure 3, consists of a plate of gray iron, machined with rectangular and spherical voids of varying depth, and inlaid with graphite inserts. A prototype of the hand held magnetic flux density sensor was developed, and is shown in Figure 4. The prototype was cut from hardwood by a CNC router and incorporated a semiconductor Hall effect sensor from Allegro Microsystems Inc. and a cylindrical neodymium-iron-boron permanent magnet. This prototype successfully and consistently distinguished varying depths of graphite inserts in the gray iron calibration standard. The depth of the graphite inserts was measurable both from the front (side with exposed graphite) and rear (unmodified gray iron surface) of the calibration standard. This demonstrates the efficacy of the magnetic flux sensing method for detecting internal as well as external graphitization corrosion. The graph given in Figure 5 shows raw data taken from the prototype sensor when applied to increasingly deep graphite inserts in the calibration standard. The data points are the output from the Hall effect sensor, measured in volts. The depth of the graphite inserts is measured in mils.

POTENTIAL EMBODIMENTS To date, work has focused on portable, hand held sensing units. Such a sensor would be manually applied, by a technician at an excavation site, to pipe under investigation. The sensing device would be attached to, or could dock with, a computerized data logging and storage system to provide analysis either on-site or after the fact. Due to the small size and high strength of rare earth permanent magnets and the small size and low power requirements of semiconductor Hall effect sensors, a version of the device could be developed for remote sensing applications and could be used in the internal inspection of pipelines. Such a device could present advantages as compared to a hand held unit. Internal inspection of a length of pipe requires at most one small excavation, whereas the entire length of pipe to be surveyed would otherwise have to be exposed. The sensor could be interfaced to existing apparatus for internal inspection of pipelines.

CONCLUSION We have demonstrated the viability of a new and robust sensing device for the detection of graphitization corrosion in gray iron. Such a sensing device will be of great value to municipalities with aging gray iron water distribution systems as it will provide them, for the first time, with a means of detecting graphitized pipe before

5

failure. The knowledge gained will enable municipalities to replace compromised pipe in a scheduled and controlled fashion, avoiding the inconvenience, lost revenue, and expense incurred by repairs performed after failure occurs.

REFERENCES 1. D.E. Krause, “Gray Iron – A Unique Engineering Material,” in Gray, Ductile and Malleable Iron Castings –

Current Capabilities, ASTM STP 455 (Philadelphia, Pennsylvania, ASTM: 1969), p. 3. 2. R.S. Elliott, Electromagnetics: History, Theory and Applications, (New York, New York: The Institute of

Electrical and Electronics Engineers, 1993), p. 412. 3. M. Zamanzadeh, E. Larkin, W. Gretz and B. Bavarian, “Case Histories of Failures in Water Mains,”

Corrosion/90, paper no. 389 (Houston, Texas: NACE, 1990).

6

Figure 1 – Full cross section photograph of a graphitized pipe.

Figure 2 – Close up photograph of the cross section of a graphitized pipe.

7

Figure 3 - Graphitization calibration standard.

Figure 4 - Prototype of the hand held graphitization sensor; implementing magnetic flux density measurement; measuring the graphitization calibration standard.

8

Figure 5 - Voltage output of the Allegro Hall sensor over graphite inserts of varying depth.

0 50 100 150 200 250 300-2.855

-2.85

-2.845

-2.84

-2.835

-2.83

-2.825

-2.82

-2.815

-2.81

-2.805

thickness of graphite layer, mils

-1*f

lux

sens

or o

utpu

t (V

olts

)

Flux - Based Measurement on Graphite Inset PlateHall-Effect Wand Sensor placed in contact with sample

9