A dynamic monitoring and early warning method for loess-mudstone interbedded landslide

Abstract

The present application relates to a kind of loess-mudstone interbedded landslide dynamic monitoring early warning method, setting mark feature point in the sliding surface that loess-mudstone interbedded landslide possibly occurs sliding, make the mark feature point can slide along with sliding body along sliding surface;Monitoring the sliding change condition of sliding mark feature point, and determine the information of the sliding surface of sliding loess-mudstone interbedded landslide;Establish the neural network algorithm of sliding loess-mudstone interbedded landslide, and sliding neural network algorithm is according to historical loess-mudstone interbedded landslide data real-time output sliding loess-mudstone interbedded landslide sliding data, and the sliding data are compared with monitoring threshold value, when exceeding sliding monitoring threshold value, send alarm.The present application has the advantages that: effectively realize the monitoring of loess-mudstone interbedded landslide, improve the accuracy of monitoring data;Realize the prediction and early warning of loess-mudstone interbedded landslide, can be managed according to the prediction result before the accident of influencing safety occurs.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of landslide monitoring, and in particular to a dynamic monitoring and early warning method for loess-mudstone interbedded landslide. BACKGROUND

[0002] Landslide refers to the soil or rock on the slope, which is affected by river erosion, groundwater activity, rainwater infiltration, earthquake and artificial cutting slope, and under the action of gravity, slides along a certain weak surface or weak zone, and the whole or dispersedly slides downward.

[0003] In the loess area, according to the material composition of the landslide body (slide body) and the development position of the sliding surface, the landslide types are mainly divided into three types, which are loess layer landslide, loess contact surface landslide and loess-mudstone interbedded landslide. Among them, the loess-mudstone interbedded landslide occurs in the slope section where the underlying rock layer is outward inclined, and the rock layer inclination is generally between 10°-20°. Under the action of gravity, the slope body slowly shears and slides along the relatively weak layer or interlayer in the underlying rock layer. Since the loess-mudstone interbedded landslide generally has low sliding speed, small sliding distance and small thickness range, the monitoring of this type of landslide is mainly periodical monitoring, that is, the state change of the landslide is manually monitored at certain time intervals, and the state of the landslide is compared with the state of the landslide obtained in the previous monitoring. When the change of the landslide is small and the overall sliding range, landslide influence and other factors are controllable, the periodical monitoring is continued. If the influence is large, the monitoring frequency is increased accordingly, and the dynamic development law is grasped in time.

[0004] Therefore, the current monitoring of loess-mudstone interbedded landslide is still mainly based on the monitoring results for judgment. However, when affected by a large amount of rainfall, irrigation water and other unexpected factors, the sliding potential energy of the loess-mudstone interbedded landslide will be further stimulated, so that the slide body may continue to slide at a faster speed and a farther distance. The existing periodical monitoring method cannot timely understand the rapid change of the loess-mudstone interbedded landslide, resulting in an increase in safety hazards. SUMMARY

[0005] The present application aims to provide a dynamic monitoring and early warning method for loess-mudstone interbedded landslide according to the shortcomings of the prior art.

[0006] The present application is achieved by the following technical solutions:

[0007] A dynamic monitoring and early warning method for loess-mudstone interbedded landslide, characterized in that the dynamic monitoring and early warning method comprises the following steps:

[0008] The marking feature points are arranged in the sliding surface where the loess-mudstone interbedded landslide can slide, and the marking feature points are connected with the sliding body above the sliding surface, so that the marking feature points can slide along the sliding surface with the sliding body, and the marking feature points are arranged in at least a relatively weak layer or interbed in the underlying rock stratum of the loess-mudstone interbedded landslide.

[0009] The sliding change of the marking feature points is monitored, and the information of the sliding surface of the loess-mudstone interbedded landslide is determined, and the sliding surface includes a main sliding surface and other sliding surfaces.

[0010] The neural network algorithm of the loess-mudstone interbedded landslide is established, the neural network algorithm outputs the sliding data of the loess-mudstone interbedded landslide in real time according to historical loess-mudstone interbedded landslide data, and compares the sliding data with a monitoring threshold, and an alarm is given when the monitoring threshold is exceeded.

[0011] A plurality of marking feature points are arranged in the sliding body, and a plurality of marking feature points are arranged in a plurality of relatively weak layers or interbeds in the underlying rock stratum in a grouped form; the plurality of marking feature points are monitored in a group form, the marking feature points with the largest sliding change are obtained, and thus the main sliding surface of the loess-mudstone interbedded landslide is determined.

[0012] A plurality of marking feature points are arranged in the sliding body, and the plurality of marking feature points are segmental rods, the rods pass through each possible sliding surface, the segment length of the rod is matched with the layering condition of the sliding body, the segments of the rod can relatively slide in the sliding direction of the sliding body, and the main sliding surface and other sliding surfaces of the loess-mudstone interbedded landslide are determined according to the sliding condition of the segments.

[0013] The rod is provided with a displacement sensor, and the displacement sensor is used to monitor the relative displacement change between the segments of the rod.

[0014] The neural network algorithm is a BP neural network algorithm, including an input layer, a hidden layer and an output layer, the output layer includes geological stratum information of the loess-mudstone interbedded landslide, sliding surface information of the loess-mudstone interbedded landslide, and external factor influence information, wherein the external factor influence information includes natural factor influence information including rainfall and artificial factor influence information including irrigation and construction.

[0015] The comparison between the sliding data and the monitoring threshold is a hierarchical comparison of the data of the plurality of marking feature points.

[0016] When multiple sliding surfaces exist, the combined effect of the multiple sliding surfaces is considered, and the sliding data when multiple sliding surfaces slide simultaneously is compared with the monitoring threshold.

[0017] Several of the marked feature points are evenly spaced on the loess-mudstone interlayer landslide and arranged in blocks on the loess-mudstone bedding landslide.

[0018] The advantages of this invention are: it effectively monitors loess-mudstone bedding landslides, improves the accuracy of monitoring data, and enables targeted treatment of loess-mudstone interbedded landslides; it enables the prediction and early warning of loess-mudstone bedding landslides, allowing for treatment before safety-affecting accidents occur based on the prediction results, ensuring true preparedness; the device used has a simple and reasonable structure, is convenient for monitoring, and is suitable for widespread application. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the arrangement structure of Embodiment 1 of the present invention;

[0020] Figure 2 This is a schematic diagram of the structure of the rod in Example 1;

[0021] Figure 3 This is a schematic diagram of the arrangement structure of Embodiment 2 of the present invention. Detailed Implementation

[0022] The features and other related features of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments, so as to facilitate understanding by those skilled in the art:

[0023] like Figures 1-3 As shown in the figure, the labels 1-10 represent: sliding body 1, sliding body 2, sliding surface 3, sliding surface 4, underlying rock layer 5, rod 6, rod segment 7, rod segment 8, rod 9, and rod 10, respectively.

[0024] Example 1: In this example, the dynamic monitoring and early warning method for loess-mudstone interlayer landslides is used to realize the dynamic monitoring of this type of landslide, so as to accurately monitor the sliding situation of this type of landslide and predict the sliding situation. Then, the predicted situation is compared with the monitoring threshold to accurately assess the safety situation of the loess-mudstone interlayer landslide, so as to facilitate the corresponding treatment and disposal of the loess-mudstone interlayer landslide.

[0025] In this embodiment, with Figure 1The shown case describes the scheme of the present embodiment, which is consistent with most cases of loess-mudstone interbedded landslide, specifically including the sliding body 2 on the underlying rock 5 and the sliding body 1, wherein the sliding body 2 is mainly mudstone, and the sliding body 1 is mainly loess covering the mudstone. There is a sliding surface 4 between the sliding body 2 and the underlying rock 5, which is generally the main sliding surface, that is, the sliding body 2 and the sliding body 1 on the sliding surface 4 will slide downward along the inclination of the sliding surface 4, and there is a sliding surface 3 between the sliding body 1 and the sliding body 2, which will make the sliding body 1 further slide downward along the inclination of the surface of the sliding body 2. The sliding range of such loess-mudstone interbedded landslide generally includes two parts, one part is the sliding range of the sliding body 2 along with the sliding body 1 sliding downward along the sliding surface 4, and the other part is the sliding range of the sliding body 1 sliding downward along the sliding surface 3 on the sliding body 2.

[0026] Specifically, the dynamic monitoring and early warning method in the present embodiment mainly includes the following steps:

[0027] 1) Mark feature points at the sliding surfaces, i.e. the sliding surface 3 and the sliding surface 4, where the loess-mudstone interbedded landslide is likely to slide, which actually uses a rod 6. The rod 6 forms a connection with the sliding body 1 and the sliding body 2, which can be a clamping connection, an adhesive connection, or other existing connection modes that can enhance the connection performance between the rod 6 and the sliding body. The main purpose is to make the part of the rod 6 corresponding to the sliding body 1 or the sliding body 2 be able to slide synchronously with the corresponding sliding body, so as to accurately reflect the sliding condition of the sliding body 2 (together with the sliding body 1) on the sliding surface 4 and the sliding condition of the sliding body 1 on the sliding surface 3.

[0028] In the present embodiment, as shown in Figure 2 , the rod 6 includes a rod segment 7 and a rod segment 8, wherein the rod segment 7 is arranged above the rod segment 8, and the rod length of the rod segment 7 corresponds to the average thickness of the sliding body 1 to ensure that the rod segment 7 can protrude above the slope surface, and the rod length of the rod segment 8 corresponds to the average thickness of the sliding body 2, so that the rod segment 8 can be entirely embedded in the sliding body 2. In actual cases, the sliding body 1 and the sliding body 2 are generally non-uniform thickness strata, so the rod lengths of the rod segment 7 and the rod segment 8 can be adaptively designed according to the results of field exploration.

[0029] In the present embodiment, the rod segment 7 and the rod segment 8 are movable, that is, they can produce relative displacement between them to represent different sliding conditions between the sliding surface 3 and the sliding surface 4. In actual design, micro guide rails, mortise and tenon slides, etc. can be used for structure building. For example, as shown in Figure 2As shown in the right middle part, when the sliding amount of the rod segment 7 along with the sliding body 1 is greater than the sliding amount of the rod segment 8 along with the sliding body 2, the relative displacement between the two rod segments is generated. Meanwhile, the displacement sensor is also arranged on the rod 6, which can be a commercially available product in the prior art, and is used to monitor the displacement of the rod segment 7 and the displacement of the rod segment 8, so as to reflect the sliding conditions on the sliding surface 3 and the sliding surface 4 respectively.

[0030] 2) The monitoring range of the loess-mudstone interbedded landslide is divided into several blocks, and several rods 6 are uniformly arranged in each block, so as to accurately evaluate the sliding conditions in the whole loess-mudstone interbedded landslide range.

[0031] 3) The monitoring data of the displacement sensor on each rod 6 is collected, and the information of the sliding surface of the loess-mudstone interbedded landslide is determined according to the monitoring information, including the main sliding surface and other sliding surfaces. For example, the sliding surface 3 is generally the main sliding surface, and the sliding surface 4 is the other sliding surface, but the sliding surface 4 can also be the main sliding surface 4, and the sliding surface 3 can be the other sliding surface. Therefore, the actual situation of the sliding surface can be judged by the rod 6 in this embodiment, so as to improve the accuracy of the subsequent prediction and early warning.

[0032] 4) The neural network algorithm of the loess-mudstone interbedded landslide is established, the neural network algorithm outputs the sliding data of the loess-mudstone interbedded landslide in real time according to the historical loess-mudstone interbedded landslide data, and compares the sliding data with the monitoring threshold value, and issues an alarm when the monitoring threshold value is exceeded.

[0033] In this embodiment, the neural network algorithm adopts the BP neural network algorithm, which includes an input layer, a hidden layer and an output layer. The output layer includes the geological stratum information of the loess-mudstone interbedded landslide, for example, the loess-mudstone interbedded landslide mainly includes the sliding body 1 and the sliding body 2, and has two sliding surfaces, i.e. the sliding surface 3 and the sliding surface 4; the sliding surface information of the loess-mudstone interbedded landslide collected by the displacement sensor, i.e. the sliding amount of the sliding surface 3 and the sliding surface 4; and external factor influence information, which includes natural factor influence information including rainfall and human factor influence information including irrigation and construction.

[0034] The above information is mainly obtained through geotechnical exploration and information collection, and cooperates with the BP neural network algorithm to establish a corresponding database. The geological stratum information is mainly obtained by geotechnical exploration, the natural factor influence information in the external factor influence information can be obtained in real time through local weather forecast information and collection of local historical weather information, and the human factor influence information needs to be obtained by collecting various projects such as surrounding construction projects and farming conditions. The historical loess-mudstone interbedded landslide data can be obtained by statistically collecting existing data.

[0035] When the data of the loess-mudstone interbedded landslide to be monitored is input as an input layer into the BP neural network, the BP neural network outputs the result in combination with the historical loess-mudstone interbedded landslide data, and the output result represents the predicted sliding condition of the loess-mudstone interbedded landslide after a period of time (which can be artificially set) under the influence of external factors, thereby solving the problem of insufficient sliding prediction when encountering a large amount of rainfall, irrigation water and other unexpected factors.

[0036] 5) The output result of the BP neural network is compared with the monitoring threshold, thereby warning the sliding condition of the loess-mudstone bedding landslide. The comparison mainly includes two parts:

[0037] First, the data of the plurality of bars 6 is compared in stages, that is, because a plurality of bars 6 is arranged in blocks on the loess-mudstone interbedded landslide, by comparing the bars 6 in each block, it can be determined that a certain block of the loess-mudstone interbedded landslide may be the most significant sliding area, if the local area is treated, to a certain extent, the sliding condition of the entire loess-mudstone interbedded landslide can be alleviated, precise construction can be achieved without treating the entire loess-mudstone interbedded landslide, and construction efficiency is improved; similarly, if multiple areas are greater than the monitoring threshold, these areas can be treated at the same time.

[0038] Second, when there are multiple sliding surfaces, such as the sliding surface 3 and the sliding surface 4 in the embodiment, the superposition effect of multiple sliding surfaces is integrated, the sliding data when multiple sliding surfaces slide at the same time is compared with the monitoring threshold, thereby accurately predicting the sliding influence range of the landslide, facilitating effective evaluation of the safety condition of the loess-mudstone interbedded landslide, and when the sliding condition and the sliding influence range of the loess-mudstone interbedded landslide exceed the monitoring threshold, the technician can make timely, accurate and targeted treatment measures according to the development trend of the loess-mudstone interbedded landslide.

[0039] When the unmanned monitoring execution mode is adopted, if the development trend of the loess-mudstone interbedded landslide is close to the monitoring threshold, the technician is warned by sending a variety of ways such as short message, telephone, email, etc. to inform the technician to handle the loess-mudstone bedding landslide in the monitoring area immediately.

[0040] 6) With the continuous development of the loess-mudstone interbedded landslide, the displacement sensor collects the corresponding displacement data, and the measured data is continuously fed back to the database for landslide prediction correction, and after the completion of the entire monitoring project, the entire measured data is stored in the training set of the database, so that the BP neural network algorithm can use the continuously updated training set to circulate, thereby improving the accuracy of the result output.

[0041] Embodiment two: the difference between this embodiment and embodiment one is that different marking feature points are used in this embodiment, i.e. the rod 9 and the rod 10 are used as marking feature points, which are different from the rod 6 used as marking feature points in embodiment one. The rod 9 and the rod 10 in this embodiment are single whole rod with displacement monitoring function, in which the rod 9 is used to monitor the sliding condition at the position of the sliding surface 4, and the rod 10 is used to monitor the sliding condition at the position of the sliding surface 3.

[0042] Compared with embodiment one, the manufacturing cost of the rod in this embodiment is cheaper and the construction is more convenient, because only a rod with a certain length needs to be manufactured and then inserted into the stratum with a corresponding depth and attached to the sliding surface; but in terms of monitoring accuracy, since the local part of the rod 9 also passes through the sliding surface 3, it will inevitably be affected by the sliding potential energy of the sliding surface 3, resulting in that the displacement monitoring data of the rod 9 may be relatively large and also prone to skewing due to the rod length direction being affected by two different sizes of sliding potential energy, therefore, the connection performance between the rod 9 and the sliding body 2 should be further strengthened during the construction process.

[0043] Although the above embodiments have been described in detail with reference to the accompanying drawings for the purpose of illustrating the inventive concept and embodiments of the present application, those skilled in the art can recognize that various improvements and changes can still be made to the present application without departing from the scope defined by the claims, therefore, they are not described here.

Claims

1. A dynamic monitoring and early warning method for loess-mudstone interbedded landslide, characterized in that: The dynamic monitoring and early warning method comprises the following steps: A mark feature point is arranged in a sliding surface of a loess-mudstone interbedded landslide which is likely to slide, the mark feature point is connected with a sliding body above the sliding surface, so that the mark feature point can slide along the sliding surface with the sliding body, and the mark feature point is arranged at least in a relatively weak layer or interbed in an underlying rock layer of the loess-mudstone interbedded landslide; Sliding change of the mark feature point is monitored, and information of a sliding surface of the loess-mudstone interbedded landslide is determined, the sliding surface comprises a main sliding surface and other sliding surfaces; A neural network algorithm of the loess-mudstone interbedded landslide is established, the neural network algorithm outputs sliding data of the loess-mudstone interbedded landslide in real time according to historical loess-mudstone interbedded landslide data, and compares the sliding data with a monitoring threshold value, and an alarm is given when the monitoring threshold value is exceeded; A plurality of mark feature points are arranged in the sliding body, the plurality of mark feature points are sectional rods, the rods pass through each sliding surface which is likely to slide, sectional lengths of the rods are matched with layering conditions of the sliding body, the sections of the rods can slide relative to each other in a sliding direction of the sliding body, and the main sliding surface and the other sliding surfaces of the loess-mudstone interbedded landslide are determined according to sliding conditions of the sections.

2. The dynamic monitoring and early warning method of loess-mudstone interbedded landslide according to claim 1, characterized in that: The plurality of mark feature points are arranged in the sliding body in a grouped form and are arranged in a plurality of relatively weak layers or interbeds in the underlying rock layer respectively; The plurality of mark feature points are monitored in a group form, a group of mark feature points with the largest sliding change is obtained, and thus the main sliding surface of the loess-mudstone interbedded landslide is determined.

3. The dynamic monitoring and early warning method of loess-mudstone interbedded landslide according to claim 1, characterized in that: A displacement sensor is arranged on the rod, and the displacement sensor is used to monitor relative displacement change between sections of the rod.

4. The dynamic monitoring and early warning method of loess-mudstone interbedded landslide according to claim 1, characterized in that: The neural network algorithm is a BP neural network algorithm, and comprises an input layer, a hidden layer and an output layer, the output layer comprises geological stratum information of the loess-mudstone interbedded landslide, sliding surface information of the loess-mudstone interbedded landslide and external factor influence information, wherein the external factor influence information comprises natural factor influence information including rainfall and artificial factor influence information including irrigation and construction.

5. The dynamic monitoring and early warning method of loess-mudstone interbedded landslide according to claim 1, characterized in that: The comparison between the sliding data and the monitoring threshold value is hierarchical comparison of data of the plurality of mark feature points.

6. The dynamic monitoring and early warning method of loess-mudstone interbedded landslide according to claim 5, characterized in that: When there are a plurality of sliding surfaces, superposition effects of the plurality of sliding surfaces are comprehensively considered, and sliding data when the plurality of sliding surfaces slide simultaneously are compared with the monitoring threshold value.

7. The dynamic monitoring and early warning method of loess-mudstone interbedded landslide according to claim 1, characterized in that: The plurality of mark feature points are uniformly and interval arranged on the loess-mudstone interbedded landslide and arranged in a block form on the loess-mudstone interbedded landslide.

Images