WEN Shipeng, XU Jishang, HU Guanghai, DONG Ping, and SHEN Hong

1)College of Marine Geo-Sciences,Ocean University of China,Qingdao266100,P. R. China

2)Key Lab of Submarine Sciences & Prospecting Techiniques,MOE,Ocean University of China,Qingdao266100,P. R. China

3)Offshore Oil Production Factory,Shengli Oilfield Subsidiary Company,Dongying257237,P. R. China

4)The First Institute of Oceanography,Qingdao266100,P. R. China

5)Division of Civil Engineering,School of Engineering,University of Dundee,DundeeDD1 4HN,U.K.

A Field Investigation on the Effects of Background Erosion on the Free Span Development of a Submarine Pipeline

WEN Shipeng1),3), XU Jishang1),2),*, HU Guanghai4), DONG Ping1),5), and SHEN Hong4)

1)College of Marine Geo-Sciences,Ocean University of China,Qingdao266100,P. R. China

2)Key Lab of Submarine Sciences & Prospecting Techiniques,MOE,Ocean University of China,Qingdao266100,P. R. China

3)Offshore Oil Production Factory,Shengli Oilfield Subsidiary Company,Dongying257237,P. R. China

4)The First Institute of Oceanography,Qingdao266100,P. R. China

5)Division of Civil Engineering,School of Engineering,University of Dundee,DundeeDD1 4HN,U.K.

The safety of submarine pipelines is largely influenced by free spans and corrosions. Previous studies on free spans caused by seabed scours are mainly based on the stable environment, where the background seabed scour is in equilibrium and the soil is homogeneous. To study the effects of background erosion on the free span development of subsea pipelines, a submarine pipeline located at the abandoned Yellow River subaqueous delta lobe was investigated with an integrated surveying system which included a Multibeam bathymetric system, a dual-frequency side-scan sonar, a high resolution sub-bottom profiler, and a Magnetic Flux Leakage (MFL) sensor. We found that seabed homogeneity has a great influence on the free span development of the pipeline. More specifically, for homogeneous background scours, the morphology of scour hole below the pipeline is quite similar to that without the background scour, whereas for inhomogeneous background scour, the nature of spanning is mainly dependent on the evolution of seabed morphology near the pipeline. Magnetic Flux Leakage (MFL) detection results also reveal the possible connection between long free spans and accelerated corrosion of the pipeline.

submarine pipeline; scour; span; corrosion

1 Introduction

For submarine pipelines located on erodible seabed, scours will develop, causing free spanning and threatening the stability of the pipelines. Some researchers have studied the mechanism of the local scour. It is recognized that the excessive seepage flow and the resulting piping are the major factors to cause the onset of scour below the pipeline (Mao, 1988; Chiew, 1990; Sumeret al., 2001; Sumer and Fredsoe, 2002), and the effect of vortices formed in the neighborhood of the pipeline is the key element in the scour process (Sumer and Fredsoe, 1990; ?evik and Yüksel, 1999). Numerous formulae were proposed for predicting the geometric parameters of maximum scour below pipelines such as scour depth (Sumer and Fredsoe, 1990; Sumeret al., 2001; ?evik and Yüksel, 1999; Dey and Singh, 2007; Yasa, 2011) and scour width (Cata?o-Lopera and García, 2007).

However, these studies are largely based on stable laboratory environment, where the background seabed is in morphological equilibrium, the soil bed is homogene-ous, and the hydrodynamic condition is uniform. However, in actual marine environment, scour processes are more complex, where background seabed can undergo irregular erosion due to heterogeneous seabed and varying hydrodynamic conditions along the pipeline. It is difficult to predict the maximum scour depth and scour width using the conventional formulae. In such situations, it is necessary to evaluate heterogeneous scours and their effects on the span development of submarine pipelines.

Corrosion could also be a major problem in pipeline engineering affecting long-term reliability of metallic submarine pipelines (DNV, 2010) and causing approximately 25% of all reported accidents (Beaverset al., 2006). Pipelines are always prone to external/internal corrosion due to the presence of chemical substances such as chloride, oxygen (O2), carbon dioxide (O2), hydrogen sulphide (H2S), and microbiological bacteria. It is recognized that corrosion rate is related to pH values, temperature, pressure, salinity, conductivity, electrical resistivity, redox potential, shrink/swell properties, materials flow rate and flow profile of the system (Adedipe, 2011; Srikanthet al., 2005; Budieaet al., 2012). No previous research has linked the corrosion of the pipeline with its spanning state.

The Shengli oilfield, located at the Huanghe River Delta, is the second largest oil field in China in terms ofannual petroleum production. The offshore oilfield of Shengli has 160 sections of submarine pipelines with 220 km in total length. Owing to the heterogeneous seabed scour background and the long-time operation of these offshore pipelines, free spans and wall corrosions are becoming the most important issues concerning the safety of the pipelines.

To meet both theoretical and practical requirements, we investigated a section of pipeline using an integrated surveying system including a multi-beam, a side-scan sonar and Magnetic Flux Leakage (MFL) sensors. Based on the field investigation data, this paper discusses the effects of background erosion on the free span development of the subsea pipelines. The engineering implications of pipeline free spans are also discussed. Field investigation of the pipeline shows that most detected inner corrosion points are distributed in a long spanning section; such a phenomenon, which has not been mentioned by any previously published literatures, indicates that long spans possibly accelerate the corrosion of the pipeline.

2 Study Area

2.1 Geological Background

The Shengli offshore oilfield is located at modern Huanghe River delta. The delta is composed of many delta lobes, because the distributaries shift frequently owing to the high sediment concentration, leading to formation of new delta lobes at the active channel mouths and land erosion at the abandoned lobes (Liet al., 2000; Yu, 2002). The investigated submarine pipeline is located at an abandoned Diaokou subaqueous delta lobe which came into being between 1953 and 1976 (Fig.1). At the time of this deltaic-lobe formation, the amount of sediments entering the sea was about 1.08×109tons annually (Milliman and Meade, 1983). In 1976, the Huanghe river channel migrated to the Qingshuigou course, and the abandoned Diaokou subaqueous delta lobe began suffering from significant erosion because of the lack of fluvial sediment supplies to compensate the sediment removal by hydrodynamic processes (Wanget al., 2006). The maximum erosion thickness is over 8.5 m in the past few years (Liet al., 2000). The evolution process of the abandoned delta was described by Liet al.(2000).

Fig.1 Location of the investigated pipeline, and the historical changes of the Huanghe River channel (modified after Liet al., 1998).

2.2 Sediment Distribution

The distributions of both sediment type and median grain size on the Yellow River subaqueous delta are heterogeneous (Liet al., 1998; Yu, 2002; Jiaet al., 2011). The sedimentary environments and associated sedimentary distribution are characterized by Liet al.(1998) and Yu (2002) as follows： 1) The sediments at the river- mouth bar are mainly fine sand, with silty clay and clayey silt depositing on both sides of the river mouth; 2) The deltaic front between the 2 and 11 m isobaths mainly consists of silty sediment; 3) The prodelta at depths ＞11 m is commonly clayey silt. The sediment at the study area tends to be easily eroded and may also subside due to sedimentary compaction processes (Chuet al., 2006). Under the load of external forces such as waves, currents, and earthquakes, the seabed at the steep subaqueous slope of the study area is extraordinarily unstable because of the formation model of the Yellow River delta which is characterized by rapid deposition, frequent swinging of watercourse and overlapping of individual delta lobes (Liet al., 2000; Yu, 2002).

2.3 Waves and Currents

The study area is a typical wave- and current- dominated coast. Waves with wave height H1/10＜0.5 m have the highest frequency of 51.1%, and waves with wave height 0.5＜H1/10＜1.5 m, 1.5＜H1/10＜3 m, and H1/10＞3.0 m have frequencies of 36.3%, 11.8%, and 0.5% respectively (Chuet al., 2006). Waves with wave height H1/10＞1.5 m in this area are generally from NE and partially from SE (Chuet al., 2006). The significant wave height generated in the deep water near the study area can reach up to 8.0 meters with corresponding wave period of 10.3 seconds (Wanget al., 2006). The measured maximum significant wave height was 5.2 m at the 7 m water depth off the Yellow River Port, approaching from the NE direction in October 1986 (Zang, 1996; Chuet al., 2006). On average, the maximum NE wave height and period are 2.5-3.5 m and 5-8 s respectively (Zang, 1996).

The tidal current has a maximum surface velocity that exceeds 1.2 m s-1along the shore (Liet al., 2000). Nonlinear wave-current interaction can enhance the bottom shear stress (Grant and Madsen, 1982). Wanget al.(2006) argued that the high stress zone locates in the depths of approximately 4-10 m from the Diaokou mouth. Beyond the ends of the Huanghe delta slope at 14-16 m depth, the bottom shear stress caused by wave action is significantly weakened. The residual current velocity along the entireYellow River Delta often ranges from 10 to 30 cm s-1, which is mainly induced by wind (Chuet al., 2006). The bottom sediments are mainly resuspended by the waves and then transported by the strong tidal currents (Wanget al., 2006; Liet al., 2000).

3 Investigation Methods

The integrated surveying system used in this research included an acoustic sensor compartment and a Magnetic Flux Leakage (MFL) sensor compartment. The combined acoustic surveying system included a multi-beam, sidescan sonar and sub-bottom profiler, to detect the suspending height, burial depth and as-laid position of the pipeline. The Magnetic flux leakage (MFL) sensors were used to detect corrosion, pitting and wall loss in the metallic pipelines.

Sonic2024 multi-beam echosounder was used to provide the burial/suspending conditions and as-laid position of the pipeline as well as the topographic changes of the local seabed. It operated at 300 kHz with 256 beams. The multi-beam data covered about 100 m on each sides of the pipeline route.

A dual-frequency side-scan sonar (EdgeTech 4200MP) was used to produce images of the seafloor and suspending condition of the pipeline. It operates at both 100 kHz and 400 kHz. The recording range is 100 m with highresolution mode.

A parametric sub-bottom profiler series SES-2000 was used to visualize sediment structures and locate embedded pipeline sections. It works at the primary frequency (PHF) of 100 kHz, and secondary frequencies (SLF) of 4, 5, 6, 8, 10 and 12 kHz, with vertical resolution up to 5 cm.

Magnetic flux leakage (MFL) tools were used to measure the change in magnetic flux lines produced by the defect and produce a signal that can be correlated to the length and depth of a defect.

4 Results

4.1 Bathymetry Along the Route of the Pipeline

The bathymetry along the route of the pipeline varies between 3.2 m and 14.6 m, with water depth deepening from the landing point to the platform (Fig.2). Water depth in the landing area is shallow, with several scour trenches and scour pits which are caused by the combination action of coastal waves and strong tidal currents. The overall topography in the middle section is relatively flat, except for some scour-induced residual highland and local scour trenches along the pipeline. The deepest seabed along the pipeline locates near the platform, with the maximum water depth of 14.6 m, which is about 3 m deeper than ambient seabed.

Fig.2 Seabed topography, along with free span and corrosion positions of the pipeline. Pink lines indicate spanning pipeline sections (being marked with SP1-SP15 respectively. See Table 1). Blue dots represent corrosion positions detected through MFL sensors (being marked with C1-C11 respectively. See Table 2). Red dots which be marked with KP1-KP6 denote the distance from the oil platform (km).

4.2 Free Spans of the Pipeline

The measurements show that 48% of the pipeline is buried, while 52% is unburied, in which 14% was free spanning. The spanning status of the pipeline is shown in Table 1. It can be seen that free spans are mainly distributed in water depth of 9-11 m.

4.3 Corrosion of the Pipeline

The corrosion condition of the pipeline was inspected successively using Magnetic Flux Leakage (MFL) sensors. Eleven corrosion points were detected, nine of which are inner corrosions (Fig.2, Table 2). For these nine inner corrosion points, eight points are distributed in the section 1069.9 m to 1073.0 m, where the pipeline is in free spanning (SP5 in Table 1 and Fig.2, with span length of 333.8 m).

5 Discussion

5.1 Effects of Background Erosion on Span Development

The mechanism of the local scour around pipelines was studied by many researches (Mao, 1988; Chiew, 1990;Sumeret al., 2001; Sumer and Fredsoe, 2002), but most of these studies had focused on stable environment, where the background seabed is in morphological equilibrium, the experimental soil is homogeneous, and the hydrodynamic condition is uniform. As the seabed at the entire abandoned subaqueous delta lobe suffered significant erosion, the mechanism causing the free spans under this background erosion must be different from what is known from previous studies where the background erosion is absent. Based on the field investigation data, the effects of background erosion on the free span development of subsea pipelines are summarized below.

Table 1 Location and spanning status of free spans

Table 2 corrosion detection result from the Magnetic Flux Leakage (MFL) sensors

5.1.1 Homogeneous background scour

For homogeneous seabed in the middle section of the pipeline route, where the overall topography is relatively flat, the morphology of scour hole below the pipeline (Fig.3) is quite similar to the scour that is induced by vortex in the wake region of pipeline (Xuet al., 2009, 2010, 2012; Chiew, 1990; Sumeret al., 2001), showing that vortex plays an important role in the span development of submarine pipelines. This is possibly due to the fact that homogeneous background erosion is accompanied by the sinking of the pipeline (Sumer and Freds?e, 1990; Sumeret al., 2001; Xuet al., 2009), making the relative distance between the pipeline and the ambient seabed unchanged in the scour process. Such a phenomenon indicates that existing formulae for predicting maximum seabed scour depth and width below pipeline (Sumeret al., 1990, 2001; ?evik, 1999; Dey and Singh, 2007; Yasa, 2011) are also applicable in the presence of homogeneous background erosion.

5.1.2 Inhomogeneous background scour

Due to the heterogeneous properties of surfacial sediments (Liet al., 1998; Yu, 2002; Jiaet al., 2011) and the varying hydrodynamic conditions along the pipeline (Wanget al., 2006; Liet al., 2000; Chuet al., 2006) in the subaqueous delta of the Yellow River, most seabed scours in the study area exhibit heterogeneous characteristics, characterized by scour trenches, scour pits, and scour- induced residual highlands along the route of the pipeline (Fig.2). Most spans of the investigated pipeline are caused by inhomogeneous background scour (Table 1). Figs.4, 5, 6, and 7 are typical multi-beam images showing free spans caused by inhomogeneous background erosion. It can be seen that the spanning status of the pipeline is mainly dependent on the evolution of seabed morphology adjacent to the pipeline.

5.1.3 Effect of background erosion on artificially supported pipeline

Underwater pile-support technology is widely used toconstrain free spanning pipelines. This technology is also used for supporting the spanning pipeline near the platform (Fig.8a). However, field investigation results show that, due to background erosion, the span height at the pilesupported pipeline can grow larger, with the maximum span height being up to 1.05 m (Fig.8). This phenomenon shows that the traditional supporting technology commonly used for protecting free spanning pipelines may not work as well as intended if the spans are largely caused by background erosion.

Fig.3 The morphology of scour hole below the pipeline. (a) Surface map, showing the flat background seabed; (b) Profile map.

Fig.4 The morphology of free spanning pipeline SP2 (see Table 1). (a) Surface map, showing the inhomogeneous background scour; (b) Profile map.

Fig.5 The morphology of free spanning pipeline SP4 (see Table 1). (a) Surface map, showing the inhomogeneous background scour; (b) Profile map.

Fig.6 The morphology of free spanning pipeline SP8 (see Table 1). (a) Surface map, showing the inhomogeneous background scour; (b) Profile map.

Fig.7 The morphology of free spanning pipeline SP15 (see Table 1). (a) Multi-beam data; (b) Side-scan sonar image.

Fig.8 The morphology of free spanning pipeline SP1 (see Table 1). (a) Surface map, showing the inhomogeneous background scour; (b) Profile map.

5.2 Effects of Background Erosion on Pipeline Stability

Seabed background erosion can influence the instability of pipeline in a number of ways. Firstly, background scour can cause long free spans (Table 1), where the unsupported pipeline could be overstressed due to selfweight, thus causing an increase in the rupture risk of the pipeline (DNV, 2006, 2008; Xuet al., 2010). Secondly, long spanning pipeline is possibly subjected to vibrations induced by vortex shedding from the span (DNV, 2006, 2008; Choi, 2001; Xuet al., 2010). These static or dynamic loads over time can lead to fatigue failure of the pipeline (DNV, 2006, 2008). Thirdly, the vibration of spans may cause the buildup of excess pore pressure, even leading to soil liquefaction (Puet al., 2013), a process which can significantly affect processes of sediment erosion (Clukeyet al., 1985; Zhouet al., 2011) as well as the stability of seabed (Sassaet al., 2006) and pipeline (Forayet al., 2006).

What’s more, Magnetic Flux Leakage (MFL) detection results show that eight of the nine inner corrosion points are located in a free spanning section (SP5 in Table 1, with span length of 333.8 m), indicating that long free spans may accelerate the corrosion of the pipeline. Although the precise reason for this phenomenon is unclear, it may be related to the large Vortex Induced Vibrations (VIV) of the spanning pipeline under the action of cur-rents and waves (DNV, 2006, 2008).

6 Conclusions

Based on the field investigation data of a submarine pipeline located at the abandoned Yellow River subaqueous delta lobe that is suffering from significant erosion, the effects of background erosion on the free span development of subsea pipelines were studied, and the effects of spans on pipeline stability were discussed. The main conclusions are as follows：

1) Seabed homogeneity and hydrodynamic unity have great influence on the free span development of subsea pipelines. For homogeneous background scours, vortex plays an important role in the span development of submarine pipelines, and the morphology of scour hole below the pipeline is quite similar to that without the background scour. For inhomogeneous background scour, the spanning status of the pipeline mainly depends on the evolution of seabed morphology near the pipeline.

2) Traditional supporting technology commonly used for protecting free spanning pipelines may not be as effective as widely believed if the spans are caused by inhomogeneous background erosion.

3) Long free spans could accelerate the corrosion of the pipeline.

Acknowledgements

This work is financially supported by the National Natural Science Foundation of China (Grant No. 41006024) and Marine Specific Research for Public ‘Forecast and appraisal for geologic hazard of inshore seafloor and study of key technology for protection’ (Grant No. 201005005).

Adedipe, O., 2011. Characterization of external induced corrosion degradation of Ajaokuta-Abuja gas pipeline system, Nigeria.International Journal of Engineering Science and Technology, 3 (11)： 8061-8068.

Beavers, J. A., Thompson, N. G., and Technologies, C. C., 2006.External Corrosion of Oil and Natural Gas Pipelines. ASM Handbook, Volume 13C, 1015-1025.

Budiea, A. M. A., Yahaya, N., and Nor, N. M., 2012. Corrosion of API X70 steel due to near shore sediment.International Journal of Civil & Environmental Engineering, 12 (3)： 84-88.

Cata?o-Lopera, Y. A., and García, M. H., 2007. Geometry of scour hole around, and the influence of the angle of attack on the burial of finite cylinders under combined flows.Ocean Engineering, 34 (5)： 856-869.

?evik, E., and Yüksel, Y., 1999. Scour under submarine pipelines in waves in shoaling conditions.Journal of Waterway, Port, Coastal and Ocean Engineering, 125 (l)： 9-19.

Chiew, Y. M., 1990. Mechanics of local scour around submarine pipelines.Journal of Hydraulic Engineering, 116 (4)： 515-529.

Choi, H. S., 2001. Free spanning analysis of offshore pipelines.Ocean Engineering, 28： 1325-1338.

Chu, Z. X., Sun, X. G., Zhai, S. K., and Xu, K. H., 2006. Changing pattern of accretion/erosion of the modern Yellow River (Huanghe) subaerial delta, China： Based on remote sensing images.Marine Geology, 227： 13-30.

Clukey, E. C., Kulhawy, F. H., Liu, P. L. F., and Tate, G. B., 1985. The impact of wave loads and pore-water pressure generation on initiation of sediment transport.Geo-Marine Letters, 5： 177-183.

Dey, S., and Singh, N. P., 2007. Clear-water scour depth below underwater pipelines.Journal of Hydro-Environment Research, 1 (2)： 157-162.

DNV, 2006.Free Spanning Pipelines. Recommended practice DNV-RP-F105, Det Norske Veritas, H?vik, Norway, 1-45.

DNV, 2008.Structural Analysis of Piping Systems.Recommended practice DNV-RP-D101, Det Norske Veritas, H?vik, Norway, 1-32.

DNV, 2010.Corroded Pipelines. Recommended practice DNVRP-F101, Det Norske Veritas, H?vik, Norway, 1-42.

Foray, P., Bonjean, D., Michallet, H., and Mory, M., 2006. Fluidsoil-structure interaction in liquefaction around a cyclieally moving cylinder.Journal of Waterway, Port, Coastal, and Ocean Engineering, 132 (4)： 289-299.

Grant, W. D., and Madsen, O. S., 1982. Movable bed roughness in unsteady oscillatory flow.Journal of Geophysical Research, 87： 469-481.

Jia, Y. G., Shan, H. X., Yang, X. J., Meng, X. M., Chang, F. Q., and Zheng, J. W., 2011.Sediment Dynamics and Geology Hazards in the Estuary of Yellow River. Science Press, Beijing, 24-130.

Li, G. X., Wei, H. L., Yue, S. H., Cheng, Y. J., and Han, Y. S., 1998. Sedimentation in the Yellow River delta, part II： Suspended sediment dispersal and deposition on the subaqueous delta.Marine Geology, 149： 113-131.

Li, G. X., Zhuang, K. L., and Wei, H. L., 2000. Sedimentation in the Yellow River delta. Part III. Seabed erosion and diapirism in the abandoned subaqueous delta lobe.Marine Geology, 168： 129-144.

Mao, Y., 1988. Seabed scour under pipelines.Proceedings of 7th International Conference on Offshore Mechanics and Arctic Engineering, Houston, TX, 33-38.

Milliman, J. D., and Meade, R. H., 1983. World-wide delivery of river sediment to the oceans.Journal of Geology, 91： 1-21.

Pu, J. J., Xu, J. S., and Li, G. X., 2013. Experimental study on damping characteristics of pipe vibration in liquefied silt.The 23rd International Offshore and Polar Engineering Conference (ISOPE), Alaska, USA, 266-271.

Sassa, S., Takayama, T., Mizutani, M., and Tsujio, D., 2006. Field observations of the build-up and dissipation of residual pore water pressures in seabed sands under the passage of storm waves.Journal of Coastal Research, 39 (Special Issue)：410-414.

Srikanth, S., Sankaranarayanan, T. S. N., Gopalakrishna, K., Narasimhan, B. R. V., Das, T. V. K., and Das, S. K., 2005. Corrosion in a buried pressurised water pipeline.Engineering Failure Analysis, 12： 634-651.

Sumer, B. M., and Fredsoe, J., 1990. Scour below pipeline in waves.Journal of Waterway, Port, Coastal and Ocean Engineering, 116 (3)： 307-323.

Sumer, B. M., and Fredsoe, J., 2002.The Mechanics of Scour in the Marine Environment. World Scientific, Singapor, 15-148.

Sumer, B. M., Truelsen, C., Sichmann T., and Freds?e, J., 2001. Onset of scour below pipelines and self-burial.Coastal Engineering, 42 (4)： 313-335.

Wang, H. J., Yang, Z. S., Li, G. X., and Jiang, W. S., 2006. Wave climate modeling on the abandoned Huanghe wave climatemodeling on the Abandoned Huanghe.Journal of Coastal Research, 22 (4)： 906-918.

Xu, J. S., Li, G. X., and Horrillo, J. J., 2009. Local scour and self-burial of a submarine pipeline on cohesive seabed.ASCE International Conference on Pipelines and Trenchless Technology (ICPTT), 361： 996-1004.

Xu, J. S., Li, G. X., Dong, P., and Shi, J. H., 2010. Bedform evolution around a submarine pipeline and its effects on wave-induced forces under regular waves.Ocean Engineering, 37： 304-313.

Xu, J. S., Li, G. X., Horrillo, J. J., Yang, R. M., and Cao, L. H., 2010. Calculation of maximum allowable free span length and safety assessment of the DF1-1 submarine pipeline.Journal of Ocean University of China, 9 (1)： 1-10.

Xu, J. S., Pu, J. J., and Li, G. X., 2012. Field observations of seabed scours around a submarine pipeline on cohesive bed.Advances in Computational Environment Science, 142： 23-33.

Yasa, R., 2011. Prediction of the scour depth under submarine pipelines - in wave condition.Journal of Coastal Research, 64： 627-630.

Yu, L. S., 2002. The Huanghe (Yellow) River： A review of its development, characteristics, and future management issues.Continental Shelf Research, 22： 389-403.

Zang, Q. Y., 1996.Nearshore Sediment Along the Yellow River Delta. Ocean Press, Beijing, 11-51.

Zhou, C. Y., Li, G. X., Dong, P., Shi, J. H., and Xu, J. S., 2011. An experimental study of seabed responses around a marine pipeline under wave and current conditions.Ocean Engineering, 38： 226-234.

(Edited by Xie Jun)

(Received August 19, 2013; revised January 12, 2014; accepted May 30, 2015)

? Ocean University of China, Science Press and Springer-Verlag Berlin Heidelberg 2015

* Corresponding author. Tel： 0086-532-66781972 E-mail： jishangxu@ouc.edu.cn