Life Cycle Monitoring of Offshore Steel Pipe Piles via UWFBG Wireless Sensor Network

. Monitoring the stress and deformation state of ofshore steel pipe piles is very necessary for complex ocean environments. However, challenges still remain in remotely capturing detailed data on ofshore projects. Herein, an ultraweak fber Bragg gratings-(UWFBGs-) based wireless sensor network was established to obtain the high-precision strain along the piles. Te No. 3 ofshore wind power safety monitoring project in South Shandong Peninsula, China, used such a fber-optic monitoring system. From January 1 to June 1, 2022, strain profles of three ofshore steel pipe piles were measured in seconds with a 1m spatial resolution and a 1 με strain resolution, which clearly show sea level and mud-water interfaces. To assess the pile’s operational condition, the displacement and rotational angle of the mud surface are further calculated. It is discovered that the horizontal deformation at sea has a positive correlation with the wind, waves, and current, meanwhile decreasing with depth. Te UWFBG-wireless sensor network can realize dynamic monitoring and early warning, which is promising for the life cycle monitoring of ofshore wind turbine steel pipe piles.


Introduction
In recent years, with the continuous growth of the wind energy market, fewer and fewer wind resources are available for development on the land. Global wind farm construction has shown a trend toward ofshore development [1][2][3]. A steel pipe pile is one of the most common forms of ofshore fan foundation, especially a large-diameter ultralong steel pipe pile which has been applied more and more widely [4]. Ofshore steel pipe piles are afected by many factors during their working life cycle, including piling, static load, and operation [5]. Teir stress and deformation states are complex and directly afect the construction and operation safety of the superstructure. Terefore, it is extremely necessary to monitor the whole life cycle of ofshore steel pipe piles [6,7].
Afected by marine hydrological and meteorological conditions, the construction window period of ofshore wind power pile foundations is short and exists in great uncertainty. During the operation period of ofshore wind power plants, it is not possible to rely solely on manual onsite inspection, and remote automated testing technology must be used to realize the monitoring [8]. Moreover, the fan foundation with a relatively long pile body is buried in the seawater, as well as the rock and soil mass of the seabed; thus, the environment around has strong corrosion and electrical conductivity [9]. Traditional monitoring techniques like vibrating-string, inclinometer, and resistive sensors require a lot of sensors and signal transmission lines [10]. It is challenging to meet the health monitoring requirements for the entire life cycle of ofshore wind power foundations due to issues like poor durability, low survival rate, poor stability, poor real-time performance, and low degree of automation.
Fiber optic sensing has attracted considerable attention in recent years because of its compactness, electromagnetic feld immunity, passive operation, multiplexing capacities, chemical stability, biocompatibility, etc. [11]. Diferent applications of fber optic sensors have been proposed, including industry [12], structural health monitoring [13], biochemistry [14], and rehabilitation [11,15]. Tis has led to the development of a large number of optical fber sensors covering physical and chemical parameters for measuring strain [16], temperature [17], pressure [18], pH value [19], and so on. Silica optical fber (SOF) and plastic optical fber (POF) are the most concerned and widely used two types of optical fbers [20]. SOF is suitable for long-distance measurement because of its high bandwidth and low attenuation [21], while POF is widely applied to short-distance measurement owing to its large core diameter, good fexibility, lightweight, low elastic modulus, and low cost [22].
Distributed strain sensing technology (DSS) adopts a strain-sensing fber-optic cable to capture the strain along the entire cable length, which is suitable for long-distance monitoring and can be applied to quantify the infrastructure's deformation [23,24]. Te feasibility of DSS for pile monitoring has been illustrated in previous studies [25][26][27][28]. However, most of the DSS techniques can not realize remote transmission and require manual sampling. Te fxed and low sampling frequencies are not conducive to exploring the dynamic efect of seawater on the ofshore steel pipe pile [29]. Fiber Bragg grating (FBG) technology is attractive for pile foundation structural health monitoring due to its low cost and remote sensing capability [30]. Timedivision multiplexing (TDM) technology and wavelengthdivision multiplexing (WDM) technology are two main multiplexing technologies for expanding the capacity of sensor networks [31]. However, due to signal crosstalk between these normal FBGs, the multiplexing capability is severely limited [32]. In addition, due to a large amount of redundant data in the raster-free area, the system has to adopt the way of low-speed inquiry to obtain the data with higher accuracy, which limits the data observation and makes it impossible to conduct long-distance real-time monitoring of ofshore steel pipe piles [33].
UWFBG is a kind of special fber Bragg grating with very weak refectivity, usually lower than −20 dB [34]. For this reason, the capacity of Bragg gratings in one fber-optic cable can be improved to thousands of Bragg gratings [35]. When connecting massive UWFBGs in series, all UWFBGs can be precisely localized based on TDM technology, because the refection time of the incident light is diferent for each UWFBG position [36]. UWFBGs efectively increase the sensing distance and multiplexing times of fber Bragg grating sensors, improves the monitoring accuracy [37], and makes it possible for remote real-time DSS of ofshore steel pipe piles.
Herein, a DSS network using UWFBGs was proposed for remote monitoring of ofshore steel pipe piles. Such a measurement system was designed detailed and applied to the No. 3 ofshore wind power safety monitoring project of the South Shandong Peninsula in China. Along with the ofshore static load test, a full-scale measurement of pile strain was conducted by a UWFBG-based sensing system, and the bearing behavior of piles was well analyzed. During the operation period, dynamic and long-term monitoring was carried out for a further understanding of the ofshore steel pipe pile working state. Tis research shows that the UWFBG measurements are the potential for the life cycle monitoring of ofshore steel pipe piles.

Methodology
2.1. Basic Principle of the UWFBG Technique. FBGs can be obtained by creating periodic variations in the refractive index of the core of an optical fber. When light is made to pass through the grating, the light refected by the varying zones of refractive indices will be in phase and amplifed at a particular wavelength, which is called the Bragg wavelength. Axial stress and temperature changes are the two factors that will cause the drift of Bragg wavelength, which satisfes the following linear relationship [38]: where Δλ B is the change of the FBG central wavelength, ε is the axial strain applied to the optical fber, ΔT is the temperature change, and K ε and K T are the wavelength sensitivity coefcients of strain and temperature, respectively, which can be calibrated in the laboratory. By rearranging equation (1), the axial strain can be obtained: Beneft from the weak refectivity, numbers of UWFBGs connected in series with the same refractive index and wavelength (λ 1 � λ 2 � ••• � λ n ) can be multiplexed in one fber core ( Figure 1). Te received wavelength shifts are used to interpret the strain changes and the refected time (equation (3)) is used to localize the sensing points [39]: where t is the time interval between receiving two UWFBG wavelengths in the fber, c is the light velocity in the fber optic, d is the interval distance of UWFBGs, and n ef is the efective refractive index.

Pile Stress and Force Calculation Based on UWFBGs.
Te test pile can be divided into a series of monitoring elements because of the small sampling intervals of the UWFBG-based sensing system. As long as the pile is installed with sensing cables correctly, the strain distribution of the pile can be calculated using equation (2) based on the measured Bragg wavelength shifts, and the steel pipe pile axial force F i can be further calculated [40]: where F i denotes the axial force of the steel pipe pile with section i, ε i is the average strain of section i, E is the elastic modulus of the pile, and A i is the section area of section i.

Structural Control and Health Monitoring
Te vertical loads along the pile propagate downward, and the compression of the upper soil layer will cause reverse shaft friction resistance. According to equations (5) and (6), shaft friction resistance R i and pile tip resistance R b of the steel pile can be obtained: where l i denotes the length between section i and section i + 1, u denotes the outer perimeter, F i denotes the pile tip axial force, and A 0 denotes the pile end area of the steel pipe pile. Meanwhile, the pile compression S s and tip-settlement S b can be expressed as follows: where S is the top settlement of the pile, which can be measured by a dial indicator installed on the top of the pile. When under horizontal load, the steel pipe pile can be simplifed as a fxed cantilever beam structure. Te strain values ε 1 and ε 2 captured at the symmetrical position of the pile at the buried depth z, respectively, can be used to calculate the bending strain value ε m (z). Note that for the strain values, positive represents tension, while negative represents compression: Ten, the rotational angle θ (z), horizontal displacement ω (z), and bending moment M (z) of the pile can be calculated using the following equations: where I denotes the moment of inertia of the pile section and C 1 and C 2 are the boundary conditions to be calculated. Under simplifed conditions, the boundary conditions of steel pipe piles are z � 0, ω � 0; z � 0, and ω′ � θ � 0. Ten, the calculated C 1 � C 2 � 0.

UWFBG Cable and Its Installation. Te FO cable with
UWFBGs used in ofshore steel pipe pile monitoring is the NZS-DDS-C03 cable produced by Suzhou NanZee Sensing Ltd., China, and its mechanical properties are provided in Table 1. Te structure of the cable is shown in Figure 2. From inside to outside, the loose-bufered cable consists of a silica FO, UWFBG, bufer, metal-clad pipe, and polyurethane (PE) jacket. Tis FO cable adopts a unique internal fxed point design, which can realize the distributed measurement of discontinuous and nonuniform strain. Te manufactured UWFBG cable is identically spaced to the 1 m UWFBG array. Before the feld investigation, the reproducibility and repeatability of the UWFBG cable need to be verifed. As shown in Figure 3, a 1 m long FO cable with a UWFBG cutting from the same UWFBG cable is glued with an interval of 1 m and horizontally fxed at the performance test instrument with a length of 1 m. Such a strain-based performance test instrument can extend the cable by using the motor with a certain velocity. During the tensile process, the displacement is increased from 0 mm to 10 mm with a gradient of 1 mm. Ten, the cable rebounds to the initial length with a gradient of −2 mm. Te NZS-QDS-A01 Figure 1: Monitoring principle of UWFBG array.

Structural Control and Health Monitoring
UWFBGs interrogator analyzes the wavelength results after each displacement, and its technical specifcations are summarized in Table 2. Te same procedure is repeated fve times. Figure 4(a) shows that both the tensile and rebound curves have a good linear relationship, with the tensile and rebound paths basically overlapping. Tere is no great fuctuation between the fve tests, proving a good reproducibility and repeatability of the cable. Finally, the strain coefcient of this cable is calibrated to be 847.03 µε/nm, which is calculated by equation (2).
In addition, to verify the stability of strain monitoring in a long-term state, the cable is stretched to 5 mm, at which point the cable achieves a strain of 5000 µε. Ten, the stretch is kept, and the wavelength values are recorded at one-hour intervals for 11 hours, with fve sets of data recorded at a time. As it is shown here, the strain diferences are basically within ±5 με during 11 h, illustrating the stability of longterm monitoring (Figure 4(b)).
In the feld, the UWFBG cable is implanted into carbon fber cloth to form carbon fber composite optical cables, which are symmetrically arranged on the inner side wall of piles in a U-shape ( Figure 5). After basic cleaning such as line fxing, grinding, and dust removal, fxed-point laying is carried out; that is, the sensing cable is laid on the grinding surface, fxed with strong glue, and fully pasted with epoxy adhesive along the laying line ( Figure 6(a)). After the adhesive curing strength reaches more than 50%, gold foil paper is pasted on the surface to prevent later welding slag from burning ( Figure 6(b)). A section of channel steel is welded at the bottom of the pile to prevent the cable from being damaged during piling ( Figure 6(c)). At the top of the pile, the cable lead is protected with the steel hose and passed through the inner platform with a hoop fxed on the hook to prevent the line from falling of. Te sensing cable then is integrated into the main optical cable ( Figure 6(d)) and connected to the reserved interface of the fan ring network. It is transmitted to the booster station through the fan ring network and then connected to the monitoring equipment ( Figure 6(e)). Finally, equipment demodulation control software and monitoring software are installed in the onshore centralized control center to realize real-time online monitoring of the pile foundation.
It is important to point out that additional temperature compensation is necessary and important in areas where seawater temperature variations are strongly infuenced by solar radiation and climate change.

UWFBG Sensor Network for Ofshore Steel Pipe Piles.
As shown in Figure 7, for ofshore steel pipe pile monitoring, a UWFBG-wireless sensor network is established. Working as the sensing module, the sensing cables with UWFBGs for pile monitoring are tested indoors frst and then installed correctly to the steel pipe pile (Figures 7(a) and 7(b)). Ten, the wavelength is interpreted in time by using the UWFBG interrogator and remotely transmitted to the onshore centralized control center via 4G/5G Internet in the transmission module within 10 s (Figure 7(c)). Subsequently, the strain profle along the entire pile is output, and the target results are calculated according to the user requirement, mainly the pile force and the pile displacement (Figure 7(d)). Finally, a suitable warning algorithm is integrated into the early warning module to analyze the critical state of the ofshore steel pipe pile and decides whether to warn or not.

Test Plan.
Before the actual application, static load tests were carried out to verify the feasibility of the system. Te anchored pile reaction method was used in the static load test. As shown in Figure 8(a), the test apparatus mainly consisted of anchored piles, reaction beams, base beams, and hydraulic jacks. Te hydraulic jacks exerted the load on the pile top through the reaction force provided by reaction beams and four anchored piles. Te settlement of the pile top was obtained by the dial gauges arranged around the top pile body. Te static load test plan referred to the code (JTS237-4-2017). Each load stage was applied for 1 hour, and the load was from the initial value of 6000 kN to a maximum of 39000 kN with a gradient of 3000 kN for the frst 7 stages and 1500 kN for the rest. Te unloading process was maintained for 15 min for every stage with a gradient of 6000 kN. After complete unloading, the system was maintained for 1 hour.
No less than 2 days after the vertical static load test, the horizontal static load test was carried out (Figure 8(b)). Te unidirectional single cycle-loading method was adopted, and the static load test plan referred to the code (JTS237-4-2017). Each load stage was applied for 1 hour, and the load was from the initial value of 100 kN to a maximum of 1000 kN with a gradient of 100 kN for the frst 4 stages and 50 kN for the rest. Te unloading process was maintained for 10 min for every stage with a gradient of 200 kN. After complete unloading, the system was maintained for 30 min.     Te FO cables with UWFBGs were connected to the instrument, and the initial Bragg wavelength shift along the cable was measured by using the UWFBGs interrogator before the static load test. During the static load test, the drift of Bragg wavelength under each load will be measured until the end of the test. According to equation (2), the strain distribution of the cables under each load can be obtained.

Vertical Static Load Test.
Te manufactured UWFBG cable is identically spaced to the 1m UWFBG array; as a result, the strain can be obtained every 1 m along the pile depth under each load (Figure 9(a)). It can be seen that the strain of the UWFBG cable increased steadily with an increase in the load and decreased with an increase in the buried depth of the pile body (Figure 9(a)). Te strain distribution along the pile can be divided into four parts: a transition zone existed at the pile head. In the seawater, the strain was quite uniform, and in the soil part, the strain decreases gradually with the depth of the pile due to the efect of the shaft friction. Te rate of strain decline was diferent due to diferent friction resistance and pressure load. It was noted that, at the depth of about −45 m, the thickness of the steel pipe pile changed from 30 mm to 24 mm, which led to a sudden increase in pile strain and then continued to decrease with the increase in depth.
Young's modulus E is 206 GPa and the cross-sectional area A is 0.07-0.09 m 2 which are all fxed values in this case. Pile axial force and shaft friction under each load were calculated by equations (4) and (5), and the distribution of the axial force is shown in Figure 9(b). Te shaft friction was further averaged for each soil layer according to the soil layer distribution as shown in Figure 9(c). Te downward transfer load of the pile decreased continuously with an increase in depth due to the shaft friction. Figure 9(c) shows that the peak value of shaft friction was between a depth of −74.5∼−75.5 m, which was located in dense Layer 5. Tis was mainly caused by the diferent densities of soil layers.
Te development of shaft friction in diferent soil layers and tip resistance with top loads are plotted in Figure 10. It indicated that the top load was shared by the shaft friction and tip resistance. Te tip resistance increased linearly with an increase in the load. Te shaft friction bored most of the top load, while the tip resistance only bored 16.89% of the top load under the maximum load. Tis test pile should be defned as an end-bearing friction pile. Te slope of the shaft friction curves of Layers 2, 3, and 4 gradually slowed down, indicating that the shaft friction of these soil layers was about to be fully realized. While the shaft friction of Layer 5 was still increasing, that is, there was still a certain space from the recommended values of the geotechnical investigation report.

Structural Control and Health Monitoring
Te pile compression S s and tip-settlement S b calculated according to equations (7) and (8) and the top settlement monitored by the dial gauges are shown in Figure 11. As the load increased, the pile compression, tip-settlement, and top settlement both gradually increased. Te top settlement was borne by the compression of the pile and the settlement of the pile end, most of which was pile compression. When the load increased to 37500 kN, the pile compression began to ease and the settlement of the pile tip came into play. In addition, it could be seen that the steel pipe pile had not reached the ultimate bearing capacity. Figure 12, the distributed strain of F1 and F2 was in an antisymmetric relationship under the horizontal force, acting in the connection direction of F1 (Fiber 1) and F2 (Fiber 2). Te results indicated that the F1 side was subjected to tensile stress; oppositely, the F2 side was subjected to compressive stress. Te strain of the steel pipe pile increased with an increase in horizontal load and reached the maximum value of 1304 με at a load of 1000 kPa, and the position was at −26 m. Te strain increased frst with an increase in the depth, reached maximum approximately at −22 m to −26 m, and then began to decrease. From −53 m to −85 m, the strain remained unchanged with the load and was approximately around 0 με. Te strain at F2 was antisymmetric with F1, showing good consistency of monitoring and indicating that the stability of the pile-bearing capacity was not lost under diferent horizontal loads.

Horizontal Static Load Test. As shown in
Te bending moment was calculated by equation (12), and the distributed displacement cure of the pile was obtained by the second integral of the corresponding curve which is shown in Figure 13. Te bending moment of the pile body was mainly distributed in the range of about 40 m below sea level. With an increase in depth, the bending moment increased frst and then decreased. Te maximum bending moment point was between Layer 1 and Layer 2 and gradually transferred to deeper. Tat was because the pile had a fexural deformation under the horizontal load. With an increase in horizontal load, the yield zone of soil around the pile gradually expanded downward, and the maximum bending moment of the pile also shifted to the lower part due to the decrease of the resistance of the upper soil. As the load increased, the absolute value of the displacement measured by the UWFBG increased. Te maximum displacement occurred at the uppermost end of the fber and generated under the 1000 kN load was 544 mm. Te horizontal displacement gradually decreased along the pile depth and was almost no longer afected by the horizontal thrust beneath a depth of −30 m.
During loading, the horizontal displacement was also measured by using the dial indicators, as shown in Figure 14.
Te two values were consistent before 600 kN, the monitored value was slightly smaller than the measured value after 600 kN due to the deformation coordination of the pile, and the sensing cable became worse under large deformation. Te maximal absolute error was 75 mm, and the average value was 16 mm. Te relative error was further analyzed by the MAPE (mean absolute percentage error) method [41] to where y t denotes the true value as a comparison, y t denotes the value to be compared, and n denotes the sample capacity. After calculation, the MAPE value, in this case, was 0.058, which showed that the full-scale distributed displacement data calculated by UWFBGs were relatively accurate and credible.

Overview of the Study Area.
Te FO wireless sensor network was deployed to the No. 3 ofshore wind project in South Shandong Peninsula, China ( Figure 15). Yellow Sea circulation is the main current in the study area. It is basically composed of the Yellow Sea warm current, carrying high temperature and high salt water northward, and two east-west coastal currents along its two sides, which migrates southward [42]. Tis coast has relatively strong tides and weak waves with an average one-tenth large wave height Structural Control and Health Monitoring [43]. Te seawater depth is shallow in this sea area, with a depth of 0-50 m.
Te frst phase of the wind farm was planned to install 58 wind power units (A01∼A58) with a single-pile steel pipe pile foundation. Te steel pile had a length of 97 m, a pile diameter of 1 m, and a wall thickness of 24-30 mm. A 220 kV ofshore booster station and an onshore centralized control center were together constructed (Figure 15(c)). Tree wind power piles A14, A41, and A45 were selected for operation period monitoring. After completing the integration of single pile lines of three steel pipe piles, the single pile lines were connected to the corresponding monitoring interfaces. Line integration of equipment access and monitoring signal output was completed at the booster station to ensure smooth transmission of monitoring data to the onshore central control center. Tree ofshore steel pipe piles had been monitored in real-time during operation. Since January 1, 2022, the monitoring frequency could be set according to the demand, even up to the second level.
As for ofshore pile foundation monitoring, the wavelength variation due to pile strain is much greater than the efect of temperature. However, the seawater levels can fuctuate frequently due to tidal action, with fuctuations within a range of 3 m in the study area which may lead to anomalies in both tensile and compressive strains on the same side of the pile. Tus, only the data 3 m below sea level were used for further analysis. Moreover, according to the China Argo Real-time Data Center, the distribution curve of seawater temperature along the depth is approximately a straight line, showing basically no change in the 0-50 m depth of this area. Tat is, the efect of temperature is relatively small, so no additional temperature compensation was made in this case.

Long-Term Monitoring Results during the Operation
Period. Te real-time monitoring data of the A41 wind power foundation were selected at 00:00 time every day from January 1, 2022, to June 1, 2022, to obtain the strain distribution curves of the symmetrical sensing cable on both sides of the pile. As shown in Figure 16, the strain curve refected the sea level and mud-water interface. In seawater, the pile strain gradually decreased with the increasing depth and presented a stable distribution near the seabed mud surface. After entering the soil layer, the strain value gradually decreased with the depth, but the pile bottom was not afected. Te maximum strain of the pile under a sea level occurred near the sea surface, which was about 250 με. During the operation of the wind turbine, the wind and waves at sea will produce horizontal loads on the pile foundation. According to the symmetrical cable strain, one side of the pile exhibits tensile strain and the other side exhibited compressive strain, and the strain values were the same. Te horizontal displacements calculated by equation (11) are shown in Figure 17, the results indicated that the maximum displacement occurred at the uppermost end of the fber, and the absolute value was 58 mm. Tere was no horizontal displacement of the pile from the pile bottom to −50 m elevation; that is, under the existing horizontal load, the pile was completely constrained after it is about 18 m into the soil.

Infuence Factors of Ofshore Steel Pipe Pile Structure
Health. Te ofshore wind turbine pile foundation is built in a complex marine environment and subjected to the combined action of long-time wind, wave, current, and other dynamic loads, which leads to the deviation and bending of the fan foundation, and even causes a series of safety problems such as deformation, liquefaction, and instability of the seabed around the fan pile foundation. Te wind load is an important load acting on the fan blades and tower cylinder, which will produce huge horizontal thrust and overturning moment, and causes the horizontal swing of the pile body. In addition, the wave is always accompanied by water fow in the process of propagation, and the wave fow load will produce a horizontal cyclic load on the foundation of the ofshore fan pile, resulting in horizontal displacement. At the same time, the wave load acting on the sea bed will generate vertical cyclic load on the sea bed surface in the process of continuous propagation and transfer to the pile foundation to generate additional horizontal displacement. Figure 18 shows the variation curve of the pile mud surface angle calculated according to equation (10) under diferent wind speeds and wave heights from January 1 to January 29. It can be found that pile swing amplitude had a signifcant correlation with wind speed and wave height. Te increase in wind speed led to the increase of horizontal thrust and overturning moment of the pile foundation; meanwhile, the larger wind speed led to the acceleration of wave velocity, which increased the wave height, aggravated the wave disturbance around the pile, and intensifed the angle of the pile mud surface.
To further analyze the swing characteristics of the pile under working conditions, the variation curves of pile strain at diferent depths of the pile foundation from 18:00 to 18:10 on January 1 are shown in Figure 19(a). It can be seen that the steel pipe pile presented a state of horizontal swing. Te oscillation amplitude was the most obvious at the sea surface and became smaller with increasing depth. Tere was no oscillation at the bottom of the pile foundation. Te variation curves of the pile strain at a depth of −2.4 m from 18:00 to 18:10 on the frst day of each month in the monitoring cycle in Figure 19(b) show that the swing amplitude of the pile also presented a certain correlation with the seasonal ocean current. In the winter of January and February, the ocean current movement was strong, the velocity was fast, and the wind and waves were large; hence, the amplitude of the pile foundation swing was also severe. In spring and summer from March to June, the movement of ocean currents was weak, so the amplitude of the pile foundation swing was relatively gentle.

Working State Evaluation and Early Warning.
According to the DNV-OS-J103 specifcation [44], ofshore wind turbine single pile foundations can be divided into four limit states, which are ultimate limit state (ULS), fatigue limit state (FLS), accidental limit state (ALS), and serviceability limit state (SLS). Previous research has shown that the failure of the foundation is mainly caused by piles exceeding the allowable rotation angle under the SLS [45]. Te wind, wave, and current loads applied to the ofshore wind turbine are not constant static but long-term cyclic forces. During the service of the wind turbine, the cycle loads of the foundation can reach 10 7 ∼10 8 times. Under the cyclic load, the soil stifness around the foundation is constantly weakened and the plastic strain accumulates, which eventually leads to excessive lateral displacement at the mud surface of the pile foundation, thus resulting in SLS failure. DNV-ST-0126 [46] specifes that the rotational angle of the mud surface caused by the cyclic load should not exceed 0.25°for a single-pile ofshore fan foundation.
Te FO wireless sensor network previously proposed is recommended for working state evaluation and early warning for the wind turbine foundation infrastructure. First, real-time monitoring of pile strains is realized by UWFBG technology. Ten, the spatialtemporal data are transmitted to the cloud module and server for processing by the transmission module and 4G/5G Internet to calculate the rotational angle of the pile at the corresponding time. Here, we set 0.25°as the serviceability limit safety threshold. Once the rotation angle at the mud surface exceeds the safety threshold, the monitoring system will issue a level I warning, predicting that the wind farm may be in a bad operating environment and should be observed continuously for a while to determine whether it reaches the SLS. If the rotation angle frequently exceeds the safety threshold with an increasing tendency, the system will determine that the wind power base reaches the SLS and gives a level II warning. At the same time, the wind turbine will not work properly.
Taking the A41 wind power foundation as an example, the real-time monitoring data from January 1, 2022, to June 1, 2022 ( Figure 20) showed that during this period, the mud surface rotational angle did not exceed the safety threshold, and the maximum value was 0.246 at 15:00 on February 1, without triggering an early warning. Te system automatically reported that the wind turbine was in a normal operation state.

Conclusions
Based on UWFBGs, a wireless sensor network for ofshore steel pipe pile monitoring was proposed, and a case study in South Shandong Peninsula, China was demonstrated. UWFBGs were adopted to automatically obtain the pile strain during the static load test as well as the operation period to achieve life cycle monitoring. For the ofshore steel pipe pile, the load transfer behavior was analyzed and the working state was evaluated. Te following conclusions can be drawn: (1) Te UWFBG-based wireless sensor network used for ofshore steel pipe pile monitoring is frst proposed, which is efcient in mapping the distributed strain along the pile and realizes the remote sensing for life cycle monitoring of ofshore steel pipe piles. maximum shaft friction is located at the layer with the highest density. In addition, the tip resistance only bores 16.89% indicating that the test pile should be defned as an end-bearing friction pile. (3) Under the horizontal static load, with an increase in depth, the strain increases frst and then decreases, the peak of pile bending moment transfers to the deeper layer, and the horizontal displacement gradually decreases. Te calculated displacement is compared with the measured data by using dial indicators, and the MAPE value is only 0.058, proving accurate and credible monitoring. (4) During operation time, the strain data obtained from the Shandong Peninsula ofshore wind power project refect the sea level and water-mud interface and show that the maximum displacement occurs at the uppermost end of the fber with an absolute value of 75 mm. (5) Te horizontal deformation of the wind turbine pile is positively correlated with the wind, wave, and current load and decreases with an increase of depth.
Half-year monitoring results show that the horizontal deformation in winter is signifcantly greater than that in other seasons owing to the seasonal ocean current. (6) Te UWFBG-based fber-optic wireless sensor network can be used for early warning of ofshore wind piles. Setting 0.25°for the rotational angle of the mud surface as a serviceability limit safety threshold, the wind turbine in the study area is in a normal operation state.
Te feld study presented in this paper will provide a valuable reference for performance monitoring and evaluation of steel pipe piles, especially ofshore overlength ones. Certainly, longer monitoring data will provide a more comprehensive basis for analysis.

Data Availability
Te data that support the fndings of this study are available on request from the corresponding author.

Conflicts of Interest
Te authors declare that they have no conficts of interest.