Go To On-Line Confrence

Corrosion Detection of Mild Steel in a Two Phase Hydrocarbon Electrolyte System Using Electrochemical Noise

J.M. Malo and O. Corona
Instituto de Investigaciones Electricas
Reforma 113, Col. Palmira
email: "Jose M. Malo"

Electrochemical current noise (ECN) measurements were carried out in mixtures of 3% NaCl electrolyte in diesel under stirred conditions (0 to 2000 rpm) using a Rotating Electrodes System (RES) which includes three mild steel concentric electrodes embedded in activated polyester resin. Chemical activation of the resin allowed electrochemical measurements in the water in oil emulsion system. A 0.2 to 15 % in volume range of 3% NaCl electrolyte additions was studied. Three distinctive noise patterns were obtained from electrochemical current noise (ECN) time-series: a) a low noise baseline for diesel in absence of electrolyte, diesel with small additions of electrolyte and/or low flow rates, b) a low noise signal with current bursts superimposed obtained from relatively small additions of electrolyte and high rotation rates and c) a high amplitude signal for high rotation rates and relatively high additions of electrolyte. Tests using a conventional mild steel RDE allowed visual interpretation of the three type of signals corresponding to absence, initiation and propagation of corrosion. For case b, the number and intensity of current bursts is indicative of proximity to cases a or c. These results contrast with experiments carried out with a conventional non-activated resin which is insensitive to the range of electrolyte additions or to stirring conditions. This method can be implemented for water in oil systems where early corrosion detection is desirable.

Introduction

The corrosion of steels in oil/water mixtures is a key problem in the oil industry [1,2]. Depending on the water cut for oil/water systems three corrosion rate behavior regions can be identified [3]. In the case of low water cuts, it has been shown that although water is entrained in the hydrocarbon phase corrosion can occur, which is attributed to wettability by some authors [4] while for others the explanation lies in conductivity of the mixture[5]. In addition, for mixtures of fluids often encountered in oil and as production equipment, local corrosion rates depend not only in the corrosivities of the two phases, but also on their flow pattern. Flow patterns can take a variety of forms such as of bubble, slug, churn or annular flow.

Attempts have been made to improve corrosion evaluation in systems where the hydrocarbon is the continuous phase and in which electrochemical techniques have limitation due to the medium high resistivity. It has been shown that an electrode probe can be developed that allows the use of linear polarization resistance methods to study corrosion in high solution resistance environments by chemically activating the dielectric surface where the electrodes are embedded [6,7]. In a hydrocarbon/water mixture the corrosion probe functions equally well if either the hydrocarbon or water is the continuous phase. In this work the application of activated probes for the corrosion detection in hydrocarbon/water mixtures is extended by examining its corrosion detection behavior using electrochemical current noise technique.

Experimental Method

Test materials

Tests were carried out for mild steel samples mounted in polyester resin, as part of a modified rotating disc setup. Instead of single working electrode three concentric samples are exposed on the rotating surface as shown in Figure 1. This configuration is a modification of the RDE, allowing simultaneous potential and current noise measurements between identical electrodes on a rotating surface.

A 300 ml volume of diesel was used as the basic test solution to which additions of 3% NaCl were made in the 0.2 to 15% vol. range.

Rotating Electrodes System (RES)

Flow tests were conducted by a modified rotating disk electrode system incorporating three isolated electrodes on the exposed surface. The electrical connection of the rotating to each of the fixed elements of the RES was made by using three carbon/silver (graphalloy) brushes, Figure 1, each in contact with a silver ring to ensure electrical continuity and a low electrical noise level. Therefore, the RES had three electrical connection, one for each electrode. The rotation speed was varied in the 0-2000 rpm range.

Probe Activation

Activation of the surface probe was attained by immersion in 10M NaOH at 80 oC for 2 hr[ ]. To remove corrosion products from activation stage the probe is cleaned by HCl 15 % at 60 oC. Resistivity between electrodes after activation is about 250 000 ohm.

Electrochemical Noise Tests

ECN measurements were recorded using ACM software and instrumentation: AutoZRA (with maximum resolution of 100 pA) controlled by a personal computer. ECN measurements were obtained at 1 reading/s to produce records of 512 points. Data processing included trend removal of the signal by minimum square linear fitting.

Figure 1. (a) Cross section of electrode mounting for the rotating system containing three electrodes and (b) electrical connection between fixed and rotating elements.

Results and Discussion

The Rotating Disk Electrode features are describe in the literature [8] however its scope of application is continuous aqueous media. In this work its application is extended to a two liquid phase system where the flow regime is controlled by the rotation speed of the RES. In the absence of stirring (rpm=0) the water phase is separated from the hydrocarbon phase and stays at the bottom of the cell as shown in Figure 2a. The probe is in contact with the hydrocarbon phase. As stirring is introduced, the electrolyte can be incorporated into the hydrocarbon phase, depending on the degree of stirring and the electrolyte content. Therefore, if the probe gets in contact with the aqueous phase corrosion can take place to some degree. Due to the hydrodynamics of an RDE based system, a fluid suction effect makes fluid ascend axially towards the surface of the electrode, on approach flows takes a radial flow combined with a centrifuge component [8]. In this way, for a low level of stirring, e.g. 150 rpm, Figure 2b, a meniscus formation occurs at the electrolyte /hydrocarbon interface, with no apparent mixing of the two phases.

For moderate stirring conditions, e.g. 500 rpm, the meniscus elongates into a bell-shaped interface which can yield water droplets that are incorporated into the hydrocarbon phase, Figure 2c. Under this condition, the probe can become in contact with the electrolyte through droplets collisions. From corrosion standpoint, water phase is the carrier of aggressive species which can produce metallic corrosion if wetting of the probe surface is attained.

For high stirring conditions, e.g. 1000 rpm, shear stress at the water/hydrocarbon interface increases and that electrolyte can come into view as a disperse phase in the hydrocarbon, Figure 2d. However, due to the more severe flow conditions, the droplet size is smaller but in higher in number than in the previous case, with the prospect of a high number of droplets impacts against the probe surface.

In this way, it is possible to simulate a wide variety of flow conditions depending on the intensity of forced convection. This is the first variable examined in this work in the 0-1500 rpm range. In addition, there is the effect of the water/oil ratio of the fluid system. For this study results are analyzed where the hydrocarbon is the continuous phase varying the electrolyte concentration in the 0 to 15 % vol. rate.

Figure 2. Flow patterns induced by different stirring rates (a) 0, (b) 150 , (c) 500 and (d) 1000 rpm.

Electrochemical Noise for an Activated Probe at 0 rpm in a Hydrocarbon Phase

Figure 3 shows the time series of the electrochemical current noise signal for an activated probe in hydrocarbon phase with no stirring. A uniform amplitude band of current fluctuations of 3x10-6 mA/cm2 size is obtained. This is the background noise for the probe representing its behavior in an electrically insulating media. This works as a reference signal pattern for comparison with other systems under study where resistivity decreases in the electrolyte presence.

Figure 3. Electrochemical current noise at 0 rpm for steel in diesel. (Size of current axis is 0.0009 mA/cm2)

Electrochemical Noise for an Activated Probe at 150 rpm

For a low stirring rate, 150 rpm, even for the higher electrolyte addition, 15% NaCl, the electrolyte phase stays in the bottom of the cell. Therefore, despite the stirring of the RES probe is in contact with the hydrocarbon phase. Figure 4 shows ECN signals obtained for the range of electrolyte additions. All electrochemical noise patterns obtained are similar to that obtained in absence of electrolyte at rpm=0, figure 4a. Standard deviation is for the signals in the range of 5.9 to 8.3 x10-6 mA/cm2.

Figure 4. Electrochemical current noise at 150 rpm for steel in diesel with additions of (a) 0%, (b) 0.2%, (c) 2%, (d) 5%, (e) 10% and (f) 15% vol. 3% of NaCl. (Size of current axis is 0.0009 mA/cm2)

Electrochemical Noise for an Activated Probe at 500 rpm

At a higher stirring rate stirring, 500 rpm, electrochemical current noise signals from the activated probe are similar to those at 150 rpm in the range of electrolyte addition from 0.2% to 10 %, as shown in Figure 5. This is explained by the absence of wetting of the probe by the electrolyte which stays as separate phase. For the 15% electrolyte addition the signal takes a different pattern, Figure 6, for which current bursts, in both current directions, appear along the time-series. Current fluctuations size can be up to 0.01 mA/cm2. This signal corresponds to a rupture of the bell shaped water-hydrocarbon interface, Figure 2c, which produces the relatively big size electrolyte bubbles formation in the hydrocarbon face, resembling a slug flow pattern found in pipelines. It can be assumed that sporadic current increases results from the contact of water bubbles with the probe surface which temporarily wet the probe surface.

Figure 5. Electrochemical current noise at 500 rpm for steel in diesel with additions of (a) 0%, (b) 0.2%, (c) 2%, (d) 5%, and (e) 10% vol. of 3% NaCl. (Size of current axis is 0.009 mA/cm2)

Figure 6. Electrochemical current noise at 500 rpm for steel in diesel with addition of 15% vol. of 3% NaCl. (Size of current axis is 0.03 mA/cm2)

Electrochemical Noise for an Activated Probe at 1000 rpm

For 1000 rpm, the low amplitude fluctuation signal is only obtained for the lowest electrolyte addition, 0.2 %, as shown in Figure 7. As the electrolyte addition increase from 2% to 15%, Figure 7-9, the current pattern is similar to the 500 rpm test with 15 % electrolyte content, presenting current bursts the number of which is related to the water content. As the water content is increased the high intensity current fluctuations take place at shorter intervals from 4 spikes at 2% to about 30 current spikes at 15% water content. As previously explained, this behavior is originated by the increase in water bubbles in the hydrocarbon as the water content is increased therefore increasing the probability of water bubbles impacts on the probe surface.

Figure 7. Electrochemical current noise at 1000 rpm for steel in diesel with additions of (a) 0%, (b) 0.2% and (c) 2% vol. of 3% NaCl. (Size of current axis is 0.009 ma/cm2)

Figure 8. Electrochemical current noise at 1000 rpm for steel in diesel with additions of (a) 5% and (b) 10% vol. 3% of Nalco.. (Size of current axis is 0.009 Ma/cm2).

Figure 9. Electrochemical current noise at 1 000 rpm for steel in diesel with additions of (a) 10% and (b) 15% vol. 3% of Nalco. (Size of current axis is 0.018 Ma/cm2)

Electrochemical Noise for an Activated Probe at 1 500 rpm

Results for the 1500 rpm stirring rate is shown for the 0.2% electrolyte addition in Figure 10. It can be noticed that even for this low water content, stirring is sufficiently intense to entrain water bubbles that produce the electrochemical current noise pattern related to from water contact with the probe.

Figure 10. Electrochemical current noise at 1500 rpm for steel in diesel with additions of (a) 0%, (b) and 0.2% vol. 3% of Nalco. (Size of current axis is 0.0018 Ma/cm2)

For an electrolyte additions from 2% to 5% an additional signal pattern emerges as shown in Figure 11. Here, a transition in the current signal occurs as a result of increased wetting of the probe surface where a general increase in the amplitude of current fluctuations amplitude take place.

Figure 11. Electrochemical current noise at 1 500 rpm for steel in diesel with additions of (a) 2% and (b) 5% vol. 3% of NaCl. (Size of current axis is 0.018 mA/cm2)

For 10 and 15 % electrolyte additions the same pattern obtained at 5% is maintained as shown in Figure 12. The two current noise signals are very similar displaying a narrow band of current fluctuations within 0.002 mA/cm2.

Figure 12. Electrochemical current noise at 1 500 rpm for steel in diesel with additions of (a) 10% and (b) 15% vol. 3% of NaCl. (Size of current axis is 0.018 mA/cm2)

Electrochemical Noise for an Activated Probe at 2000 rpm

A pattern with current spikes results for the higher stirring rate studied and the lower electrolyte addition, Figure 13b. This result is indicative of the activated probe sensitivity. It can be assumed that due to the high stirring the electrolyte is fully incorporated as a separate phase in the hydrocarbon phase, allowing, despite the low water content, the surface wetting.

Figure 13. Electrochemical current noise at 2 000 rpm for steel in diesel with additions of (a) 0% and (b) 0.2% vol. 3% of NaCl. (Size of current axis is 0.0015 mA/cm2)

Figures 14-16 correspond to higher electrolyte additions, from 2% to 15 %. The curves for these cases belong to the same pattern of current fluctuations which in previous results appeared at high stirring rates and relatively high water additions. A subtle reduction in the slow amplitude variations of the current level take place as the water content is increased.

Figure 14. Electrochemical current noise at 2 000 rpm for steel in diesel with addition of 2% vol. 3% of NaCl. (Size of current axis is 0.0045 mA/cm2)

Figure 15. Electrochemical current noise at 2 000 rpm for steel in diesel with addition of 5% vol. 3% of NaCl. (Size of current axis is 0.00225 mA/cm2)

Figure 16. Electrochemical current noise at 2 000 rpm for steel in diesel with additions of (a) 10% and (b) 15% vol. 3% of NaCl. (Size of current axis is 0.0045 mA/cm2)

Flow Pattern Classification

Considering the combined effect of flow rate and water content results can be classified in various ways, according to their standard deviation for instance. However, in this work the current pattern is proposed as the most important attribute of the signals studied. Due the nature of the activated probe where levels of activation can vary from case to case, the standard deviation, a measure of current fluctuations about the media, does not necessarily lead to the best conclusions. Alternatively, the form of the signal follows a well defined behavior as water content and mixing of the two phases are varied. In fact, a signal with current bursts is the first indication of the presence of corrosion conditions in a hydrocarbon/water mixture. In addition, the number of current burst gives a relative measure of how intense wetting is. At one extreme case current bursts are so close in the time scale that that pattern tuns into a uniform amplitude signal indicting a more extensive corrosion level, associated to a general probe surface wetting. Three patterns can be identified the classification of which is summarized in Figure 17. These classification can work as a guideline for corrosion monitoring purposes where intensity of corrosion is related to a signal pattern. Artificial intelligence, particularly neural networks, are suitable analysis tools for making the pattern identification part of an automatic process where evaluation and alarms can be implemented based of the electrochemical noise pattern signals.

Figure 17. Classification of electrochemical noise patterns for 3% NaCl additions and rotation speed of ERS.

Results with a conventional rotating disc electrode confirmed results obtained with the activated probe. That is the case for the 2000 rpm and 0.2 % water addition for instance, where uniformly distributed corrosion spots are formed on the disk surface for an 8 hr immersion test, in contrast with the lack of surface attack for the a similar test with a 100 % hydrocarbon phase.

Additional work is necessary to validate these results in terms of corrosion inhibition phenomena where the effectiveness of this approach to corrosion protection should modify the current noise accordingly reducing the intensity of current fluctuations.

Conclusions

Results obtained show that an activated probe in combination with the electrochemical current noise technique can be effectively used for corrosion detection of hydrocarbon/water(<15% v.) mixtures. Depending on the water content and the stirring rate of the mixture three electrochemical current noise patterns are obtained which correspond to the intensity of corrosion activity. These patterns can serve as an tool to assess the level of corrosion attack in corrosion monitoring activities.

References

1. J. P. Brill, Multiphase flow in wells, Journal of Petroleum Technology, January, 15(1987).

2. J. S. Smart and G. L. Smith, Pigging and chemical treatment of pipelines, Pipes and Pipelines International, May June, 19(1991).

3. U. Lotz, L. van Bodegom and C. Ouwehand, Corrosion/90, paper no. 41 (Houston, Tx, NACE, 1990).

4. K.D. Efird, Materials Performance, 30, No. 3, 63(1991).

5. B. D. Craig, Corrosion in Oil/Water Systems, Materials Performance, No. 8., 35, 61(1996).

6. R. Jasinski and K. Efird, Electrochemical Corrosion Measurements in Crude Oil, Corrosion, No. 8, 43, 476 (1987).

7. R. Jasinski and K. Efird, Electrochemical Corrosion Probe for High Resistivity Hydrocarbon/Water Mixtures, Corrosion, 44, No. 9, 659(1988).

8. A. Bard, Electrochemical Methods, John Wiley and Sons, New York, 1979.