Shedding Light on the Failure Factors of Subsea Critical Fastener Bolts

Offshore facilities, such as oil and gas rigs, wind turbines, and related subsea equipment, typically use flanges fastened using bolts and nuts as the main connectors. In this study, multidisciplinary parameters, namely the preload torque used to tighten bolts, simulated subsea water currents, water temperature, and impressed current cathodic protection, were applied to ASTM A193 B7 bolts. An experimental supervisory control and data acquisition system was designed to obtain measurements every 5 min throughout a 21-day experiment. Finite element analysis was performed to predict the structurally vulnerable areas of the bolts. A strong correlation was found between the reference electrode readings and the measured electrical current, tightening torque, and water temperature. As the water temperature rises during the day, the reference electrode reading becomes less negative and the electrical current decreases. Subsea water currents cause about a four-time increase in the bolt corrosion rate, with unprotected bolts suffering a nine-time-higher corrosion rate than protected bolts. A unique supply–demand interaction is observed; less protection is supplied to areas with lower corrosion rates (lower demand for protection). Finally, scanning electron microscopy examination reveals new insights into the failure mechanisms of subsea bolts.

More than a quarter of today's oil and gas supply is produced offshore.In recent years, the natural gas output from offshore fields has risen by more than 50%, and offshore electricity generation, mainly by wind power, has increased rapidly. 1 The European Union is moving toward an electricity system dominated by wind, with onshore and offshore sources together accounting for just over 50% of the total electricity generation by 2050. 2 Offshore oil and gas standards and best practices involve the use of offshore wind turbines, and relevant organizations are working on new standards for the coming years. 2 Bolt flange connections, a critical part of offshore operations, are vulnerable links in wind turbine power systems. 3Thousands of bolting components, such as threaded bolts, studs, and nuts, are used in a wide range of applications. 4Unexpected bolt failures have been occurring in the oil and gas industry for more than a decade and have occurred in offshore wind turbines in recent years. 5Offshore equipment, especially critical bolt connectors, encounter severe service conditions.Several parameters affect the corrosion process in subsea applications, such as bolt flange connections.The primary parameters are the sea conditions, including salinity, subsea water currents, and water temperature, and the different levels of torque used to tighten bolts.The sea water temperature is the main parameter that influences current density, thereby affecting the properties of the final deposited layer.A temperature rise can cause a sharp increase in current density at the moment of temperature change. 6The water current also needs to be considered, as it affects the corrosion rate.A strong water current can destroy passivation films and reveal new, bare steel, thus increasing the corrosion rate.The water current also controls the oxidation-diffusion level, which influences the corrosion rate of steel. 7The time of exposure to the environment with cathodic protection is critical. 8he torque applied to tighten bolts and nuts greatly affects the corrosion behavior of metal by changing its electrochemical behavior.The applied stress increases the anodic and cathodic reactions of the steel.In particular, the cathodic reaction rate under loads is considerably higher than that under no-load conditions.Pressure can help increase the Hydrogen evolution reaction (HER). 9mpressed current cathodic protection (ICCP), which is supposed to eliminate or at least delay bolt corrosion, is a major parameter affecting the service conditions of subsea equipment.Cathodic protection is a corrosion control system used to prevent corrosion of metal structures.Cathodic protection is implemented by the system by supplying electrons through current circulation between an electrode (anode) in contact with the environment and the configuration (cathode).Under subsea conditions, given data from a Pourbaix diagram reflecting a pH of 7.8-8.2and a voltage of approximately −1 V, the anodic reaction will be oxygen evolution and, in the presence of chlorides, chlorine evolution, whereas the cathodic reaction (on the configuration surface) will be hydrogen evolution. 10However, in the use of ICCP, overprotection must be avoided, as the hydrogen evolution on the metal surface may cause hydrogen embrittlement (HE). 11n particular, a certain combination of mutual influence of the abovementioned parameters causes the unexpected failure of subsea critical bolts.Therefore, a long-duration experiment should be conducted while monitoring the main variables such as: water temperature, water current, torque applied, and the ICCP system electrical current, to understand these mutual effects.
Health monitoring systems should be used because offshore applications are always far away from the mainland.Supervisory control and data acquisition (SCADA) systems are usually used to monitor several parameters related to offshore environments, such as the water temperature and current.The measured signals, recorded at a low sampling frequency (usually at 10 min intervals), can be quickly processed by many data analysis tools to detect correlations, clarify the causes of problems, and automatically provide operators and managers with real time online feedback. 12he various parameters of service conditions constantly and simultaneously affect one another and thus the processes that occur underwater.Moreover, because corrosion is slow and affected by several parameters, theoretically calculating the corrosion process is almost impossible. 13Only a database approach can elucidate the underwater corrosion process.

Materials and Methods
Threaded fasteners for offshore applications are manufactured according to various specifications, including API Spec 20E 14 and API Spec 20 F. 15 Each API specification details requirements for materials, properties, manufacturing, testing, inspection, and record keeping.
In this study, the main specimens were ASTM A193 16 B7 bolts with a chemical composition of AISI 4140, a diameter of 5/8 inches, z E-mail: amir@sce.ac.ilECS Advances, 2024 3 021501 and a length of 4 3/4 inches, as specified by API Spec 20E.ASTM A105 17 flanges were selected because they help rapidly simulate bolts when applying different torques.Based on the root area of the thread, the maximum allowable tensile load that the flange bolts could accommodate was approximately 83% of yield strength, in line with API Spec 6 A 18 and ISO 10,432. 19Usually, flange bolts must be preloaded to 67%-73% of the bolt yield stress.
The bolts were tightened at three torque magnitudes: 455 Nm (100% of G Y P . .), 341 Nm (75% of G Y P . .), and 114 Nm (25% of G Y P . .).These torque values, as percentages of G , Y P . .and a color illustration of the bolts tightened at different levels of torque are presented in Fig. 1.
The finite element analysis software SolidWorks Simulation was used to identify the model parts, such as the nuts, bolts, and flanges, and the level of torque reached in each bolt.This analysis showed that the most vulnerable area of the bolt was the area right below the nut, as shown in Fig. 2. 20,21 For the test, a SCADA system capable of collecting various measurement data was created.Data were obtained in real time and logged directly into a computer in Excel format.The ICCP system detailed in a previous article 22 was integrated into the SCADA system, and an ammeter was installed to register the ICCP current in addition to the collected data.The flanges and bolts were connected to the ICCP system at six different locations; in other words, six bolts were tightened at different preload torques.The reference electrode readings from the various locations in the configuration were used to determine the electrical current distribution across the configuration.
A water heater was used to control the water temperature, which was measured and recorded using a thermometer.The SCADA system was packaged into a palm-sized device containing all the necessary connections and cables.The package was printed using a 3D printer, and a screen was added; the interface with the user occurred in real time, and the data being measured were shown live on the screen, as depicted in Fig. 3.
After the bolts were fastened to the flanges according to the values presented in Fig. 1, the configurations were immersed for 21 days in an aquarium containing a 3.5% NaCl solution.The flanges and bolts were connected to the ICCP system at six different locations (six bolts tightened at different preload torques).Two bolts with the same tightening torque were connected to the reference electrode.Measurement was performed every 5 min, and the data acquired during the 21 days of the experiment were logged directly into a computer in Excel format in more than 6,000 rows.Each experiment was conducted several times, and the results were averaged.Analysis of the data collected from the experimental SCADA system revealed an interesting correlation between the water temperature and the reference electrode readings.Subsequently, the flanges were retrieved from the aquarium, as shown in Fig. 4, and the bolts were removed from the flanges.
The bolts were cleaned according to designation C3.3 (Table A1.1), ASTM G1 standard. 23The bolts were cleaned and then sliced in the vulnerable area, as shown in Fig. 5.Each sample surface was checked using an Ivium multichannel analyzer and software, which automatically calculated the corrosion rate of the samples.Finally, the bolts were subjected to microscopic evaluation via scanning electron microscopy (SEM).
Excellent corrosion behavior was defined as ⩽0.02 mm y −1 , whereas 0.02-0.1 mm y −1 was very good, 0.5-1 mm y −1 was average, and 1-5 mm y −1 was poor. 24The bolts were weighed to check the mass lost during the 21-day experiment in comparison   with their weights before service, allowing the calculation of their corrosion rates.

Results
The first set of results is related to the corrosion rates in different locations on the bolts and the effects of cathodic protection on the corrosion rate.These results are the average of bolts tightened at the same torque in several experiment rounds.Figure 6 displays the corrosion rates of the edges and vulnerable areas of the bolts tightened at different torques while the water pump (used to simulate subsea water currents) was turned on and off.Also shown are the increases in the corrosion rates against the three tightening torques.At the highest tightening torque, the vulnerable area with the water pump on demonstrates the highest corrosion rate (22 mm y −1 ), but the edge with the water pump on has a corrosion rate of only 11.175 mm y −1 .The corrosion rate of the vulnerable area is approximately twice (about 1.35 times) that of the edge when the water pump is turned on (off) at all three tightening torque levels.
With the baseline corrosion rate being 5.04 mm y −1 , which was obtained by immersing a bolt in the aquarium for 21 days without stress or the water pump, Fig. 7 reveals the influence of the different parameters on the corrosion rate.The corrosion rate increases with the tightening torque, regardless of the subsea current or the location of the evaluated area on the bolt.The water current (pump turned on) even further accelerates corrosion, with the vulnerable area exhibiting the highest corrosion rate at all three tightening torques.The combined effects of high subsea water currents and high torque in the vulnerable area enhance the corrosion rate by 436% (4.3 times) relative to the baseline, as shown in Fig. 7.
The torque level significantly affects the corrosion rate of the bolts, especially in their vulnerable area.Specifically, the corrosion rate increases with the tightening torque.This effect considerably influences the electrochemical behavior of steel, especially in the vulnerable area of the bolt.Pre-strain can enhance the anodic and cathodic reactions of steel in a simulated marine atmosphere. 25oreover, the torque level can cause stress concentration at the crack tip, especially in the vulnerable area of the bolt.This local stress concentration can induce plastic strain in the crack tip area and promote local anodic dissolution into the environment. 26In particular, the cathodic reaction rate under loads is considerably higher than that under no-load conditions, indicating that elastic tensile stress can contribute more to an increase in the HER than anodic steel dissolution. 27he effects of ICCP and subsea water currents on the corrosion rates of the bolts tightened at different torque levels can be isolated.The effect of the subsea water current on the corrosion rate of the   bolts is shown in Fig. 8.A correlation is observed between the tightening torque and the corrosion rate, even with the ICCP system applied.With the water pump turned off, the unprotected bolt tightened at the lowest torque (114 Nm, 25% of G Y P . . ) shows a corrosion rate of 0.102 mm y −1 , whereas the protected bolt tightened at the same torque exhibits a corrosion rate of 0.011 mm y −1 .With the water pump turned on, the protected bolt tightened at 114 Nm demonstrates a corrosion rate of 0.045 mm y −1 , whereas the unprotected one at the same tightening torque has a corrosion rate of 0.42 mm y −1 .As for the bolts tightened at the highest torque (451 Nm, 100% of G Y P . . ) and subjected to water currents, the protected bolt shows a corrosion rate of 0.091 mm y −1 (which is higher than those of the protected bolts at other tightening torques), whereas the unprotected bolt has a corrosion rate of 0.889 mm y −1 .The corrosion rates of the bolts in the presence of subsea water currents are approximately four times higher than those of the bolts not subjected to subsea water currents.Strong water currents affect the corrosion rate in several ways: by dismantling passivation films and revealing new, bare steel, thereby increasing the corrosion rate; by controlling the oxidation-diffusion level; and by potentially inducing an erosion-corrosion state. 28e effects of ICCP and water currents on the corrosion rate of the bolts tightened at different torque levels can be isolated.The corrosion rates in the presence of the ICCP system are more than nine times higher than those of the exposed bolts, depending on the torque levels, as shown in Fig. 9.The ICCP system protects bare metal (i.e., metal not covered by deposits).When water currents destroy passivation films and reveal new, bare steel surfaces, ICCP immediately covers these surfaces with an electron shield, thus preventing further corrosion.
The data measured and collected via the experimental SCADA system were analyzed comprehensively.During the 21 days of the test, the system collected more than 6,000 Excel sheet rows and 10 columns.The first correlation examined was that between the different tightening torques and the reference electrode readings.Data from the reference electrode readings at the three torque levels were plotted as shown in Fig. 10   [V] show different reference electrode readings, especially in the first 10 days.The readings for E6[V] are less negative than the readings of E3[V], and E1[V], or less protection to the lowest tightening torque.
After 10 days of tests, the water temperature was raised artificially to 28 °C, which was maintained until the end of the experiment.This was to check the correlation between the water temperature and the other ICCP parameters measured using the SCADA system.The temperature dimension data during the experiment are shown in Fig. 11.These temperatures were measured inside the aquarium during the three weeks of the experiment.The day's high temperature is marked 1 to 7 in the first week of the test for an easy understanding of the day-night temperature profile.Point 10 shows the starting day of the artificial water heating.
Integration of the reference electrode readings and temperature profile readings (Fig. 12) shows an interesting connection.This strong correlation was also described in a previous article. 22For the first 10 days, the water temperature profile ranges from the highest temperature at noon to the lowest at night.The reference electrode readings follow the same profile, increasing (less negative) until the highest temperature at noon and decreasing (more negative) until the lowest temperature at night.From the time the water temperature was increased artificially to 28 °C, the reference electrode readings follow the abovementioned path, rising (becoming less negative) and then stabilizing at the highest (least negative) values.
The ammeter readings I in and the reference electrode readings are plotted in Fig. 13.A strong correlation is observed; when the ICCP electrical current reduces, the reference electrode readings become less negative.Reasonably, when the current strength decreases, the reference electrode readings are less negative because the electrical circuit has fewer electrons, resulting in fewer electrons across the protected surface (i.e., less protection for the configuration).The last connection that needs to be checked is that between the water temperature and the electrical current I .
in When the temperature rises, the reference electrode reading becomes less negative and the electrical current reduces, so a logical analogy can be made: when the temperature rises, the electrical current drops.
The relationship between the water temperature and the electrical circuit current I in is presented in Fig. 14.A strong correlation is seen, especially after the water is heated; when the temperature rises, the ammeter readings decline.8. less negative readings reflect less cathodic protection.The protection provided by the ICCP system is uniform; in other words, the same protection (electrons) is supplied to the protected configuration surface.However, a relationship between the lowest corrosion rate (or the lowest tightening torque) and the reference electrode values is seen.The lowest corrosion rate bolt receives the lowest protection, as inferred from the reference electrode readings.In other words, the demand meets the supply: less protection is supplied to areas with lower corrosion rates (i.e., less protection demand).After 10 days, the water temperature increases artificially, and the reference electrode readings follow the temperature trend; The readings become less negative with no significant difference between the readings of the bolts tightened at different torques levels, as shown in Fig. 13.

Discussion
A strong connection is seen between the three parameters: the reference electrode readings (E ref ), the electrical current (I in ), and the temperature (T) measured in the aquarium.When the water temperature rises during the day, the reference electrode readings become less negative and the electrical current decreases.ICCP efficiency is related to the protection current density; as the current density increases, the corrosion rate of steel decreases. 29Current density at higher temperatures is lower than that at lower temperatures.The oxygen reduction reaction decreases according to the growth of calcareous deposits on a specimen surface. 6The sea water temperature is the main parameter that influences current density and accordingly affects the properties of the final deposited layer. 30A higher temperature increases both anodic and cathodic reactions. 31mperature strongly affects water resistivity; as temperature increases, water resistivity decreases. 32Therefore, the main effect of water heating is a change in water conductivity.This will boost the consumption of the electrons protecting the metal surface.In particular, the absolute electron count on the metal surface will decrease (limited by the applied current), and the reference electrode reading will be less negative.
When the cathodic protection system was activated, gas bubbles began forming on the configuration surface.The gas bubbles that accumulated across the flange and the bolt are seen in Fig. 15.The gas bubbles began forming immediately after the activation of the cathodic protection system.In experiments without cathodic protection systems, such gas bubbles are not formed at all and therefore, cannot be seen.According to a Pourbaix diagram of water and iron, 10 considering the potential and the pH of the environment, the resulting cathodic reaction can be an HER, such as a H 2 molecule or a H atom appearing on the metal surface.Given that the abovementioned gas bubbles may be hydrogen atoms or molecules, they may indicate HE.This point is meaningful to industries, as HE can cause unexpected failure.
Figures 16 and 17 present a comparison of the protected and unprotected bolts.The morphologies of the cracks significantly differ between the different bolts.Unlike the protected bolts, the unprotected bolts exhibit high corrosion rates and corrosion cracks  that resemble corrosion stresses.The crack depth in the protected bolts appears to be smaller, but an in-depth examination reveals HE morphological cracks.
The effect of water temperature is related to the intensity of the chemical reactions.Water reduction (HER) is greater at higher temperatures. 33This elevates H 2 molecules or atoms on the metal surface.Combined with the presence of stress (torque) and susceptible steel, hydrogen atoms create the conditions for HE.The crack morphology [Error!Reference source not found.]and gas bubbles suggest that the protected bolts suffer from HE and the unprotected bolts suffer from stress corrosion cracking.Although the corrosion rates of the protected bolts are extremely low relative to those of the unprotected ones, HE cracks are far more dangerous and can cause unexpected failure.

Conclusions
The corrosion process occurring in subsea facilities is complex.Multidomain parameters are involved, and the relationships between these parameters are not fully understood.These relationships can be elucidated by simulating the terms of service of a subsea facility, such as a critical connector, and performing data collection and analysis.
• The wide range of distribution of the applied force across the bolt causes various corrosion rates in different locations.The corrosion rate in the vulnerable area is, on average, four times that at the edge.
• Subsea water currents, simulated in this study using aquarium water pumps, have a crucial influence on the corrosion rate of subsea critical connectors.Such currents enhance the corrosion rate by approximately four times.
• The corrosion rates of the bolts protected by the ICCP system are extremely low compared with those of the exposed bolts.The unprotected bolts have higher nighttime corrosion rates than the protected bolts.Thus, the ICCP system slows down the corrosion process.
• A relationship is seen between the reference electrode values and the tightening torque.The bolt with the lowest torque applied (or lowest corrosion rate) gains the lowest (least negative) reference electrode readings (i.e., the least protection), whereas the bolt with the highest corrosion rate obtained high (more negative) reference electrode readings (i.e., high protection).
• A strong correlation is observed between the three parameters: reference electrode readings, electrical current, and water temperature.This relationship is evident when the water is heated to 28 °C and kept at this elevated temperature.As the temperature increases, the reference electrode readings become less negative (i.e., less protection is applied) while the electrical current declines.
• ICCP activation produces hydrogen molecules and atoms that are absorbed by the metal via small corrosion cracks.HE cracks are seen in the bolts tightened at high torques.Ofer's curiosity, passion, and keen interest led him to delve into this field.He spoke about it with great fervor, his eyes sparkled with enthusiasm, imbuing him with a sense of liveliness and essentiality.
Ofer's desire to address an issue of profound ecological damage shines through in this paper, illuminating practical methods for understanding the mechanism of complex corrosion and thereby resolving the issue to prevent future occurrences.With Ofer's passing prematurely, we remain assured of his desire for ongoing research and integration of this subject within the industry.
This research stands as a testament to Ofer's enduring legacy, reminding us of the boundless possibilities that await those who dare to explore.

Figure 1 .
Figure 1.(A) Torque values as percentages of G ; Y P . .(B) color illustration of bolts tightened at different torque levels.

Figure 2 .
Figure 2. Vulnerable area of bolt, identified via finite element analysis.

Figure 4 .
Figure 4. (A) SCADA experimental system at work; (B) flanges and bolts after removal from aquarium.

Figure 5 .
Figure 5. Slice in vulnerable part of bolt.

Figure 6 .
Figure 6.Corrosion rates in different locations on bolts.

Figure 8 .
Figure 8.Effect of water current on corrosion rate of bolts tightened at different torque levels.

Figure 9 .
Figure 9.Effect of water currents on corrosion rate of bolts tightened at different torque levels.

Figure 10 .
Figure 10.Reference electrode readings for three bolts.

Figure 12 .
Figure 12.Reference electrode and temperature profile readings.

Figure 10
Figure 10 indicates that the same readings can be discerned between the bolts tightened at 341 and 455 Nm (75% and 100% of G ; Y P . .yellow and red, respectively) and the reference electrode.Bolt no.6 E6[V], tightened at 114 Nm (25% of G ; Y P . .blue) demonstrates different recorded data in the first 10 days of the test (before the artificial water heating).The values of the reference electrode are less negative for the 114 Nm bolt than those for the 341 and 455 Nm bolts, whereas the corrosion rates are 0.045 mm/y for the 114 Nm bolt and 0.0709 and 0.0907 mm/y for the 341 and 455 Nm bolts, respectively, as shown in Fig.8.less negative readings reflect less cathodic protection.The protection provided by the ICCP system is uniform; in other words, the same protection (electrons) is supplied to the protected configuration surface.However, a relationship between the lowest corrosion rate (or the lowest tightening torque) and the reference electrode values is seen.The lowest corrosion rate bolt receives the lowest protection, as inferred from the reference electrode readings.In other words, the demand meets the supply: less protection is supplied to areas with lower corrosion rates (i.e., less protection demand).After 10 days, the water temperature increases artificially, and the reference electrode readings follow the temperature trend; The readings become less negative with no significant difference between the readings of the bolts tightened at different torques levels, as shown in Fig.13.A strong connection is seen between the three parameters: the reference electrode readings (E ref ), the electrical current (I in ), and the temperature (T) measured in the aquarium.When the water temperature rises during the day, the reference electrode readings become less negative and the electrical current decreases.ICCP efficiency is related to the protection current density; as the current density increases, the corrosion rate of steel decreases.29Current density at higher temperatures is lower than that at lower temperatures.The oxygen reduction reaction decreases according to the growth of calcareous deposits on a specimen surface.6The sea water temperature is the main parameter that influences current density and accordingly affects the properties of the final deposited layer.30A higher temperature increases both anodic and cathodic reactions.31

Figure 14 .
Figure 14.Water temperature and electrical circuit current I in .

Figure 15 .
Figure 15.Bubbles accumulated on flange surface during experiment.