Intelligent exploration system for oil exploration

Abstract

The invention belongs to the technical field of oil exploration, and discloses an intelligent exploration system for oil exploration, and the system comprises a data collection module which is used for collecting multi-source information data of an exploration region; the data processing module is used for preprocessing the multi-source information data to obtain a multi-source feature data set; converting the multi-source feature data set into a three-dimensional geological model by using a full-waveform inversion algorithm, and visualizing the three-dimensional geological model to obtain a three-dimensional geological map; the region prediction module is used for training according to the three-dimensional geological map to obtain an oil and gas reservoir region prediction model, and predicting an oil and gas reservoir distribution region based on the oil and gas reservoir region prediction model; the path planning module is used for formulating a preliminary oil drilling path according to the predicted oil and gas reservoir distribution area; collecting real-time drilling data, and dynamically adjusting the preliminary oil drilling path according to the real-time drilling data to obtain a final oil drilling path; the intelligent drilling process is achieved, human intervention is reduced, and the overall operation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of oil exploration, and more specifically, to an intelligent exploration system for oil exploration. Background Art

[0002] A patent with the publication number CN115559712A discloses an intelligent oil exploration system and its application method. The system includes: a detection module, an information transmission module, a central processing module, and an interaction module. The detection module is used to detect downhole data in real time, divide the downhole data into important data and secondary data, and store the secondary data. The information transmission module is used to transmit the important data to the central processing module. The central processing module is used to analyze the important data and correct the drilling parameters in real time. The interaction module is used to feedback the downhole situation to the staff in real time and perform manual correction according to the downhole situation. In this application, a relay station is added to ensure the accuracy and effectiveness of downhole signal transmission. At the same time, the data is classified according to the actual working conditions on site, and the important data is analyzed in real time to correct the drilling parameters. At the same time, considering the error of intelligent operation, this application adopts a combination of manual and intelligent methods to process the on-site situation.

[0003] The existing oil exploration systems mainly have the following main problems:

[0004] The elastic wave equation is not adopted, which may not accurately reflect the propagation behavior of seismic waves in complex underground media; ignoring the influence of factors such as wave velocity, density, and medium heterogeneity may lead to a large difference between the simulation results and the actual observed waveforms; the optimization algorithms such as gradient descent are not adopted, and the parameters in the three-dimensional geological model cannot be effectively adjusted to make the simulated waveform gradually approach the actual observed waveform; without the propagation operator constraint formula, the influence of the source point and the receiving point on the propagation operator cannot be effectively adjusted. The influence of different source points and receiving points may be treated equally, resulting in too much contribution from some irrelevant or noisy receiving points to the final model, interfering with the optimization process of the model; not considering the dynamic adjustment of the weight factor of the receiving point may ignore the time shift problem of the received waveform. Receiving points with a large time shift usually contain more important wave propagation information, but if the weights of these receiving points are not high enough, it may lead to the loss of important information or error accumulation, affecting the accuracy of the final model; dynamically adjusting the weight factor of the receiving point may ignore the time shift problem of the received waveform. Receiving points with a large time shift usually contain more important wave propagation information, but if the weights of these receiving points are not high enough, it may lead to the loss of important information or error accumulation, affecting the accuracy of the final model;

[0005] Without discretizing the reservoir distribution area into voxels and providing multi-source information for each voxel, the model may not fully reflect the heterogeneity and complexity of the underground geology; the path cost function is optimized only based on the distance between nodes, while ignoring geological factors such as rock hardness and formation pressure, which may lead to the selection of inappropriate paths; the actual challenges that may be encountered during drilling are not quantified through a comprehensive path cost formula, which may lead to path planning that does not fully consider these challenges; in the path search process, if the remaining path cost from the target node to the current node is not reasonably predicted, the path selection may not be forward-looking. Simply relying on the cost of the current path may miss the actual optimal path, resulting in the inability to effectively control drilling costs and time;

[0006] Without using measurement while drilling technology to collect drilling data in real time, the dynamic changes in the drilling process cannot be reflected in real time; without optimization based on real-time data, the drilling path may not be adjusted in time, and the geological conditions may change during the drilling process; if the path planning is not updated in time, it may lead to the selection of unsuitable paths during the drilling process, increase drilling time and cost, and even lead to the failure to achieve the predetermined goals; without using path risk calculation and risk assessment formulas, it is impossible to comprehensively consider key factors such as friction coefficient, drill bit pressure, drilling speed, etc., which may lead to insufficient assessment of the safety of the drilling path; it is impossible to optimize the drilling path and parameters through real-time data, resulting in a significant reduction in drilling efficiency.

[0007] In view of this, the present invention proposes an intelligent exploration system for oil exploration to solve the above problems. Summary of the invention

[0008] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: an intelligent exploration system for oil exploration, comprising:

[0009] Data acquisition module, used to collect multi-source information data of the exploration area;

[0010] The data processing module is used to pre-process the multi-source information data to obtain a multi-source feature data set; convert the multi-source feature data set into a three-dimensional geological model using a full waveform inversion algorithm, and visualize the three-dimensional geological model to obtain a three-dimensional geological map;

[0011] The regional prediction module is used to obtain the regional prediction model of oil and gas reservoirs according to the three-dimensional geological map training, and predict the distribution area of ​​oil and gas reservoirs based on the regional prediction model of oil and gas reservoirs;

[0012] The path planning module is used to formulate a preliminary oil drilling path according to the predicted oil and gas reservoir distribution area; collect real-time drilling data, dynamically adjust the preliminary oil drilling path according to the real-time drilling data, and obtain the final oil drilling path;

[0013] The communication transmission module is used to transmit the final oil drilling path to the intelligent exploration terminal in real time through satellite communication technology; each module is connected by wired and / or wireless means.

[0014] Furthermore, the multi-source information data includes geophysical data, seismic wave data and geological data; the geophysical data includes gravity, magnetism, resistivity, electromagnetic method and natural potential; the seismic wave data includes seismic wave reflection data, seismic wave propagation velocity, seismic wave amplitude and seismic wave frequency; the geological data includes core samples, surface sediment samples and fluid samples.

[0015] Furthermore, the method of preprocessing the multi-source information data to obtain the multi-source feature data set includes:

[0016] By using a density clustering algorithm, outliers existing in the multi-source information data including geophysical data, seismic wave data and geological data are identified and eliminated, thereby obtaining a geophysical feature data set, a seismic wave feature data set and a geological feature data set;

[0017] The geophysical feature data set, seismic wave feature data set and geological feature data set are feature extracted through principal component analysis, and the geophysical feature data set, seismic wave feature data set and geological feature data set after feature extraction are normalized by standard deviation and converted into standard normal distribution with mean 0 and standard deviation 1 to obtain normalized geophysical feature data set, seismic wave feature data set and geological feature data set; the normalized geophysical feature data set, seismic wave feature data set and geological feature data set are fused through a weighted model to obtain a multi-source feature data set.

[0018] Furthermore, the method of converting a multi-source feature data set into a three-dimensional geological model using a full waveform inversion algorithm includes:

[0019] S41. The three-dimensional geological model is preset as m. The three-dimensional geological model m includes two physical properties of underground media, which are seismic wave velocity and density. The distribution of the seismic wave velocity and density in the three-dimensional geological model is preset as follows: m(x,y,z)=(v′(x,y,z),ρ(x,y,z)); wherein v′(x,y,z) is the seismic wave velocity at a certain position (x,y,z) in the three-dimensional geological model; ρ(x,y,z) is the density at a certain position (x,y,z) in the three-dimensional geological model; (x,y,z) is the position coordinates where seismic wave propagation occurs in the three-dimensional geological model;

[0020] S42. According to the current three-dimensional geological model, the propagation process of seismic waves is simulated by the elastic wave equation; the elastic wave equation is: in, is the second derivative of the seismic wave displacement u(x,t) with respect to time t, i.e., the acceleration of the seismic wave; u(x,t) is the seismic wave displacement at a certain position (x,y,z) in the three-dimensional geological model and at time t; c 2 (x,y,z) is the wave velocity function, representing the propagation velocity of seismic waves at each point in the three-dimensional geological model; is the Laplace operator, representing wave propagation in the three-dimensional geological model; t is the time variable, representing the evolution of the change of the wave during propagation over time.

[0021] S43. The simulated waveform of the seismic wave is obtained by solving the elastic wave equation through the finite difference method, and the simulated waveform of the seismic wave is: b pe = F[m,(x ce ,y ce ,z ce ),(x a ,y a ,z a )]; where, b pe is the simulated waveform data of the seismic wave; F is the propagation operator of the seismic wave; (x ce ,y ce ,z ce ) is the source position coordinate of the seismic wave; (x a ,y a ,z a ) is the receiving point position coordinate of the seismic wave;

[0022] S44. The propagation operator of the seismic wave is restricted by the propagation operator restriction formula, and the propagation operator restriction formula is: where, F′ is the propagation operator of the seismic wave after restriction; N su is the number of source points of the seismic wave; N re is the number of receiving points of the seismic wave; ∈ is a constant controlling the influence of the source point and the receiving point on the propagation operator;

[0023] S45. Construct an L2 norm loss function to quantify the difference between the simulated waveform data of the seismic wave and the actual observed waveform data; the L2 norm loss function is: where, J(m) is the L2 norm loss function; b os ((x a ,y a ,z a ),t a ) is the actual observed waveform data of the seismic wave at the a-th receiving point; b pe ((x a ,y a ,z a ),t a, m) is the simulated waveform data of seismic waves calculated through the three-dimensional geological model m; t a is the time shift of the a-th receiving point; ω a is the weight factor for each receiving point; M is the total number of receiving points of seismic waves; a is the index of the receiving point, a = 1, 2,..., M; the weight factor ω of each receiving point is dynamically adjusted and designed through the weight factor adaptive formula a ;

[0024] S46. Update the parameters of the three-dimensional geological model through the gradient descent update formula. The gradient descent update formula is: where, m k+1 is the three-dimensional geological model at the (k + 1)-th iteration; m k is the three-dimensional geological model at the k-th iteration; γ is the learning rate; is the gradient of the L2 norm loss function J(m) with respect to the three-dimensional geological model m k ;

[0025] S47. Preset the loss threshold of the L2 norm loss function, and continuously update the three-dimensional geological model through repeated iterations until the L2 norm loss function is less than or equal to the preset loss threshold of the L2 norm loss function, and then stop to obtain the final three-dimensional geological model m * .

[0026] Furthermore, the method for dynamically adjusting and designing the weight factor ω of each receiving point through the weight factor adaptive formula a includes:

[0027] The weight factor adaptive formula is: where, T max is the maximum time of seismic wave propagation; T M is the total time required to receive seismic waves.

[0028] Furthermore, the method for obtaining the three-dimensional geological map includes:

[0029] Use different filters in ParaView software to cut the three-dimensional geological model according to any geometric shape, view different parts of the three-dimensional geological model, create a section, and display the slice passing through the model to view the geological structure at different depths underground; extract the surface of the three-dimensional geological model from the three-dimensional grid data to view the external form of the underground structure; set different colors for different regions according to seismic wave velocity and density to distinguish different underground rock layers; adjust the transparency of different parts to view the internal structure of the three-dimensional geological model, and present the underground hierarchical structure by adjusting transparency and color, and finally obtain the three-dimensional geological map.

[0030] Furthermore, the training method of the oil and gas reservoir area prediction model includes:

[0031] Divide the dataset into a training set, a validation set, and a test set; construct a reservoir area prediction model, which includes an input layer, a 3D convolutional layer, a 3D pooling layer, a fully connected layer, and an output layer; the input data is a historical 3D geological map, and the output data is the reservoir distribution area; use ReLU as the activation function; the reservoir area prediction model is a 3D convolutional neural network model;

[0032] Use the mean squared error as the loss function of the model to measure the difference between the predicted value and the actual value of the model; use the training set to train the model, and minimize the loss function through the Adam optimizer; use the validation set to evaluate the performance of the reservoir area prediction model, and calculate the accuracy index to measure the performance of the model; tune the hyperparameters of the model, calculate the gradient of each hyperparameter through the backpropagation algorithm, and optimize the model hyperparameters through the gradient descent method to improve the model performance;

[0033] Evaluate the performance of the model in the prediction task through the test set. When the performance of the model in the prediction task reaches the preset performance threshold, stop the test to obtain the reservoir area prediction model; use the trained reservoir area prediction model to predict the current 3D geological map to obtain the reservoir distribution area.

[0034] Furthermore, the method for obtaining the preliminary oil drilling path includes:

[0035] Discretize the reservoir distribution area into Q voxels; a voxel is a small volume unit in the reservoir distribution area, containing multi-source information data corresponding to the reservoir distribution area; use the spatial position of each voxel as a node and the feasible path between adjacent voxels as an edge to construct a graph structure; the graph structure is G=

[0036] {V, E}; where V is the set of nodes; E is the set of edges;

[0037] Preset the initial node as v s , and the target node as v o ;

[0038] Define a path cost function to represent the cumulative cost from the initial node to the current node; the path cost function is where v i is the current node; v j is the previous adjacent node of the current node v i ; g(v j ) is the cumulative cost from the initial node to the node v j ; w ij is the path cost between adjacent nodes, that is, the current node v i and the previous adjacent node v jThe path cost between; i and j are indices of node numbers; N′ is the set of adjacent nodes of the current node v i ;

[0039] The path cost w between the adjacent nodes ij is obtained through the comprehensive path cost formula; the comprehensive path cost formula is w ij =α 1 ·d(v i , v j ) + α 2 ·|H(v i ) - H(v j )| + α 3 ·|P(v i ) - P(v j )|; where d(v i , v j ) is the distance between the current node v i and the previous adjacent node v j ; H(v i ) is the rock formation hardness of the current node v i ; H(v j ) is the rock formation hardness of the previous adjacent node v j ; P(v i ) is the formation pressure of the current node v i ; P(v j ) is the formation pressure of the previous adjacent node v j ; α 1 is the weight coefficient for adjusting the influence of distance on the path cost; α 2 is the weight coefficient for adjusting the influence of rock formation hardness on the path cost; α 3 is the weight coefficient for adjusting the influence of formation pressure on the path cost;

[0040] Define a heuristic function to estimate the remaining path cost from the current node to the target node; the heuristic function is h(v i ) = ‖v i - v o ‖ - δ·S′; where ‖v i - v o ‖ is the Euclidean distance from the current node v i to the target node v o ; S′ is the oil and gas reserves of the current node v i ; δ is the influence coefficient for adjusting the influence of oil and gas reserves on the heuristic function;

[0041] Combine the path cost function and the heuristic function to obtain a comprehensive evaluation function; the comprehensive evaluation function is f(v i ) = g(v i ) + h(vi );Adopt Algorithm A * to perform path search. Starting from the initial node v s , gradually select the node that minimizes the comprehensive evaluation function as the current node until the target node v o is selected and then stop; collect all the selected nodes, and then obtain the preliminary oil drilling path.

[0042] Furthermore, the method for obtaining the final oil drilling path includes:

[0043] Perform actual drilling according to the preliminary oil drilling path, and collect real-time drilling data during the actual drilling process through measurement-while-drilling technology; the real-time drilling data includes bit pressure, drilling speed, drilling inclination angle, drilling depth, and adjacent formation pressure; based on the real-time drilling data, use the genetic algorithm to iteratively optimize the preliminary oil drilling path; preset the risk threshold of the oil drilling path, and stop the iteration when the risk value of the adjusted oil drilling path is less than or equal to the preset risk threshold of the oil drilling path, and then obtain the final oil drilling path;

[0044] The risk value of the adjusted oil drilling path is obtained through the path risk calculation formula; the path risk calculation formula is: where r i′ is the risk value of the adjusted oil drilling path; is the weight factor of the risk value of the adjusted oil drilling path; i′ is the index of the oil drilling path type; the risk value r of the adjusted oil drilling path i′ is obtained through the risk assessment formula; the risk assessment formula is where μ is the friction coefficient; p bit is the bit pressure; υ drill is the drilling speed; △p is the adjacent formation pressure difference; △d is the drilling depth difference; n is the total number of risk values of the adjusted oil drilling path.

[0045] Furthermore, the method for real-time transmitting the final oil drilling path to the intelligent exploration terminal through satellite communication technology includes:

[0046] Install satellite communication equipment on the drilling platform or drilling equipment. The satellite communication equipment includes a mobile satellite antenna and a satellite modem; use the satellite communication equipment to establish a real-time communication link with the satellite through the mobile satellite antenna. The satellite communication equipment establishes a connection according to the MQTT protocol and uploads the final oil drilling path to the intelligent exploration terminal through the real-time communication link.

[0047] The technical effects and advantages of an intelligent exploration system for oil exploration according to the present invention:

[0048] The present invention simulates the propagation of seismic waves through the elastic wave equation, which can accurately reflect the behavior of seismic waves in complex subsurface media, including the effects of factors such as wave velocity, density, and medium heterogeneity. This method can provide high simulation accuracy; by using optimization algorithms such as gradient descent, it is possible to gradually approximate the simulated waveform to the actual observed waveform by continuously adjusting the parameters in the three-dimensional geological model; by controlling the influence of the source point and the receiving point, the propagation operator constraint formula enables the model to adapt to different seismic events and different scales of seismic sources; for different seismic sources and propagation paths, the constrained propagation operator can help the model better adapt to various data and conditions; through the control of the L2 norm loss function, the geological model can be automatically optimized in repeated iterations, so that the final model can have better prediction ability in practical applications, especially suitable for seismic wave simulation in real-time monitoring and dynamic environments; dynamically adjusting the weight factor of the receiving point can optimize the contribution of each point to the final loss function according to the time shift of different receiving points; receiving points with larger time shifts of the received waveform will receive more attention, while points with smaller time shifts will receive relatively lower weights, which can more accurately capture the details of wave propagation;

[0049] By discretizing the oil and gas reservoir distribution area into voxels and providing multi-source information for each voxel, the complex geological structure underground can be better reflected; the path cost function not only considers the distance between adjacent nodes, but also combines important factors such as rock formation hardness and formation pressure; through the comprehensive path cost formula, the actual challenges that may be encountered during the drilling process can be more accurately quantified, avoiding selecting paths with unfavorable or unstable geological conditions, thereby optimizing the drilling effect; by combining the path cost function with a heuristic function to form a comprehensive evaluation function, it can be ensured that the process of path search not only considers the actual cost of the current path, but also can reasonably predict the remaining path cost to the target;

[0050] By using measurement-while-drilling technology to collect drilling data in real time (such as bit pressure, drilling speed, drilling inclination angle, etc.), various dynamic changes encountered during the drilling process can be reflected in real time; this optimization based on real-time data can enable the path planning to be adjusted in a timely manner to ensure a smoother and more efficient drilling process; the path risk calculation and risk assessment formula can comprehensively consider key factors such as the friction coefficient, bit pressure, and drilling speed, and ensure the safety of the drilling path by presetting a risk threshold. By continuously optimizing the drilling path, the efficiency of the drilling process can be improved. Brief Description of the Drawings

[0051] Figure 1 It is a schematic structural diagram of an intelligent exploration system for oil exploration according to the present invention;

[0052] Figure 2 It is a schematic flow diagram of an intelligent exploration method for oil exploration according to the present invention; Detailed Embodiments

[0053] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0054] Embodiment 1

[0055] Please refer to Figure 1 As shown, an intelligent exploration system for oil exploration in this embodiment includes:

[0056] A data acquisition module for acquiring multi-source information data of the exploration area;

[0057] A data processing module for preprocessing the multi-source information data to obtain a multi-source feature data set; using the full waveform inversion algorithm to convert the multi-source feature data set into a three-dimensional geological model, and visualizing the three-dimensional geological model to obtain a three-dimensional geological map;

[0058] A regional prediction module for training and obtaining an oil and gas reservoir area prediction model based on the three-dimensional geological map, and predicting the distribution area of the oil and gas reservoir based on the oil and gas reservoir area prediction model;

[0059] A path planning module for formulating a preliminary oil drilling path according to the predicted distribution area of the oil and gas reservoir; collecting real-time drilling data, and dynamically adjusting the preliminary oil drilling path according to the real-time drilling data to obtain a final oil drilling path;

[0060] A communication transmission module for real-time transmitting the final oil drilling path to the intelligent exploration terminal through satellite communication technology; each module is connected by wired and / or wireless means.

[0061] The multi-source information data includes geophysical exploration data, seismic wave data and geological data; the geophysical exploration data includes gravity, magnetism, resistivity, electromagnetic method and spontaneous potential; the seismic wave data includes seismic wave reflection data, seismic wave propagation velocity, seismic wave amplitude and seismic wave frequency; the geological data includes core samples, surface sediment samples and fluid samples;

[0062] By analyzing magnetic forces, it helps to detect the distribution and structural characteristics of underground magnetic minerals. In complex tectonic regions, it can effectively assist in identifying the geological background of oil and gas reservoirs; resistivity helps to judge the porosity and permeability of oil and gas reservoirs and is commonly used to identify the reserves of oil and gas reservoirs; underground rock formations with low resistivity usually indicate large porosity or high water content, while rock formations with high resistivity indicate dry rock formations or containing oil and gas; electromagnetic methods can penetrate deep underground rock formations and are used to reveal underground oil and gas reservoir layers and aquifers, suitable for detecting deeply buried oil and gas resources; spontaneous potential refers to the potential difference generated by underground rock formations under natural conditions, usually caused by natural phenomena such as groundwater flow and electrochemical reactions of minerals, and is used to detect the distribution of groundwater and oil and gas reservoirs.

[0063] Seismic waves will reflect at the interfaces of formations with different densities and elastic moduli. Detecting the time difference (reflection time difference) of the reflected seismic waves can construct an image of the underground rock formations; seismic wave reflection data can reveal the depth, shape, and structure of underground rock formations, helping to identify geological structures such as oil and gas reservoirs and faults; the propagation speeds of seismic waves in different geological layers are different. Seismic waves propagate faster in harder and denser underground rock formations and slower in soft or aquifer layers; the amplitude of seismic waves reflects the characteristics of the underground rock formation interfaces. A stronger seismic wave amplitude indicates that the underground rock interface is harder and denser, while a weaker seismic wave amplitude indicates that the underground rock formation interface is soft and porous; the frequency of seismic waves helps to distinguish different types of underground rock formations. High-frequency seismic waves can detect shallow underground rock formations, while low-frequency seismic waves are suitable for penetrating deeper underground rock formations.

[0064] Core samples are cylindrical rocks extracted from underground wells and are used to provide detailed rock composition data; surface sediment samples help to understand the sedimentary environment and geological history, and thus infer the formation background of underground oil and gas resources; fluid samples refer to natural gas, crude oil, groundwater, etc. extracted from oil and gas reservoirs. By analyzing fluid samples, the properties, reserves, and fluidity of oil and gas reservoirs can be understood, providing an important basis for formulating development plans.

[0065] Methods for preprocessing multi-source information data to obtain a multi-source feature data set include:

[0066] Identifying and removing outliers in geophysical exploration data, seismic wave data, and geological data included in multi-source information data through density clustering algorithms to obtain geophysical exploration feature data sets, seismic wave feature data sets, and geological feature data sets;

[0067] Feature extraction is performed on the geophysical exploration feature dataset, seismic wave feature dataset, and geological feature dataset through principal component analysis, and standard deviation normalization processing is performed on the geophysical exploration feature dataset, seismic wave feature dataset, and geological feature dataset after feature extraction, converting them into a standard normal distribution with a mean of 0 and a standard deviation of 1, obtaining the normalized geophysical exploration feature dataset, seismic wave feature dataset, and geological feature dataset; the normalized geophysical exploration feature dataset, seismic wave feature dataset, and geological feature dataset are fused through a weighted model to obtain a multi-source feature dataset.

[0068] The method for converting the multi-source feature dataset into a three-dimensional geological model using the full waveform inversion algorithm includes: S41. Preset the three-dimensional geological model as m. The three-dimensional geological model m includes two physical properties of underground media, and the two physical properties of underground media are seismic wave velocity and density. Preset the distribution of seismic wave velocity and density in the three-dimensional geological model as: m(x, y, z) = (v′(x, y, z), ρ(x, y, z)); where, v′(x, y, z) is the seismic wave velocity at a certain position (x, y, z) in the three-dimensional geological model; ρ(x, y, z) is the density at a certain position (x, y, z) in the three-dimensional geological model; (x, y, z) is the position coordinate where seismic wave propagation occurs in the three-dimensional geological model.

[0069] S42. According to the current three-dimensional geological model, simulate the propagation process of seismic waves through the elastic wave equation; the elastic wave equation is: Among them, is the second derivative of the seismic wave displacement u(x, t) with respect to time t, that is, the acceleration of the seismic wave; u(x, t) is the seismic wave displacement at a certain position (x, y, z) and time t in the three-dimensional geological model, representing the change of the wave field.

[0070] c 2 (x, y, z) is the wave velocity function, representing the propagation velocity of seismic waves at each point in the three-dimensional geological model. The wave velocity is a part of the physical properties of the medium and usually changes with the change of underground media. For example, the wave velocities of different rock layers, minerals, or liquids are different, so c 2 (x, y, z) may be a function of a certain position (x, y, z) in the three-dimensional geological model. is the Laplace operator, representing wave propagation in the three-dimensional geological model; t is the time variable, representing the evolution of the change of the wave with time during propagation. The right side and the left side of the wave equation both involve time evolution, reflecting the propagation and change of the wave with time.

[0071] S43. Solve the elastic wave equation through the finite difference method to obtain the simulated waveform of the seismic wave. The simulated waveform of the seismic wave is: b pe = F[m, (x ce , yce , z ce ), (x a , y a , z a )]; where, b pe is the simulated waveform data of the seismic wave; F is the propagation operator of the seismic wave; (x ce , y ce , z ce ) is the source position coordinate of the seismic wave; (x a , y a , z a ) is the receiving point position coordinate of the seismic wave;

[0072] S44. Restrict the propagation operator of the seismic wave through the propagation operator restriction formula. The propagation operator restriction formula is: where, F′ is the propagation operator of the restricted seismic wave; N su is the number of source points of the seismic wave; N re is the number of receiving points of the seismic wave; ∈ is a constant that controls the influence of the source point and the receiving point on the propagation operator;

[0073] For example, if the propagation operator F of the seismic wave is 0.8, the number of source points N su of the seismic wave is 10, the number of receiving points N re of the seismic wave is 20, and the constant ∈ that controls the influence of the source point and the receiving point on the propagation operator is 0.1, then the propagation operator of the restricted seismic wave

[0074] S45. Construct an L2 - norm loss function to quantify the difference between the simulated waveform data and the actual observed waveform data of the seismic wave; The L2 - norm loss function is: where, J(m) is the L2 - norm loss function; b os ((x a , y a , z a ), t a ) is the actual observed waveform data of the seismic wave at the a - th receiving point; b pe ((x a , y a , z a ), t a , m) is the simulated waveform data of the seismic wave calculated through the three - dimensional geological model m; t a is the time shift at the a - th receiving point; ω a is the weight factor for each receiving point; M is the total number of receiving points of the seismic wave; a is the index of the receiving point, a = 1, 2,..., M; Dynamically adjust the weight factor ω a for each receiving point through the weight factor adaptive formula;

[0075] S46. Update the parameters of the three-dimensional geological model through the gradient descent update formula, and the gradient descent update formula is: where, m k+1 is the three-dimensional geological model for the (k + 1)-th iteration; m k is the three-dimensional geological model for the k-th iteration; γ is the learning rate, which controls the step size of each update; is the gradient of the L2-norm loss function J(m) with respect to the three-dimensional geological model m k .

[0076] S47. Preset the loss threshold of the L2-norm loss function, and continuously update the three-dimensional geological model through repeated iterations until the L2-norm loss function is less than or equal to the preset loss threshold of the L2-norm loss function, and then stop to obtain the final three-dimensional geological model m * .

[0077] The method for dynamically adjusting the weight factor ω a of each receiving point through the weight factor adaptive formula includes:

[0078] The weight factor adaptive formula is: where, T max is the maximum time for seismic wave propagation, representing the maximum time required for seismic waves to propagate in space; T M is the total time required to receive seismic waves, representing the transmission time from the seismic source to all receiving points.

[0079] In the propagation of seismic waves, the attenuation of seismic waves occurs as the propagation time increases. The longer the propagation time, the weaker the signal received by the receiving point, because during the propagation of seismic waves in the medium, absorption, reflection and other effects are usually encountered. Therefore, the weight of the receiving point with a longer receiving time should be smaller to avoid the influence of too distant receiving points on the model being too large;

[0080] For example, assume that the maximum time for seismic wave propagation is 50 seconds, the total number M of receiving points of seismic waves is 10, the total time T M required to receive seismic waves is 100 seconds, the transmission time from the seismic source to the first receiving point is 10 seconds, the transmission time from the seismic source to the second receiving point is 20 seconds, the transmission time from the seismic source to the third receiving point is 15 seconds, the transmission time from the seismic source to the fourth receiving point is 30 seconds, and the transmission time from the seismic source to the fifth receiving point is 25 seconds; calculate the weight factor of each receiving point:

[0081] The weight factor of the first receiving point: The weight factor of the second receiving point: The weight factor of the third receiving point: Weight factor of the 4th receiving point: Weight factor of the 5th receiving point:

[0082] The method for obtaining a 3D geological map includes:

[0083] Use different filters in ParaView software to cut the 3D geological model according to any geometric shape, view different parts of the 3D geological model, create a section, and display the slices passing through the model to view the geological structures at different depths underground; extract the surface of the 3D geological model from the 3D grid data to view the external morphology of the underground structure; set different colors for different regions according to seismic wave velocity and density to distinguish different underground rock layers; adjust the transparency of different parts to view the internal structure of the 3D geological model, and present the underground hierarchical structure by adjusting transparency and color, and finally obtain a 3D geological map.

[0084] The training method of the oil and gas reservoir area prediction model includes:

[0085] Divide the dataset into a training set, a validation set, and a test set; construct an oil and gas reservoir area prediction model, which includes an input layer, a 3D convolutional layer, a 3D pooling layer, a fully connected layer, and an output layer; the input data is the historical 3D geological map, and the output data is the oil and gas reservoir distribution area; use ReLU as the activation function; the oil and gas reservoir area prediction model is a 3D convolutional neural network model;

[0086] Use the mean squared error as the loss function of the model to measure the difference between the predicted value and the actual value of the model; use the training set to train the model, and minimize the loss function through the Adam optimizer; use the validation set to evaluate the performance of the oil and gas reservoir area prediction model, and calculate the accuracy index to measure the performance of the model; tune the hyperparameters of the model, calculate the gradient of each hyperparameter through the backpropagation algorithm, and optimize the model hyperparameters through the gradient descent method to improve the model performance;

[0087] Evaluate the performance of the model in the prediction task through the test set, stop the test when the performance of the model in the prediction task reaches the preset performance threshold to obtain the oil and gas reservoir area prediction model; use the trained oil and gas reservoir area prediction model to predict the current 3D geological map to obtain the oil and gas reservoir distribution area.

[0088] The method for obtaining the preliminary oil drilling path includes:

[0089] Discretize the oil and gas reservoir distribution area into Q voxels; a voxel is a small volume unit in the oil and gas reservoir distribution area, containing multi-source information data corresponding to the oil and gas reservoir distribution area; use the spatial position of each voxel as a node and the feasible path between adjacent voxels as an edge to construct a graph structure; the graph structure is G =

[0090] {V, E}; where V is a set of nodes representing each possible drilling location in the oil and gas reservoir distribution area; E is a set of edges representing the connection relationships between adjacent nodes;

[0091] The preset initial node is v s , which is a certain point on the ground or a drilling platform; the target node is v o , which is the center point or high-probability area of the oil and gas reservoir distribution area;

[0092] Define a path cost function to represent the cumulative cost from the initial node to the current node; the path cost function is where v i is the current node; v j is the previous adjacent node of the current node v i ; g(v j ) is the cumulative cost from the initial node to node v j ; w ij is the path cost between adjacent nodes, that is, the path cost between the current node v i and the previous adjacent node v j ; i and j are the indexes of the node numbers; N′ is the set of adjacent nodes of the current node v i ;

[0093] The path cost w ij between adjacent nodes is obtained through the comprehensive path cost formula; the comprehensive path cost formula is w ij =α 1 ·d(v i , v j ) + α 2 ·|H(v i ) - H(v j )| + α 3 ·|P(v i ) - P(v j )|; where d(v i , v j ) is the distance between the current node v i and the previous adjacent node v j ; H(v i ) is the rock formation hardness of the current node v i ; H(v j ) is the rock formation hardness of the previous adjacent node v j ; P(v i ) is the formation pressure of the current node v i ; P(v j ) is the formation pressure of the previous adjacent node v j ; α 1is the weight coefficient for adjusting the influence of distance on the path cost; α 2 is the weight coefficient for adjusting the influence of rock formation hardness on the path cost; α 3 is the weight coefficient for adjusting the influence of formation pressure on the path cost;

[0094] Define a heuristic function, which is used to estimate the remaining path cost from the current node to the target node; the heuristic function is h(v i ) = ‖v i - v o ‖ - δ·S′; where, ‖v i - v o ‖ is the Euclidean distance from the current node v i to the target node v o ; S′ is the oil and gas reserve of the current node v i ; δ is the influence coefficient for adjusting the influence of oil and gas reserves on the heuristic function;

[0095] Combine the path cost function and the heuristic function to obtain a comprehensive evaluation function; the comprehensive evaluation function is f(v i ) = g(v i ) + h(v i ); Use the A * algorithm to perform path search. Starting from the initial node v s , gradually select the node that minimizes the comprehensive evaluation function as the current node until the target node v o is selected and then stop; collect all the selected nodes, and then obtain the preliminary oil drilling path.

[0096] The method for obtaining the final oil drilling path includes:

[0097] Perform actual drilling according to the preliminary oil drilling path, and collect real-time drilling data during the actual drilling process through measurement-while-drilling technology; the real-time drilling data includes bit pressure, drilling speed, drilling inclination angle, drilling depth, and adjacent formation pressure; based on the real-time drilling data, use the genetic algorithm to iteratively optimize the preliminary oil drilling path; preset the risk threshold of the oil drilling path. When the risk value of the adjusted oil drilling path is less than or equal to the preset risk threshold of the oil drilling path, stop the iteration, and then obtain the final oil drilling path;

[0098] The risk value of the adjusted oil drilling path is obtained through the path risk calculation formula; the path risk calculation formula is: where, r i′ is the risk value of the adjusted oil drilling path; is the weight factor of the risk value of the adjusted oil drilling path; i′ is the index of the oil drilling path type; the risk value r i′Obtained through a risk assessment formula; the risk assessment formula is where μ is the friction coefficient; p bit is the bit pressure; υ drill is the drilling speed; △p is the adjacent formation pressure difference; △d is the drilling depth difference; n is the total number of adjusted petroleum drilling path risk values.

[0099] The method for real-time transmitting the final petroleum drilling path to the intelligent exploration terminal through satellite communication technology includes:

[0100] Install satellite communication equipment on the drilling platform or drilling equipment. The satellite communication equipment includes a mobile satellite antenna and a satellite modem; use the satellite communication equipment to establish a real-time communication link with the satellite through the mobile satellite antenna. The satellite communication equipment establishes a connection according to the MQTT protocol and uploads the final petroleum drilling path to the intelligent exploration terminal through the real-time communication link.

[0101] The preset petroleum drilling path risk threshold is set by the staff. Different petroleum drilling path risk values are collected through the intelligent exploration terminal, and the average value of multiple petroleum drilling path risk values is taken as the preset petroleum drilling path risk threshold.

[0102] In this embodiment, by simulating the propagation of seismic waves through the elastic wave equation, it can accurately reflect the behavior of seismic waves in complex underground media, including the influence of factors such as wave velocity, density, and medium heterogeneity. This method can provide a high simulation accuracy; by using optimization algorithms such as gradient descent, the parameters in the three-dimensional geological model can be continuously adjusted to make the simulated waveform gradually approach the actual observed waveform; by controlling the influence of the source point and the receiving point, the propagation operator constraint formula enables the model to adapt to different seismic events and different scales of sources; for different seismic sources and propagation paths, the constrained propagation operator can help the model better adapt to various data and conditions; through the control of the L2 norm loss function, the geological model can be automatically optimized in repeated iterations, so that the final model can have better prediction ability in practical applications, especially suitable for seismic wave simulation in real-time monitoring and dynamic environments; dynamically adjusting the weight factor of the receiving point can optimize the contribution of each point to the final loss function according to the time shift of different receiving points; receiving points with larger time shifts of the received waveform will receive more attention, while points with smaller time shifts will receive relatively lower weights, and the details of wave propagation can be captured more accurately;

[0103] By discretizing the oil and gas reservoir distribution area into voxels and providing multi-source information for each voxel, the complex underground geological structure can be better reflected; the path cost function not only considers the distance between adjacent nodes, but also combines important factors such as rock formation hardness and formation pressure; through the comprehensive path cost formula, the actual challenges that may be encountered during the drilling process can be more accurately quantified, avoiding selecting paths with unfavorable or unstable geological conditions, thereby optimizing the drilling effect; by combining the path cost function with the heuristic function to form a comprehensive evaluation function, it can ensure that the path search process not only considers the actual cost of the current path, but also can reasonably predict the remaining path cost to the target;

[0104] By using measurement-while-drilling technology to collect drilling data in real time (such as bit pressure, drilling speed, drilling inclination angle, etc.), various dynamic changes encountered during the drilling process can be reflected in real time; this optimization based on real-time data can enable the path planning to be adjusted in a timely manner, ensuring a smoother and more efficient drilling process; the path risk calculation and risk assessment formula can comprehensively consider key factors such as the friction coefficient, bit pressure, and drilling speed, and ensure the safety of the drilling path by presetting risk thresholds. By continuously optimizing the drilling path, the efficiency of the drilling process can be improved.

[0105] Embodiment 2

[0106] Please refer to Figure 2 As shown, for the parts not described in detail in this embodiment, refer to the description content of Embodiment 1. Provide an intelligent exploration method for oil exploration, including:

[0107] S1. Collect multi-source information data of the exploration area;

[0108] S2. Preprocess the multi-source information data to obtain a multi-source feature data set; use the full waveform inversion algorithm to convert the multi-source feature data set into a three-dimensional geological model, and visualize the three-dimensional geological model to obtain a three-dimensional geological map;

[0109] S3. Train and obtain an oil and gas reservoir area prediction model based on the three-dimensional geological map, and predict the oil and gas reservoir distribution area based on the oil and gas reservoir area prediction model;

[0110] S4. According to the predicted oil and gas reservoir distribution area, formulate a preliminary oil drilling path; collect real-time drilling data, and dynamically adjust the preliminary oil drilling path according to the real-time drilling data to obtain the final oil drilling path;

[0111] S5. Transmit the final oil drilling path to the intelligent exploration terminal in real time through satellite communication technology;

[0112] Since the electronic device introduced in this embodiment is the electronic device adopted in the intelligent exploration system for oil exploration in the embodiments of the present application, based on the intelligent exploration system for oil exploration introduced in the embodiments of the present application, those skilled in the art can understand the specific implementation manners and various variations of the electronic device in this embodiment. Therefore, the specific implementation of how this electronic device realizes the method in the embodiments of the present application will not be described in detail herein. As long as those skilled in the art implement the electronic device adopted in the intelligent exploration system for oil exploration in the embodiments of the present application, it falls within the protection scope of the present application.

[0113] The above formulas are all dimensionless and take their numerical values for calculation. The formula is a formula obtained by collecting a large amount of data for software simulation to approximate the real situation. The preset parameters and threshold selection in the formula are set by those skilled in the art according to the actual situation.

[0114] The above description is only the preferred implementation manner of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions within the concept of the present invention belong to the protection scope of the present invention. It should be noted that for those ordinary users in the technical field, several improvements and refinements made without departing from the principle of the present invention should also be regarded as within the protection scope of the present invention.

Claims

1. An intelligent exploration system for oil exploration, characterized in that: include: Data acquisition module, used to collect multi-source information data of the exploration area; A data processing module is used to pre-process multi-source information data to obtain a multi-source feature data set; The multi-source feature data set is converted into a three-dimensional geological model using a full waveform inversion algorithm, and the three-dimensional geological model is visualized to obtain a three-dimensional geological map; The regional prediction module is used to obtain the regional prediction model of oil and gas reservoirs according to the three-dimensional geological map training, and predict the distribution area of ​​oil and gas reservoirs based on the regional prediction model of oil and gas reservoirs; The path planning module is used to develop a preliminary oil drilling path based on the predicted oil and gas reservoir distribution area; Collect real-time drilling data, dynamically adjust the preliminary oil drilling path according to the real-time drilling data, and obtain the final oil drilling path; The communication transmission module is used to transmit the final oil drilling path to the intelligent exploration terminal in real time through satellite communication technology; each module is connected by wired and / or wireless means.

2. The intelligent exploration system for oil exploration according to claim 1, characterized in that: The multi-source information data includes geophysical data, seismic wave data and geological data; the geophysical data includes gravity, magnetism, resistivity, electromagnetic method and natural potential; the seismic wave data includes seismic wave reflection data, seismic wave propagation velocity, seismic wave amplitude and seismic wave frequency; the geological data includes core samples, surface sediment samples and fluid samples.

3. The intelligent exploration system for oil exploration according to claim 2, characterized in that: The method of preprocessing multi-source information data to obtain a multi-source feature data set includes: By using a density clustering algorithm, outliers existing in the multi-source information data including geophysical data, seismic wave data and geological data are identified and eliminated, thereby obtaining a geophysical feature data set, a seismic wave feature data set and a geological feature data set; The geophysical feature data set, seismic wave feature data set and geological feature data set are feature extracted through principal component analysis, and the geophysical feature data set, seismic wave feature data set and geological feature data set after feature extraction are normalized by standard deviation and converted into standard normal distribution with mean 0 and standard deviation 1 to obtain normalized geophysical feature data set, seismic wave feature data set and geological feature data set; the normalized geophysical feature data set, seismic wave feature data set and geological feature data set are fused through a weighted model to obtain a multi-source feature data set.

4. The intelligent exploration system for oil exploration according to claim 3, characterized in that: The method of converting a multi-source feature data set into a three-dimensional geological model using a full waveform inversion algorithm comprises: S41. The three-dimensional geological model is preset as m. The three-dimensional geological model m includes two physical properties of underground media, which are seismic wave velocity and density. The distribution of the seismic wave velocity and density in the three-dimensional geological model is preset as follows: m(x,y,z)=(v′(x,y,z),ρ(x,y,z)); wherein v′(x,y,z) is the seismic wave velocity at a certain position (x,y,z) in the three-dimensional geological model; ρ(x,y,z) is the density at a certain position (x,y,z) in the three-dimensional geological model; (x,y,z) is the position coordinates where seismic wave propagation occurs in the three-dimensional geological model; S42. According to the current three-dimensional geological model, the propagation process of seismic waves is simulated by the elastic wave equation; the elastic wave equation is: in, is the second-order derivative of the seismic wave displacement u(x, t) relative to time t, that is, the acceleration of the seismic wave; u(x, t) is the seismic wave displacement at a certain position (x, y, z) and time t in the three-dimensional geological model; c 2 (x, y, z) is the wave velocity function, which represents the propagation velocity of seismic waves at each point in the three-dimensional geological model; is the Laplace operator, which represents the wave propagation in the three-dimensional geological model; t is the time variable, which represents the evolution of the wave changes during the propagation process over time; S43. The elastic wave equation is solved by the finite difference method to obtain the simulated waveform of the seismic wave. The simulated waveform of the seismic wave is: b pe =F[m,(x ce ,y ce ,z ce ),(x a ,y a ,z a )]; where b pe is the simulated waveform data of seismic waves; F is the propagation operator of seismic waves; (x ce ,y ce ,z ce ) is the coordinate of the earthquake source position of the seismic wave; (x a ,y a ,z a ) are the coordinates of the receiving point of the seismic wave; S44. The propagation operator of the seismic wave is restricted by a propagation operator restriction formula, and the propagation operator restriction formula is: Where F′ is the propagation operator of seismic waves after restriction; N su is the number of earthquake source points; N re is the number of receiving points for seismic waves; ∈ is a constant that controls the influence of the source point and receiving point on the propagation operator; S45, constructing an L2 norm loss function to quantify the difference between the simulated waveform data of the seismic wave and the actual observed waveform data; the L2 norm loss function is: Among them, J(m) is the L2 norm loss function; b os ((x a ,y a ,z a ),t a ) is the actual observed waveform data of the seismic wave at the ath receiving point; b pe ((x a ,y a ,z a ),t a , m) is the simulated waveform data of seismic waves calculated by the three-dimensional geological model m; t a is the time offset of the ath receiving point; ω a is the weight factor of each receiving point; M is the total number of receiving points for seismic waves; a is the index of the receiving point, a=1,2,...,M; the weight factor ω of each receiving point is adjusted by the weight factor adaptive formula a Conduct dynamic adjustment design; S46, updating the parameters of the three-dimensional geological model by using a gradient descent update formula, the gradient descent update formula is: Among them, m k+1 is the three-dimensional geological model of the k+1th iteration; m k is the three-dimensional geological model of the kth iteration; γ is the learning rate; is the L2 norm loss function J(m) with respect to the 3D geological model m k The gradient of S47, preset a loss threshold of the L2 norm loss function, and continuously update the three-dimensional geological model through repeated iterations until the L2 norm loss function is less than or equal to the loss threshold of the preset L2 norm loss function, and then obtain the final three-dimensional geological model m * .

5. The intelligent exploration system for oil exploration according to claim 4, characterized in that: The weight factor ω of each receiving point is adjusted by the weight factor adaptive formula a Methods for dynamically adjusting designs include: The adaptive formula of the weight factor is: Among them, T max is the maximum time for seismic wave propagation; T M The total time required to receive seismic waves.

6. The intelligent exploration system for oil exploration according to claim 5, characterized in that: The method for obtaining the three-dimensional geological map includes: Use different filters in the ParaView software to cut the 3D geological model into any geometric shape, view different parts of the 3D geological model, create a section and display its slices through the model, and view the geological structure at different depths underground; extract the surface of the 3D geological model from the 3D grid data to view the external morphology of the underground structure; set different colors for different areas according to the seismic wave velocity and density to distinguish different underground rock layers; adjust the transparency of different parts to view the internal structure of the 3D geological model, and present the underground hierarchical structure by adjusting the transparency and color, and finally obtain a 3D geological map.

7. The intelligent exploration system for oil exploration according to claim 6, characterized in that: The training method of the oil and gas reservoir region prediction model includes: The data set is divided into a training set, a validation set and a test set; an oil and gas reservoir area prediction model is constructed, and the oil and gas reservoir area prediction model includes an input layer, a 3D convolution layer, a 3D pooling layer, a fully connected layer and an output layer; the input data is a historical three-dimensional geological map, and the output data is the oil and gas reservoir distribution area; ReLU is used as an activation function; the oil and gas reservoir area prediction model is a 3D convolutional neural network model; Use mean square error as the loss function of the model to measure the difference between the model's predicted value and the actual value; use the training set to train the model and use the Adam optimizer to minimize the loss function; use the validation set to evaluate the performance of the reservoir area prediction model and calculate the accuracy index to measure the performance of the model; tune the model's hyperparameters, calculate the gradient of each hyperparameter through the back propagation algorithm, and optimize the model hyperparameters through the gradient descent method to improve the model performance; The performance of the model in the prediction task is evaluated through the test set. When the performance of the model in the prediction task reaches the preset performance threshold, the test is stopped to obtain the oil and gas reservoir area prediction model; the trained oil and gas reservoir area prediction model is used to predict the current three-dimensional geological map to obtain the oil and gas reservoir distribution area.

8. The intelligent exploration system for oil exploration according to claim 7, characterized in that: The method for obtaining the preliminary oil drilling path includes: The oil and gas reservoir distribution area is discretized into Q voxels; a voxel is a small volume unit in the oil and gas reservoir distribution area, containing multi-source information data corresponding to the oil and gas reservoir distribution area; the spatial position of each voxel is used as a node, and the feasible path between adjacent voxels is used as an edge to construct a graph structure; the graph structure is G = {V,E}; where V is the node set; E is the edge set; The default initial node is v s , the target node is v o ; Define the path cost function to represent the cumulative cost from the initial node to the current node; the path cost function is Among them, v i is the current node; v j is the current node v i The previous adjacent node of j ) is from the initial node to node v j The cumulative cost of ij is the path cost between adjacent nodes, that is, the current node v i and the previous adjacent node v j The path cost between them; i and j are the indexes of the node numbers; N′ is the current node v i The set of adjacent nodes of ; The path cost w between the adjacent nodes ij It is obtained through the comprehensive path cost formula; the comprehensive path cost formula is w ij =α1·d(v i ,v j )+α2·|H(v i )-H(v j )|+α3·|P(v i )-P(v j )|; where d(v i ,v j ) is the current node v i and the previous adjacent node v j The distance between i ) is the current node v i The hardness of the rock formation; H(v j ) is the previous adjacent node v j The rock hardness; P(v i ) is the current node v i The formation pressure; P(v j ) is the previous adjacent node v j α1 is the weight coefficient for adjusting the influence of distance on path cost; α2 is the weight coefficient for adjusting the influence of rock formation hardness on path cost; α3 is the weight coefficient for adjusting the influence of formation pressure on path cost; Define a heuristic function to estimate the remaining path cost from the current node to the target node; the heuristic function is h(v i )=‖v i -v o ‖-δ·S′; where ‖v i -v o ‖ is the current node v i To the target node v o The Euclidean distance of the current node v i The oil and gas reserves; δ is the influence coefficient of adjusting the oil and gas reserves on the heuristic function; The path cost function and the heuristic function are combined to obtain a comprehensive evaluation function; the comprehensive evaluation function is f(v i )=g(v i )+h(v i ); using A * The algorithm performs path search, starting from the initial node v s At the beginning, the node that minimizes the comprehensive evaluation function is gradually selected as the current node until the target node v o Stop when selected; collect all selected nodes to obtain a preliminary oil drilling path.

9. The intelligent exploration system for oil exploration according to claim 8, characterized in that: The method for obtaining the final oil drilling path includes: Actual drilling is carried out according to the preliminary oil drilling path, and real-time drilling data is collected during the actual drilling process through measurement while drilling technology; the real-time drilling data includes drill bit pressure, drilling speed, drilling inclination, drilling depth and adjacent formation pressure; based on the real-time drilling data, the preliminary oil drilling path is iteratively optimized using a genetic algorithm; a risk threshold of the oil drilling path is preset, and the iteration is stopped when the adjusted oil drilling path risk value is less than or equal to the preset oil drilling path risk threshold, thereby obtaining the final oil drilling path; The adjusted oil drilling path risk value is obtained through the path risk calculation formula; the path risk calculation formula is: Among them, r i′ is the adjusted oil drilling path risk value; is the weight factor of the adjusted oil drilling path risk value; i′ is the index of the oil drilling path type; the adjusted oil drilling path risk value r i′ It is obtained through the risk assessment formula; the risk assessment formula is Where μ is the friction coefficient; p bit is the drill bit pressure; drill is the drilling speed; △p is the pressure difference between adjacent formations; △d is the drilling depth difference; and n is the total number of adjusted oil drilling path risk values.

10. The intelligent exploration system for oil exploration according to claim 9, characterized in that: The method of transmitting the final oil drilling path to the intelligent exploration terminal in real time through satellite communication technology includes: Satellite communication equipment is installed on the drilling platform or drilling equipment. The satellite communication equipment includes a mobile satellite antenna and a satellite modem. The satellite communication equipment is used to establish a real-time communication link with the satellite through the mobile satellite antenna. The satellite communication equipment establishes a connection according to the MQTT protocol and uploads the final oil drilling path to the intelligent exploration terminal through the real-time communication link.

Images