Application of the LM-BP neural network approach for 1 landslide risk assessments 2

Landslide disaster is one of the main risks involved with the operation of long-distance oil and 14 gas pipelines. Because previously established disaster risk models are too subjective, this paper presents 15 a quantitative model for regional risk assessment through an analysis of the laws of historical landslide 16 disasters along oil and gas pipelines. Using the Guangyuan section of the Lanzhou-Chengdu-Chongqing 17 (LCC) Long-Distance Products Oil Pipeline (82km) in China as a case study, we successively carried out 18 two independent assessments: a hazard assessment and a vulnerability assessment. We used an entropy 19 weight method to establish a system for the vulnerability assessment, whereas a Levenberg Marquardt20 Back Propagation (LM-BP) neural network model was used to conduct the hazard assessment. The risk 21 assessment was carried out on the basis of two assessments. The first, the system of the vulnerability 22 assessment, considered the pipeline position and the angle between the pipe and the landslide (pipeline 23 laying environmental factors). We also used an interpolation theory to generate the standard sample 24 matrix of the LM-BP neural network. Accordingly, a landslide hazard risk zoning map was obtained 25 based on hazard and vulnerability assessment. The results showed that about 70% of the slopes were in 26 high-hazard areas with a comparatively high landslide possibility and that the southern section of the oil 27 pipeline in the study area was in danger. These results can be used as a guide for preventing and reducing 28 regional hazards, establishing safe routes for both existing and new pipelines and safely operating 29 pipelines in the Guangyuan section and other segments of the LCC oil pipeline. 30


33
By the year 2020, the total mileage of long-distance oil and gas pipelines is expected to exceed 160,000 34 km in China. This represents a major upsurge in the mileage of multinational long-distance oil and gas 35 pipelines (Huo, Wang, Cao, Wang, & Bureau, 2016). The rapid development of pipelines is associated 36 with significant geological hazards, especially landslides, which increasingly threaten the safe operation 37 of pipelines (Wang et al., 2012;Yun & Kang, 2014;Zheng, Zhang, Liu, & Wu, 2012). Landslide disasters 38 cause great harm to infrastructure and human life. Moreover, the wide impact area of landslides restricts 39 the economic development of landslide-prone areas (Ding, Heiser, Hübl, & Fuchs, 2016;Hong, Pradhan, 40 Xu, & Bui, 2015). A devastating landslide can lead to casualties, property losses, environmental damage 41 and long-term service disruptions caused by massive oil and gas leakages (G. Li, Zhang, Li, Ke, & Wu, 42 2016;Zheng et al., 2012). Generally, pipeline failure or destruction caused by landslides is much more 43 deleterious than the landslides themselves, which makes it important to research the risk assessment of  Based on the theory of the LM-BP neural network, a standard sample matrix was developed using the 71 interpolation theory after an analysis of the distribution characteristics of landslides that occurred in the 72 study area was performed and a regional landslide hazards assessment was completed. Considering the 73 interaction between landslide disasters and the pipeline itself, the pipeline vulnerability evaluation in the 74 landslide area was realized using the entropy weight method. This paper established a risk assessment 75 model and methods for assessing landslide geological hazards of oil pipelines by comprehensively 76 utilizing GIS and RS technology, which together improved the quantitative degree of the assessment.

78
The study area was Guangyuan City in the Sichuan province, which was further restricted to the area 79 from 105°15´ to 106°04É and 32°03to 32°45Ń, straddling 19 townships in five counties from south to Longmen Mountain's Piedmont Fault Zone) make the area geologically unstable and prone to frequent 90 geological hazards (Shiyuan Li et al., 2012). Guangyuan, through which the pipeline passes, has a high 91 incidence of landslides, some of which have happened 300 times in the Lizhou and Chaotian districts 92 (Zhang, Shi, Gan, & Liu, 2011). In this area, landslide geological hazards seriously threaten the safe 93 operation of the LCC oil pipeline.

105
The location of the middle line of the pipeline was detected through the direct connection method (i.e., 106 the transmitter's output line was directly connected to the metal pipeline) using an RD8000 underground 107 pipeline detector. Pipeline midline coordinates were measured using total network Real Time Kinematic 108 technology, and simultaneously, the coordinates of the pipe ancillary facilities (including test piles, 109 mileage piles and milestones) were acquired. Mileage data obtained through inner pipeline detection was 110 derived from the China Petroleum Pipeline Company. Division precision and the scale of the slope unit (i.e., the basic element for a regional landslide hazard 114 assessment) were in keeping with the results of the evaluation (Qiu, Niu, ZhaoYannan, & Wu, 2015). A 115 total of 315 slope units were divided using hydrologic analysis in ArcGIS (v. 10.4) (Fig. 2a). The 116 irrational unit was artificially identified and modified by comparing GF-1 satellite remote sensing 117 images. Boundary correction, fragment combination and fissure filling were used for modification.

118
The object of the pipeline vulnerability assessment in the landslide area was the pipeline. Considering 119 both previous research and the particulars of the research object, we used a comprehensive 120 segmentation method based on GIS to divide the pipelines in our study. A total of 180 pipes were 121 divided in the study area, of which the longest was about 1.7 km, and the shortest was only about 10 m 122 ( Fig. 2b). showed a significant correlation between MAP and CVAR (R = 0.99) and between NDWI and NDVI (R 132 = 0.87). Based on correlation and standard deviation, CVAR and NDWI were eliminated from the 133 original evaluation system for landslide hazard assessment in the pipeline area (Table 1).

134
Generally, the evaluation index of pipeline vulnerability as it relates to the relationship between a pipeline 135 and its surrounding environment is rarely considered. The evaluation indicators in this paper were refined 136 to include pipeline parameters and the spatial relationship between a pipeline and landslide. The pipelines 137 in the study area were based in mountainous areas and had been running for many years. All of these 138 pipelines consisted of high-pressure pipes that were made of steel tubes and had a diameter of 610 mm 139 for conveying oil. In keeping with the theory of the entropy weight method, these indicators (e.g., 140 pressure, materials, diameter and media) were not included in the final evaluation system used to 141 determine pipeline vulnerability.

159
The training samples and test samples were evaluated using similar construction methods but with 160 different sample sizes. Finally, the indicator data was normalized, it was entered into the LM-BP neural 161 network for simulation and 315 slope unit landslide hazard values were output. The vulnerability evaluation model of pipelines in the landslide area was established using the entropy 164 weight method, which overcame the shortcomings of the traditional weight method that does not consider 165 the different evaluation indexes and the excessive human influence on the process of evaluation (Gao, 166 Li, Wang, Li, & Lin, 2017;Pal, 2014). Pipeline defect density was obtained from the pipeline internal 167 inspection data, which consisted of both mileage data that needed to be converted into three-dimensional 168 coordinate data and pipeline center line coordinate data obtained through C# programming. In addition, 169 the main slide direction of the landslide was replaced by the slope direction that was extracted by DEM.

170
The coordinate azimuth of the pipe section was extracted using the linear vector data of each pipe section, 171 and the angle between the pipeline and the slope was calculated using the mathematical method. The 172 calculation process was solved in the VB language on ArcGIS using second development functions. safe section (low hazard) was located in the central part of the study area. The dangerous (high hazard) 188 section was located north and south (Fig. 4). In the study area, most of the exposed rock was dominated 189 by shale, which belonged to the easy-slip rock group.

190
Average altitude ranged from 450 m to 1400 m, and the relative height difference was greater than 80 191 m, with the slope between 15° and 35°. Based on an overlay analysis of historic landslides within the 192 study area, and hazard zonation maps, we surmised that the probability of landslides in the study area 193 was extremely high, and that 87% of the landslides occurred in the medium-, high-, and extremely high-194 hazard areas. Among these landslides, three were located in low-hazard areas, which accounted for 13% 195 of the landslide disaster sites, five occurred in medium-hazard areas (accounting for 21.7% of disaster 196 sites), seven occurred in high-hazard areas (accounting for 30.4% of sites) and eight occurred in 197 extremely high-hazard areas (accounting for 34.8% of sites). The evaluation results were found to 198 accurately reflect the trends and rules of distribution of landslides in the study area. The number and area 199 of slopes in high-hazard and extremely high-hazard areas accounted for about 70% of the total (Table 3).

200
The probability of landslide occurrence in the study area was generally high, which was consistent with 201 the fact that the region was landslide-prone. The equal interval of 0.25 was used to divide the pipeline vulnerability level into four grades to obtain 204 the pipeline vulnerability zonation of the study area (Fig. 5). The pipeline in the northern part of the study 205 area was given a low vulnerability grade, while the situation in the south of the region is more serious.

206
The number, length and percentage of pipeline segments with different grade vulnerabilities are shown 207 in Table 4. The number and length of pipeline segments in highly vulnerable areas (III) and extremely 208 vulnerable areas (IV) accounted for about 12% of the total. Large or huge landslides were common in areas that we categorized as extremely high risk, which we 225 defined as those that were geologically evolving or had experienced obvious deformations within the last 226 2 years with still visible cracks. These pipelines were subject to dangers at any time, as the pipelines 227 within the areas prone to landslides were found to contain many defects or extensive damage. These Hicher, 2015). The exposed rocks in this area were mainly shale and belonged to the sliding-prone 240 rock group. Rock type and interlayer structure were found to be important internal indicators that a 241 landslide could occur (Guzzetti, Cardinali, & Reichenbach, 1996;Xiang et al., 2010;Xin, Chong, & 242 Page 7 Nat. Hazards Earth Syst. Sci. Discuss., https://doi.org/10.5194/nhess-2018-360 Manuscript under review for journal Nat. Hazards Earth Syst. Sci. Discussion started: 6 December 2018 c Author(s) 2018. CC BY 4.0 License. Dai, 2009). The distance between the fault and the pipeline in the section was about 2 km with a 243 NDVI of about 0.75 and MAP of about 970 mm. Faulted zones and nearby rock and earth masses 244 that were destroyed in a geologic event reduced the integrity of a slope, and the faults and important 245 groundwater channels could also cause deformation and damage of a slope (Yinghui Liu, 2009). The 246 pipelines in these areas exhibited many defects. Most pipelines passed through the slope in an inclined 247 or horizontal way, an attribute that typically increased the risk of a landslide occurring.

248
In high-risk areas, small or moderate landslides commonly occurred in areas that we categorized as 249 high risk. They were in deformation, or had obvious deformation recently (within 2 years), such as 250 obvious cracks, subsidence or tympanites on the landslide and even shear. The pipelines in these areas 251 had defects and were buried at a shallow depth. If a landslide occurred in this pipeline area, it could cause 252 pipe suspension, floating and damage. It could also contribute to a small to moderate leakage of the 253 medium. However, damaged pipes can be welded or repaired. Monitoring is critical in high-risk areas.

254
In our study, the pipeline high-risk area was defined by the following areas: (1) K635-K642). This area was located in the south of the pipe, which was buried in the study area.

257
The altitude of the study area was between 450 m and 800 m, the relative elevation difference was over 258 100m and the slope was between 15° and 40°. Most of the outcrops in this area were quartz sandstone, 259 which belonged to the easy-sliding rock group. The pipes in this area were about 2.5 km away from faults.

260
The NDVI was about 0.6 to 0.8, and MAP was about 970 nm. Pipes showed many defects, most of them 261 either crossing the slope or lying in the center of slope. All of the above factors provided sufficient 262 conditions for the formation of landslide.

263
In the medium-risk areas, only small landslides were found to occur, and we observed no sign of 264 deformation. But through the analysis of geological structure, topography and landform, we found the 265 area to demonstrate a tendency for developing landslides. The pipes in this risk area exhibited almost no 266 faults and were buried deep beneath the ground. However, under bad conditions, the landslides in these 267 areas could also affect the pipes' safety, causing the pipes to become exposed or deformed. These areas 268 need simple monitoring. For our study, medium-risk areas were defined as follows: (1)

272
In the low-risk areas, landslides didn't occur under ordinary conditions, but they could occur if a strong 273 earthquake hit or if the area experienced continuous or heavy rain. The pipes in low-risk areas showed 274 no defects and were buried very deep. They were also located far away from areas affected by landslides.

275
Therefore, landslides in these areas caused no obvious damage to the pipes, and few threatened the safety 276 of pipes. However, regular inspection is necessary to ensure that the pipes continue to operate safely. The 277 pipe low-risk area were defined as follows: (1)              The landslide won't happen under ordinary conditions, but it will occur when strong earthquake, long continuous rain or extremely heavy rain happened.
The pipes in low risk areas have no any defects and buried very deep. Meanwhile, they are far away from the area affected by landslide.
Landslides have no obvious damage to the pipes, and few threats to pipes' safety.

Medium (II)
Small landslide mainly occur, and no sign of deformation. But through analyzing geological structure, topography and landform, there is a tendency of landslide.
The pipes in risk areas have almost no faults and buried deep. However, under bad condition, the landslide may also affect the pipes' safety.
The landslide may make the pipes exposed or deformation.
simple monitoring

High (III)
Landslides are most in medium-model and little-model, and they are in deformation, or have obvious deformation recently, such as obvious cracks, subsidence or tympanites on the landslide and even shear.
The pipeline has defects, and buried shallow. Once landslides occurred in the pipeline area, pipes' safety will be threatened The safety of pipeline will be threatened and may suffer from pipe suspension, floating, and damage etc. Therefore it will contribute to a small amount of medium leakage. Fortunately, the pipe can be welded or repaired.

Main monitoring
Extremely high (IV) Large or huge landslide is common in the area with extremely high risk, which is changing or has experienced obvious deformation recently with visible cracks.
The pipelines are subject to dangers at any time as the pipelines within the area prone to landslide have been spotted with many defects or much damage.
There are great threats, for example pipeline rupture or break and may lead to considerable leakage of media or serious deformation even transportation failure.