Since the beginning of the 20th century, with the gradual implementation of the national strategy of western development, the development of transportation in western China has been rapidly promoted, and tunnel projects and underground projects built in karst areas in western China have sprung up like mushrooms after rain. Due to the special topography and complex geological conditions in these areas, the tunnels built in these karst areas are extremely complicated and difficult to predict [1]. Therefore, the construction of large railway tunnels in areas with high karst development will inevitably encounter a large number of karst disasters and adverse geological problems. In the process of construction, disasters such as water inrush, mud outburst, collapse and gas burning often occur, which bring great difficulties to the site construction and cause a large number of casualties and a large number of engineering economic losses. Combined with the existing technical and economic conditions, various unfavorable geological disasters in the process of tunnel construction have become a difficult problem that has not been well solved by the engineering geology and tunnel engineering circles at home and abroad [4, 9]. In view of the above situation, in order to ensure the safety of tunnel construction in highly developed karst areas, reduce the losses caused by unfavorable geological disasters such as water inrush and mud outburst, improve the prediction accuracy and reduce the project cost, taking Yinjiayan Tunnel in bid section B of Xuda Railway as the engineering background, the comprehensive advanced geological prediction method is adopted, and the geological radar is used to continuously track and predict the highly developed karst tunnels, which effectively guides the tunnel construction. 1 Project Overview The entrance of Yinjiayan Tunnel is located in Longshan Town and Shiping Township, Gulin County, with a central mileage of K7 1+658 and a total length of 3,050 m, so the traffic is extremely inconvenient. This area belongs to the transition zone between the southern margin of Sichuan Basin and Yunnan-Guizhou Plateau, and belongs to the low-mid-mountain tectonic erosion landform, with high mountains and deep valleys and large topographic relief. Generally, the mountains are low in the north, high in the south and low in the east, mostly in the east-west direction, which is basically consistent with the tectonic line. The elevation of this area is mostly 350 ~ 2000m, and the relative height difference is 200 ~ 700 m. The valleys basically develop along the structural line, and the steep slopes are generally V-shaped valleys and U-shaped parts. Wide Gu Duo is a farmland, with obvious karst landform and serious surface dissolution in limestone areas. Typical limestone landforms such as stone forest and stone pillars are characterized by many dissolved depressions, karst caves, underground rivers and beaded depressions. Vegetation is well developed on hillsides and ridges, and most areas are sparsely populated. There are many valleys in the mountains, and most of them are dry land and paddy fields. The tunnel site passes through Permian Liangshan Formation and Qixia Formation limestone (P 1l+q), Silurian Hanjiadian Formation shale, limestone, mudstone, sandstone (S 1-2HN), Shiniulan Formation and Longmaxi Formation shale and limestone (S 1l+s). The main fault structure in the tunnel is Liujiagou fault (steep fault, which affects the bandwidth) with a gentle dip angle of about 20. The surrounding rocks except the entrance are mainly Grade III and IV, with developed joints and fissures. There are colluvial rubble and dangerous stones produced by differential weathering at the entrance slope along the bedding plane and at the exit. Limestone (especially Qixia Formation in Liangshan) has a strong karst development. There are many caves (2 ~ 6m deep) at the entrance and middle of the exploration hole, and the water level is high (higher than the cave body). The buried depth of the tunnel is shallow, and the upper rock mass is mostly V-class surrounding rock or water-rich rock mass, so the stability is poor. To sum up, the main unfavorable geological phenomena of Yinjiayan Tunnel are karst and fault fracture zones and their influence zones determined by stratum lithology, as well as possible collapse in other fractured rock masses and possible falling blocks in intact rock masses. 2 Comprehensive advanced geological prediction of karst tunnels At present, more and more attention has been paid to the advanced geological prediction of tunnels, and there are many methods, but each has its own characteristics. It is very important to choose suitable methods according to different geological conditions and environments. In order to better understand the tunnel geological conditions and reduce the construction risk, on the basis of geological analysis, combined with the comprehensive geological prediction method, the comprehensive geological advance prediction is made for the entrance section of Yinjiayan tunnel DK70+803 ~ DK70+83 1. 2. 1 Geological analysis method and geophysical detection method At present, the geological situation in front of tunnel face is regarded as an important link in tunnel safety production in tunnel construction at home and abroad. According to a large number of engineering practices, there are many methods to detect bad geological bodies in front of tunnel face. According to whether instruments are used or not, it can be divided into two categories: geological analysis method and geophysical prospecting method, among which geological analysis method is a basic method in tunnel advance prediction, and the common ones are: ground geological survey, tunnel excavation geological logging, advanced drilling, fault prediction method and geological experience method. Geophysical detection methods are mainly based on electromagnetic reflection wave theory, and the main instruments are TSP tunnel seismic detection, geological radar detection and transient electromagnetic method [1 1- 12]. The above advanced geological prediction methods and the present situation at home and abroad have their own advantages and disadvantages, so improving the accuracy and timeliness of prediction is still an academic problem to be solved urgently in tunnel engineering geology at home and abroad. Therefore, in order to guide tunnel construction more timely and effectively and improve information construction technology, it is urgent and necessary to study comprehensive advanced geological prediction technology. The advantages and disadvantages of different geological advance prediction methods are limited by different geological environment conditions. Table 1 gives the prediction range, theoretical basis and evaluation accuracy of different prediction methods, and Table 2 gives the comparison of advantages and disadvantages of different prediction methods [2- 10]. According to different disaster classification and grade sections, combined with geological conditions, different forecasting methods are reasonably used to forecast the geological conditions in front of the tunnel face, including the flow chart of comprehensive advanced geological forecasting for karst tunnels (Figure 1) and the working procedure block diagram of comprehensive advanced geological forecasting for karst tunnels (Figure 2). 2.2 fuzzy neural network prediction method in the process of karst tunnel construction, the number of possible geological disasters, the location of geological disasters, the scale of a single geological disaster and its impact on the project are random and uncontrollable, so it is irregular. Therefore, the fuzzy neural network method is used to establish a prediction model for the target tunnel in order to achieve better prediction purposes. Firstly, the main factors for disaster risk assessment of high-risk karst tunnels are determined, that is, the factor set Y = {Y 1, Y2, Y3, …, Yn}, among which the main influencing factors for engineering geological disaster risk of karst tunnels are Y 1- geological survey data in design stage; Y2—— the development characteristics of tunnel face dissolution; Y3—— development characteristics of joints and fractures in working face; Y4—— tunnel heading fault investigation; Y5—— characteristics of groundwater seepage in tunnel face; Y6—— strength index of rock mass in working face; D7—— tunnel heading lithology; D8—— reflection spectrum of geological radar. In view of the randomness and uncontrollability of the combination of risk factors in karst tunnels, the whole comprehensive advance prediction process still needs to be further improved and perfected to make it more operable. In this paper, the adaptive fuzzy neural network (ANFIS) method is used to establish the fuzzy neural network comprehensive prediction model of the above influencing factors of karst tunnel, and the comprehensive prediction is made [6, 14]. Schematic diagram of adaptive fuzzy neural network (ANFIS) model structure (Figure 3). 3 Comprehensive geological advance prediction technology and application examples of tunnel construction in karst area Because of the complex geological environment in Yinjiayan tunnel construction area, it is necessary to find out the unfavorable geological factors such as faults, cracks, karst and groundwater in the rock stratum in front of tunnel excavation. Combined with the comprehensive geological advance prediction method and working system proposed in this paper, the tunnel is continuously analyzed and predicted during the tunnel construction. Judging from the prediction results, the prediction accuracy is high, which provides a basis for changing the construction scheme in time, effectively reduces the occurrence of geological disasters caused by tunnel construction, ensures the safe production of construction, and also brings economic benefits to engineering construction. 3. 1 Establish a comprehensive prediction and risk assessment model for geological disasters. According to various influencing factors of karst tunnel, an evaluation model based on adaptive fuzzy neural network (ANFIS) method is established, that is, the actual influencing factors in the construction process of karst tunnel are comprehensively evaluated to determine the types and existence of disasters. That is, the existence of a specific disaster; (2) the scale of the disaster; ③ The scale of the disaster is small; Some disasters don't exist. Take karst as an example, as shown in Table 3. The types of disasters can be roughly divided into eight categories: karst cave, gas, mud outburst, water inrush, broken rock mass, weak rock mass, landslide and fault. According to the eight defined disaster types, the disaster grade is determined by the prediction and evaluation model, and the disaster risk grade is divided into four types: I, II, III and IV. See Table 4 for details. 3.2 Comprehensive geological advance forecast preliminarily judges that this forecast is located at DK70+803 at the entrance of the tunnel, 28m ahead of the forecast heading. SIR-3000 geological radar is used in the field test, and the forecast is realized by combining the ANFIS comprehensive advanced forecast method proposed in this paper. Among them, the interpretation results of radar data collected on site are shown in Figure 4, and the training samples of ANFIS are taken from the samples constructed according to the standards in Table 3. Combined with the actual detection results of Yinjiayan tunnel shown in Table 5, the training samples of ANFIS model are established. The judgment result by using sample data is ideal, which is completely consistent with the result obtained by the model. After the model network sample is debugged, the rock mass karst in this prediction section is comprehensively predicted, and the results are shown in Table 6.
For more information about project/service/procurement bidding, and to improve the winning rate, please click on the bottom of official website Customer Service for free consultation:/#/? source=bdzd