A method of oil exploration using an intelligent oil exploration robot system

Abstract

The present application belongs to the technical field of oil exploration, and discloses an oil exploration method using an intelligent oil exploration robot system; the method comprises the following steps: collecting geological information data; analyzing the geological information data and constructing a three-dimensional geological model; planning an exploration path based on the three-dimensional geological model; controlling the exploration robot to perform drilling operations and collect drilling data according to the exploration path; analyzing the drilling data and evaluating oil resources; the present application can realize the automation and intelligentization of the oil exploration process, significantly improve the efficiency and accuracy of oil exploration, reduce the operation risk and cost, and realize the systematic evaluation of the overall oil resources, thereby providing a reliable basis for the subsequent development and utilization of oil.

Description

Technical Field

[0001] This invention relates to the field of petroleum exploration technology, and more specifically, to a petroleum exploration method employing an intelligent petroleum exploration robot system. Background Technology

[0002] With the continuous growth of global energy demand, oil exploration technology is facing unprecedented challenges. Traditional oil exploration methods rely on manual operation and fixed equipment. Although they have made significant progress in the early stages, as oil resources gradually enter complex geological environments and hard-to-reach areas, traditional methods have revealed some shortcomings. For example, drilling is high-risk, inefficient, and costly, and manual operation may lead to data errors and operational instability. In addition, the geological conditions of oil and gas reservoirs are becoming increasingly complex, usually requiring accurate exploration data and real-time decision support to ensure the safety and economy of extraction.

[0003] Against this backdrop, the introduction of intelligent technology has become a significant breakthrough in the field of oil exploration. With the development of robotics, artificial intelligence, the Internet of Things, big data analytics, and automation, intelligent oil exploration methods are receiving increasing attention. These methods can effectively optimize decision-making during the exploration process and significantly improve exploration efficiency. For example, patent application CN115559712A discloses an intelligent oil exploration system and its application method, comprising: 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 and classify the downhole data into important data and secondary data, storing the secondary data. The information transmission module is used to transmit important data to the central processing module. The central processing module is used to analyze important data and correct drilling parameters in real time. The interaction module is used to provide real-time feedback on downhole conditions to staff and make manual corrections based on the downhole conditions. This invention ensures the accuracy and effectiveness of downhole signal transmission by incorporating relay stations, while also classifying data based on actual working conditions and analyzing important data in real time to correct drilling parameters.

[0004] However, while the aforementioned technologies can detect downhole data in real time and ensure the accuracy of signal transmission through relay stations, and classify data and correct drilling parameters in real time based on actual working conditions, they mainly focus on the real-time correction of drilling parameters and the guarantee of data transmission. They lack a systematic assessment and optimization plan for the overall oil resources and cannot fully support resource assessment and long-term planning during the exploration process. This leads to insufficient resource assessment, unscientific mining planning, and a mismatch between technology and cost during the exploration process, which in turn affects exploration efficiency and economic benefits.

[0005] In view of this, the present invention proposes an oil exploration method using an intelligent oil exploration robot system to solve the above problems. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art and achieve the above objectives, the present invention provides the following technical solution: a petroleum exploration method employing an intelligent petroleum exploration robot system, comprising:

[0007] S1: Collect geological information data;

[0008] S2: Analyze geological information data and construct a three-dimensional geological model;

[0009] S3: Based on a three-dimensional geological model, plan the exploration route;

[0010] S4: Control the exploration robot to perform drilling operations and collect drilling data according to the exploration path;

[0011] S5: Analyze drilling data to assess oil resources.

[0012] Furthermore, the geological information data includes seismic data, magnetic field data, and gravity field data; the seismic data is the seismic wave reflection time distribution beneath the surface of the area to be explored, and the area to be explored is the area where oil exploration is planned; the magnetic field data is the magnetic field intensity distribution on the surface of the area to be explored; and the gravity field data is the gravity distribution on the surface of the area to be explored.

[0013] The steps for constructing a three-dimensional geological model include:

[0014] S201: Preprocessing geological information data;

[0015] S202: Perform time-depth conversion on the preprocessed seismic data to obtain depth data;

[0016] S203: Construct a three-dimensional geological model based on depth data;

[0017] S204: Perform seismic inversion on the preprocessed seismic data to obtain velocity data;

[0018] S205: Perform magnetic field inversion on the preprocessed magnetic field data to obtain magnetization data;

[0019] S206: Perform gravity inversion on the preprocessed gravity field data to obtain density data;

[0020] S207: Update the three-dimensional geological model based on velocity, magnetization, and density data.

[0021] Furthermore, in step S201, the method for preprocessing the seismic data includes:

[0022] A rectangular coordinate system is constructed for the area to be explored and marked as the regional coordinate system. A grid area is preset, and the area to be explored is divided into grids according to the grid area. The coordinates of each grid point and seismic point are obtained according to the regional coordinate system. The grid point is the center point of each grid, and the seismic point is the location point in the area to be explored corresponding to each reflection time in the seismic data. The Euclidean distance between each grid point and each seismic point is calculated based on the coordinates of each grid point and each seismic point, and is used as the influence distance.

[0023] A preset influence factor is used. Based on the influence factor and influence distance, a weighting coefficient between each grid point and each seismic point is calculated. The expression for the weighting coefficient is: In the formula, qz(p,q) is the weight coefficient between the p-th grid point and the q-th seismic point, yx(p,q) is the influence distance between the p-th grid point and the q-th seismic point, a is the influence factor, p∈[1,P], q∈[1,Q], P is the number of grid points, and Q is the number of seismic points. Based on the weight coefficient between each grid point and each seismic point, the reflection time of each seismic point is weighted and summed to calculate the reflection time corresponding to each grid point, and this summed time is marked as the weighted time. The expression for the weighted time is: In the formula, js p For the weighted time of the p-th grid point, fs q Let q be the reflection time of the q-th earthquake point;

[0024] Methods for preprocessing magnetic field data include:

[0025] The number of magnetic field intensities in the magnetic field data is counted and marked as the sampling number; the background intensity is obtained by summing each magnetic field intensity in the magnetic field data and dividing by the sampling number; the background intensity is subtracted from each magnetic field intensity in the magnetic field data to obtain the magnetic anomaly data corresponding to each magnetic field point. The magnetic field point is the location point in the area to be explored corresponding to each magnetic field intensity in the magnetic field data.

[0026] Methods for preprocessing gravity field data include:

[0027] Obtain the exploration elevation and calculate the topographic influence value; subtract the topographic influence value from each gravity in the gravity field data to obtain the gravity correction value corresponding to each gravity point. The gravity point is the location point in the area to be explored corresponding to each gravity in the gravity field data; the expression for the topographic influence value is: dy = ρgh; where dy is the topographic influence value, ρ is the crustal density, g is the standard gravity, and h is the exploration elevation.

[0028] Further, in step S202, the method for obtaining depth data includes:

[0029] Preliminary velocity data is obtained, which is the estimated wave velocity distribution in the subsurface of the area to be explored. Based on the preliminary velocity data, the estimated wave velocity of each grid point is obtained. The estimated wave velocity of each grid point is multiplied by the corresponding weighted time and then divided by 2 to obtain the depth value of each grid point. The depth value of each grid point is used as the depth data.

[0030] In step S203, the method for constructing a three-dimensional geological model includes:

[0031] Based on the coordinates and depth values ​​of each grid point, construct the three-dimensional coordinates of each grid point; based on the three-dimensional coordinates of each grid point, use a three-dimensional grid reconstruction algorithm to convert it into a three-dimensional geological model;

[0032] In step S207, the method for updating the three-dimensional geological model includes:

[0033] Velocity, magnetization, and density data are used as analytical data. This data is input into a trained geological analysis model to predict corresponding geological labels. Geological labels are numerical tags corresponding to the geological data; different geological data correspond to different labels. Geological data includes stratum type, stratum depth, oil reservoir area, and oil reservoir depth. Specifically, stratum type refers to the type of each stratum in the area to be explored; stratum depth is the vertical distance from the surface of the area to be explored to the top of each stratum; oil reservoir area is the grid containing oil in the area to be explored; oil reservoir depth is the vertical distance from the surface of the area to be explored to the bottom of each oil-bearing stratum. Based on the geological labels, corresponding geological data is obtained, and the 3D geological model is updated. The training process of the geological analysis model includes:

[0034] A set of b sets of analysis data is collected in advance, and corresponding geological labels are assigned to each set of b sets of analysis data. The analysis data and corresponding geological labels are converted into a set of feature vectors. Each set of feature vectors is used as input to the geological analysis model. The geological analysis model outputs a set of predicted geological labels corresponding to each set of analysis data, and uses the actual geological labels corresponding to each set of analysis data as the prediction target. The actual geological labels are the pre-set geological labels corresponding to the analysis data. The training objective is to minimize the sum of prediction errors of all analysis data. The geological analysis model is trained until the sum of prediction errors converges, at which point training stops. The geological analysis model is a deep neural network model.

[0035] Furthermore, the steps for planning the exploration path include:

[0036] Step S301: Treat all grids as nodes and the oil reservoir area as an oil node;

[0037] Step S302: Randomly select an oil node and mark it as the initial node; starting from the initial node, explore all the adjacent nodes of the initial node, and select one node from all the adjacent nodes as the successor node according to the preset filtering rules, and mark the successor node as the filtering node.

[0038] Step S303: Explore all neighboring nodes of the successor node, and select one node from all neighboring nodes as the update node according to the preset filtering rules, update the successor node as the update node, and mark the update node as the filter node.

[0039] Step S304: Repeat step S303 until all adjacent nodes corresponding to the successor node are marked as filter nodes, then the loop ends and the initial path is obtained;

[0040] Step S305: Determine whether the initial path includes all oil nodes; if yes, retain the initial path; otherwise, delete the initial path.

[0041] Step S306: Repeat steps S302 to S305 to obtain m1 initial paths, where m1 is an integer greater than 1, and each of the m1 initial paths is different.

[0042] Step S307: Calculate the cumulative depth gradient of each initial path, and select m2 initial paths. Mark all m2 initial paths as intermediate paths, where 1 < m2 < m1.

[0043] Step S308: Calculate the path distance of each intermediate path, select the best path, and use the best path as the exploration path.

[0044] Furthermore, in step S302, the filtering rule is as follows: if there is an oil node among the adjacent nodes, then one of the oil nodes is randomly selected as the successor node; if there is no oil node among the adjacent nodes, then one of the nodes is randomly selected as the successor node.

[0045] In step S307, the expression for accumulating the depth gradient is: In the formula, ls d Let sd be the cumulative depth gradient of the d-th initial path, d∈[1,m1], sd dk Let sd be the oil reservoir depth of the k-th oil node in the d-th initial path. dk+1 Let K be the oil reservoir depth of the (k+1)th oil node in the d-th initial path, where k∈[1,K-1] and K is the number of oil nodes;

[0046] Methods for selecting m2 initial paths include:

[0047] Sort the cumulative depth gradients of each initial path from smallest to largest, select the first m2 cumulative depth gradients in ascending order, and mark them as the filtering depths; select the initial paths corresponding to the m2 filtering depths from the m1 initial paths.

[0048] In step S308, the path distance of the intermediate path is the number of all nodes traversed by the intermediate path.

[0049] Methods for selecting the best path include:

[0050] Sort the path distances of each intermediate path from smallest to largest, select the path distance that is first in the ascending order, and mark it as the shortest distance; then select the intermediate path corresponding to the shortest distance from m2 intermediate paths.

[0051] Furthermore, the method for controlling the exploration robot to perform drilling operations includes:

[0052] The oil reservoir area where drilling operations are conducted is marked as the current area. Based on the oil reservoir depth in the current area, the maximum drilling depth in the current area is determined. The current drilling depth is acquired in real time, which is the vertical distance between the drill bit and the surface to be explored. The rock strata depths corresponding to the current area are marked as comparison depths. All comparison depths are sorted with the current drilling depth in ascending order. The comparison depth preceding the current drilling depth is obtained in ascending order and marked as the determined depth. The rock strata type corresponding to the determined depth is marked as the determined type. The drilling parameters of the exploration robot are adjusted in real time based on the current drilling depth and the determined type. When the current drilling depth reaches the maximum drilling depth, the exploration robot is controlled to stop drilling operations.

[0053] Furthermore, the step of real-time adjustment of the drilling parameters of the exploration robot includes:

[0054] Step S401: Construct N parameter sets, set different numerical labels for the N parameter sets, and mark them as parameter labels;

[0055] Step S402: Construct a population S containing n individuals, generate the position of each individual, and assign each individual position a one-to-one correspondence with the parameter label. Set the iteration count t to 0.

[0056] Step S403: Determine the attraction function and the iteration threshold T;

[0057] Step S404: Calculate the attractiveness of each individual in population S and update the position of each individual;

[0058] Step S405: Determine whether the iteration number t is less than the iteration threshold T. If yes, let t = t + 1 and return to step S404. If no, proceed to step S406.

[0059] Step S406: Calculate the attraction degree corresponding to each individual in population S, obtain the parameter label corresponding to the individual with the highest attraction degree, obtain the corresponding parameter set based on the obtained parameter label, and adjust the drilling parameters of the exploration robot in real time based on the obtained parameter set.

[0060] Further, in step S401, the method for constructing N parameter sets is as follows: obtain the parameter range, which includes the range of each parameter in the drilling parameters; the drilling parameters include drilling speed, drill bit rotation speed, drill bit pressure and drill bit temperature; randomly select a value from each range in the parameter range and construct a parameter set, and construct a total of N parameter sets, all of which are different;

[0061] In step S402, the expression for each individual location is: In the formula, Let C be the position of the i-th individual. i Let C be the random coefficient of the i-th individual. i ∈[0, 1], i∈[1, n];

[0062] In step S403, the expression for the attraction function is: f = tz; where f is the attraction and tz is the drilling quality. The drilling quality is obtained by setting different numerical labels for different rock strata types and marking them as rock strata labels; obtaining the corresponding parameter set based on the parameter label corresponding to the individual location; using the current drilling depth, the rock strata label corresponding to the determined type, and the parameter set corresponding to the individual location as evaluation data, inputting the evaluation data into the trained quality evaluation model, and evaluating the corresponding drilling quality. Drilling quality is a comprehensive indicator of drilling efficiency, drilling safety, and drilling cost. The training process of the quality evaluation model is consistent with the training process of the geological analysis model, and both are deep neural network models.

[0063] In S404, the method for updating the location of each individual includes:

[0064]

[0065] In the formula, Let v be the position of the i-th individual after the update, v be the maximum attraction, y be the decay factor, e be the natural constant, and l be the position of the i-th individual after the update. ib Let be the distance between the i-th individual and the optimal individual. The position of the optimal individual. Let L be the position of the i-th individual before the update, L be the perturbation factor, R be a random number between [0, 1], and the optimal individual is the individual with the highest attraction in the population S.

[0066] Furthermore, the drilling data includes lithological data, porosity data, and permeability data; the lithological data refers to the actual type of the oil-bearing rock formation; the porosity data is the proportion of pore volume to total volume in the oil-bearing rock formation; the permeability data is the ability of the rock in the oil-bearing rock formation to allow fluid to pass through; the oil resources include oil reserves and recoverability.

[0067] Methods for assessing oil resources include:

[0068] The 3D geological model is divided into grids based on the grid area, and the divided grids are marked as cells. Drilling data is added to the corresponding cells to obtain the simulated geological model. Numerical simulation is used to dynamically simulate the simulated geological model to obtain oil reserves and exploitability.

[0069] The technical effects and advantages of the oil exploration method using an intelligent oil exploration robot system of the present invention are as follows:

[0070] By integrating multi-source geological information data such as seismic, magnetic field, and gravity data, and constructing a three-dimensional geological model and planning the optimal exploration path, the system automates and intelligentizes the oil exploration process. Furthermore, by adjusting drilling parameters in real time, it intelligently controls exploration robots to perform drilling operations, thereby significantly improving the efficiency and accuracy of oil exploration and reducing operational risks and costs. Simultaneously, the system uses dynamic simulation calculations to accurately assess the oil reserves and exploitability of oil-bearing strata based on drilling data collected during drilling operations, achieving a systematic assessment of overall oil resources and providing a reliable basis for subsequent oil development and utilization. Attached Figure Description

[0071] Figure 1 This is a flowchart of an oil exploration method using an intelligent oil exploration robot system according to Embodiment 1 of the present invention;

[0072] Figure 2 This is a flowchart of the exploration path planning method in Embodiment 1 of the present invention. Detailed Implementation

[0073] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0074] Example 1

[0075] Please see Figure 1 As shown in the figure, this embodiment describes an oil exploration method using an intelligent oil exploration robot system. The method includes:

[0076] S1: Collect geological information data.

[0077] Geological information data includes seismic data, magnetic field data, and gravity field data;

[0078] The seismic data represents the time distribution of seismic wave reflections in the area to be explored, which is the area where oil exploration is planned. The seismic data is acquired by generating seismic waves using a seismic source device (such as a mechanical source or a hammer source) built into the exploration robot, and then receiving and recording the reflected seismic waves using a seismic detector built into the robot. The exploration robot is an oil exploration robot. It should be understood that acquiring seismic data helps to infer the structure and characteristics of underground rock strata, thereby inferring the shape, scale, and location of underground oil reservoirs.

[0079] Magnetic field data refers to the distribution of magnetic field intensity on the surface of the area to be explored. Magnetic field data is acquired through magnetometers (such as superconducting quantum interference devices, laser magnetometers, and electronic magnetometers) built into the exploration robot. It should be understood that collecting magnetic field data helps to identify underground magnetic minerals, rock formations, and oil reservoirs, thereby assisting in oil exploration.

[0080] Gravity field data represents the gravity distribution on the surface of the area to be explored. This data is acquired using gravimeters (such as point mass gravimeters and gyro gravimeters) built into the exploration robot. It should be understood that collecting gravity field data helps to infer the characteristics of underground structures, such as faults and salt domes, thereby revealing the spatial distribution and potential reserves of oil reservoirs and providing important support for oil exploration.

[0081] S2: Analyze geological information data and construct a three-dimensional geological model.

[0082] The steps for constructing a three-dimensional geological model include:

[0083] S201: Preprocessing geological information data;

[0084] S202: Perform time-depth conversion on the preprocessed seismic data to obtain depth data;

[0085] S203: Construct a three-dimensional geological model based on depth data;

[0086] S204: Perform seismic inversion on the preprocessed seismic data to obtain velocity data;

[0087] S205: Perform magnetic field inversion on the preprocessed magnetic field data to obtain magnetization data;

[0088] S206: Perform gravity inversion on the preprocessed gravity field data to obtain density data;

[0089] S207: Update the three-dimensional geological model based on velocity, magnetization, and density data.

[0090] In step S201 above, the method for preprocessing seismic data includes:

[0091] A rectangular coordinate system is constructed for the area to be explored and marked as the regional coordinate system. The origin of the regional coordinate system is selected by those skilled in the art based on the actual situation. A grid area is preset, which is pre-set by those skilled in the art based on the area to be explored, and the area to be explored is obtained by those skilled in the art through on-site measurement. The area to be explored is divided into grids according to the grid area. The coordinates of each grid point and seismic point are obtained according to the regional coordinate system. The grid point is the center point of each grid, and the seismic point is the location point in the area to be explored corresponding to each reflection time in the seismic data. The Euclidean distance between each grid point and each seismic point is calculated based on the coordinates of each grid point and each seismic point, and is used as the influence distance.

[0092] A preset impact factor is established, which is pre-set by those skilled in the art based on actual conditions. Based on the impact factor and the impact distance, a weighting coefficient between each grid point and each seismic point is calculated. The expression for the weighting coefficient is: In the formula, qz(p,q) is the weight coefficient between the p-th grid point and the q-th seismic point, yx(p,q) is the influence distance between the p-th grid point and the q-th seismic point, a is the influence factor, p∈[1,P], q∈[1,Q], P is the number of grid points, and Q is the number of seismic points. Based on the weight coefficient between each grid point and each seismic point, the reflection time of each seismic point is weighted and summed to calculate the reflection time corresponding to each grid point, and this summed time is marked as the weighted time. The expression for the weighted time is: In the formula, js p For the weighted time of the p-th grid point, fs q Let q be the reflection time of the q-th earthquake point.

[0093] Methods for preprocessing magnetic field data include:

[0094] The number of magnetic field intensities in the magnetic field data is counted and marked as the sampling number; the background intensity is obtained by summing each magnetic field intensity in the magnetic field data and dividing by the sampling number; the background intensity is subtracted from each magnetic field intensity in the magnetic field data to obtain the magnetic anomaly data corresponding to each magnetic field point. The magnetic field point is the location point in the area to be explored corresponding to each magnetic field intensity in the magnetic field data.

[0095] Methods for preprocessing gravity field data include:

[0096] The exploration elevation is obtained and the topographic influence value is calculated. The exploration elevation is obtained through the GPS positioning system built into the exploration robot. The topographic influence value is subtracted from each gravity point in the gravity field data to obtain the gravity correction value corresponding to each gravity point. The gravity point is the location point in the area to be explored corresponding to each gravity in the gravity field data. The expression for the topographic influence value is: dy = ρgh; where dy is the topographic influence value and ρ is the crustal density, usually 2.67 g / cm³. 3 g is the standard gravitational force, typically 9.81 m / s². 2 h represents the exploration elevation; crustal density and standard gravity were obtained by those skilled in the art through consulting relevant geological research data or literature.

[0097] In step S202 above, the method for obtaining depth data includes:

[0098] Preliminary velocity data is obtained, which is the estimated wave velocity distribution in the subsurface of the area to be explored. The preliminary velocity data is obtained by those skilled in the art based on geological exploration data of adjacent areas of the area to be explored. Based on the preliminary velocity data, the estimated wave velocity of each grid point is obtained. The estimated wave velocity of each grid point is multiplied by the corresponding weighted time and then divided by 2 to obtain the depth value of each grid point. The depth value of each grid point is used as the depth data.

[0099] In step S203 above, the method for constructing a three-dimensional geological model includes:

[0100] Based on the coordinates and depth values ​​of each grid point, construct the three-dimensional coordinates of each grid point; based on the three-dimensional coordinates of each grid point, use a three-dimensional grid reconstruction algorithm (such as Delunay triangulation algorithm, Marching Cubes algorithm, etc.) to transform it into a three-dimensional geological model.

[0101] In step S204 above, the seismic inversion method is such as least squares inversion, Tikhonov regularization inversion, etc.; the velocity data is the actual wave velocity distribution in the subsurface of the area to be explored.

[0102] In step S205 above, the magnetic field inversion method is such as least squares inversion, adaptive filtering inversion, geomagnetic tensor inversion, etc.; the magnetization data is the magnetic susceptibility distribution of the underground area to be explored.

[0103] In step S206 above, gravity inversion methods include, for example, least squares inversion, wave inversion, boundary element method, etc.; density data is the density distribution of the underground area to be explored.

[0104] In step S207 above, the method for updating the three-dimensional geological model includes:

[0105] Velocity, magnetization, and density data are used as analytical data. This data is input into a trained geological analysis model to predict corresponding geological labels. Geological labels are numerical tags corresponding to the geological data; different geological data have different labels. Geological data includes stratum type, stratum depth, oil reservoir area, and oil reservoir depth. Specifically, stratum type refers to the type of each stratum in the area to be explored, such as sandstone, limestone, and shale; stratum depth is the vertical distance from the surface of the area to be explored to the top of each stratum; oil reservoir area is a grid of oil-bearing areas in the area to be explored; oil reservoir depth is the vertical distance from the surface of the area to be explored to the bottom of each oil-bearing stratum, which is the stratum containing oil. Based on the geological labels, corresponding geological data are obtained, and the 3D geological model is updated.

[0106] The training process for a geological analysis model includes:

[0107] B sets of analysis data are collected in advance, and corresponding geological labels are set for each set of analysis data, where b is an integer greater than 1. The analysis data and corresponding geological labels are converted into a set of feature vectors. The geological labels corresponding to the analysis data are collected by those skilled in the art during the construction of historical 3D geological models. Under the conditions of each set of analysis data, the exploration area corresponding to each set of analysis data is explored to obtain the geological data corresponding to each exploration area. According to the geological data corresponding to each exploration area, the corresponding geological labels are set for the b sets of analysis data in sequence.

[0108] Each set of feature vectors is used as input to the geological analysis model, which outputs a set of predicted geological labels corresponding to each set of analyzed data. The actual geological labels corresponding to each set of analyzed data are used as the prediction target; these actual geological labels are the pre-set geological labels corresponding to the analyzed data. The training objective is to minimize the sum of prediction errors for all analyzed data. The prediction error is calculated using the formula η. w =(β) w -ε w ) 2 , where η w Let w be the prediction error, w be the group number of the feature vector corresponding to the analyzed data, and β be the prediction error. w Let ε be the predicted geological label corresponding to the w-th set of analysis data. w Let w be the actual geological label corresponding to the w-th set of analysis data; train the geological analysis model until the sum of prediction errors converges and then stop training.

[0109] The geological analysis model described above is specifically a deep neural network model, which includes an input layer, hidden layers, and an output layer. Each hidden layer contains multiple neurons, and each neuron is connected to the neurons in the next layer. The connections contain weights that determine the importance and influence of the data transmitted in the neural network. An activation function is applied to each neuron between the hidden layer and the output layer. The activation function introduces non-linearity, allowing the network to learn more complex patterns and features.

[0110] S3: Based on a three-dimensional geological model, plan the exploration route.

[0111] like Figure 2 As shown, the steps for planning an exploration path include:

[0112] Step S301: Treat all grids as nodes and the oil reservoir area as an oil node;

[0113] Step S302: Randomly select an oil node and mark it as the initial node; starting from the initial node, explore all the adjacent nodes of the initial node, and select one node from all the adjacent nodes as the successor node according to the preset filtering rules, and mark the successor node as the filtering node.

[0114] Step S303: Explore all neighboring nodes of the successor node, and select one node from all neighboring nodes as the update node according to the preset filtering rules, update the successor node as the update node, and mark the update node as the filter node.

[0115] Step S304: Repeat step S303 until all adjacent nodes corresponding to the successor node are marked as filter nodes, then the loop ends and the initial path is obtained;

[0116] Step S305: Determine whether the initial path includes all oil nodes; if yes, retain the initial path; otherwise, delete the initial path.

[0117] Step S306: Repeat steps S302 to S305 to obtain m1 initial paths, where m1 is an integer greater than 1, and each of the m1 initial paths is different.

[0118] Step S307: Calculate the cumulative depth gradient of each initial path, and select m2 initial paths. Mark all m2 initial paths as intermediate paths, where 1 < m2 < m1.

[0119] Step S308: Calculate the path distance of each intermediate path, select the best path, and use the best path as the exploration path.

[0120] In step S302 above, the selection rule is as follows: if there is an oil node among the adjacent nodes, then one of the oil nodes is randomly selected as the successor node; if there is no oil node among the adjacent nodes, then one of the nodes is randomly selected as the successor node.

[0121] In step S307 above, the expression for the accumulated depth gradient is: In the formula, ls d Let sd be the cumulative depth gradient of the d-th initial path, d∈[1,m1], sd dk Let sd be the oil reservoir depth of the k-th oil node in the d-th initial path. dk+1 Let K be the oil reservoir depth of the (k+1)th oil node in the d-th initial path, where k∈[1,K-1] and K is the number of oil nodes;

[0122] Methods for selecting m2 initial paths include:

[0123] Sort the cumulative depth gradients of each initial path from smallest to largest, select the first m2 cumulative depth gradients in ascending order, and mark them as the filtering depths; select the m2 initial paths corresponding to the filtering depths from the m1 initial paths.

[0124] In step S308 above, the path distance of the intermediate path is the number of all nodes traversed by the intermediate path.

[0125] Methods for selecting the best path include:

[0126] Sort the path distances of each intermediate path from smallest to largest, select the path distance that is first in the ascending order, and mark it as the shortest distance; then select the intermediate path corresponding to the shortest distance from m2 intermediate paths.

[0127] S4: Control the exploration robot to perform drilling operations and collect drilling data according to the exploration path.

[0128] Methods for controlling exploration robots to perform drilling operations include:

[0129] The oil reservoir area to be drilled is marked as the current area. Based on the oil reservoir depth in the current area, the maximum drilling depth for the current area is determined. The current drilling depth is acquired in real time; the current drilling depth is the vertical distance of the drill bit relative to the ground to be explored. The drill bit is integrated into the drilling equipment within the exploration robot, and the current drilling depth is obtained through a depth sensor integrated into the drill bit. The rock strata depths corresponding to the current area are marked as comparison depths. All comparison depths are sorted with the current drilling depth in ascending order. The comparison depth preceding the current drilling depth is obtained according to the ascending order and marked as the determined depth. The rock strata type corresponding to the determined depth is marked as the determined type. The drilling parameters of the exploration robot are adjusted in real time based on the current drilling depth and the determined type. When the current drilling depth reaches the maximum drilling depth, the exploration robot is controlled to stop drilling operations.

[0130] The steps for real-time control of drilling parameters of an exploration robot include:

[0131] Step S401: Construct N parameter sets, set different numerical labels for the N parameter sets, and mark them as parameter labels;

[0132] Step S402: Construct a population S containing n individuals, generate the position of each individual, and assign each individual position a one-to-one correspondence with the parameter label. Set the iteration count t to 0.

[0133] Step S403: Determine the attraction function and the iteration threshold T;

[0134] Step S404: Calculate the attractiveness of each individual in population S and update the position of each individual;

[0135] Step S405: Determine whether the iteration number t is less than the iteration threshold T. If yes, let t = t + 1 and return to step S404. If no, proceed to step S406.

[0136] Step S406: Calculate the attraction degree corresponding to each individual in population S, obtain the parameter label corresponding to the individual with the highest attraction degree, obtain the corresponding parameter set based on the obtained parameter label, and adjust the drilling parameters of the exploration robot in real time based on the obtained parameter set.

[0137] In step S401 above, the method for constructing N parameter sets is as follows: Obtain the parameter range, which includes the range of each parameter in the drilling parameters. The parameter range is obtained by those skilled in the art based on the technical parameters of the drilling equipment integrated into the exploration robot. The drilling parameters include drilling speed, drill bit rotation speed, drill bit pressure, and drill bit temperature. Drilling speed is the speed at which the drill bit penetrates the rock formation; drill bit rotation speed is the number of rotations of the drill bit per unit time (e.g., per second, per minute); drill bit pressure is the vertical pressure applied to the drill bit; and drill bit temperature is the heat generated by the drill bit during drilling. Randomly select a value from each range in the parameter range and construct a parameter set, resulting in a total of N parameter sets, all of which are distinct.

[0138] In step S402 above, the expression for the location of each individual is: In the formula, Let C be the position of the i-th individual. i Let C be the random coefficient of the i-th individual. i ∈[0,1],i∈[1,n].

[0139] In step S403 above, the expression for the attraction function is: f = tz; where f is the attraction and tz is the drilling quality. The drilling quality is obtained by setting different numerical labels for different rock strata types and marking them as rock strata labels; obtaining the corresponding parameter set based on the parameter label corresponding to the individual location; using the current drilling depth, the rock strata label corresponding to the determined type, and the parameter set corresponding to the individual location as evaluation data, inputting the evaluation data into the trained quality evaluation model to evaluate the corresponding drilling quality. The drilling quality is a comprehensive index of drilling efficiency, drilling safety, and drilling cost. The training process of the quality evaluation model is consistent with the training process of the geological analysis model, and both are deep neural network models. The drilling quality corresponding to the evaluation data is recorded by a person skilled in the art when historically adjusting drilling parameters. Under the conditions of the current drilling depth and determined type in each set of evaluation data, the exploration robot is controlled to perform drilling operations according to the parameter set, and the drilling operation results are analyzed through three dimensions: drilling efficiency, drilling safety, and drilling cost to evaluate the corresponding drilling quality. A corresponding drilling quality is set for each set of evaluation data.

[0140] In S404 above, the method for updating the location of each individual includes:

[0141]

[0142] In the formula, Let v be the position of the i-th individual after the update, v be the maximum attraction, y be the decay factor, e be the natural constant, and l be the position of the i-th individual after the update. ib Let be the distance between the i-th individual and the optimal individual. The position of the optimal individual. The position of the i-th individual before the update is given by L, which is the perturbation factor, and R is a random number between [0, 1]. The optimal individual is the individual with the highest attraction in the population S. The maximum attraction, decay factor, and perturbation factor are preset by those skilled in the art according to the actual situation.

[0143] Drilling data includes lithological data, porosity data, and permeability data;

[0144] The lithological data represents the actual type of the oil-bearing strata; the lithological data was obtained through methods such as gamma-ray logging and sonic logging.

[0145] Porosity data is the proportion of pore volume to total volume in an oil-bearing rock formation; porosity data is obtained through methods such as neutron porosity logging and density porosity logging.

[0146] Permeability data represents the ability of rocks in oil-bearing formations to allow fluids to pass through; permeability data are obtained using the sonic transit time method and the differential pressure method.

[0147] It should be noted that drilling data is crucial for assessing oil reserves and recoverability. Among these data, lithological data determines the type and distribution of oil-bearing strata, porosity data determines the oil storage capacity of the strata, and permeability data determines the fluidity and extraction efficiency of the oil within the strata. Through comprehensive analysis of drilling data, a complete assessment of oil-bearing strata can be achieved, accurately evaluating oil reserves and recoverability, which helps in developing optimal extraction plans.

[0148] S5: Analyze drilling data to assess oil resources.

[0149] Oil resources include oil reserves and recoverability.

[0150] Methods for assessing oil resources include:

[0151] The 3D geological model is divided into grids based on the grid area, and the divided grids are marked as cells. Drilling data is added to the corresponding cells to obtain the simulated geological model. Numerical simulation methods (such as reservoir simulation software such as Eclipse, CMG, and Schlumberger) are used to dynamically simulate the simulated geological model to obtain oil reserves and recoverability.

[0152] This embodiment integrates multi-source geological information data such as seismic, magnetic field, and gravity data. By constructing a three-dimensional geological model and planning the optimal exploration path, it achieves automation and intelligence in the oil exploration process. Furthermore, by adjusting drilling parameters in real time, it intelligently controls the exploration robot to perform drilling operations, thereby significantly improving the efficiency and accuracy of oil exploration and reducing operational risks and costs. At the same time, it uses dynamic simulation calculations to accurately assess the oil reserves and exploitability of oil-bearing strata based on the drilling data collected during the drilling operations, achieving a systematic assessment of overall oil resources and providing a reliable basis for subsequent oil development and utilization.

[0153] Example 2

[0154] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code that, when executed by the one or more processors, can perform an oil exploration method employing an intelligent oil exploration robot system.

[0155] The method or system according to the embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as a ROM or hard disk, may store an oil exploration method using an intelligent oil exploration robot system provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components of the electronic device shown in this application may be omitted according to actual needs.

[0156] Example 3

[0157] Please refer to the accompanying drawings. One embodiment of this application discloses a computer-readable storage medium. The computer-readable storage medium stores computer-readable instructions. When the computer-readable instructions are executed by a processor, an oil exploration method employing an intelligent oil exploration robot system, as described in the above-described embodiments of this application, can be performed. The storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0158] Furthermore, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, this application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be executed by a processor to perform instructions corresponding to the method steps provided in this application, such as an oil exploration method employing an intelligent oil exploration robot system. When this computer program is executed by a central processing unit (CPU), it performs the functions defined in the method of this application.

[0159] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0160] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method of oil exploration using an intelligent oil exploration robot system, characterized by, The method comprises the following steps: S1: collecting geological information data; the geological information data comprises seismic data, magnetic field data and gravity field data; the seismic data is the seismic wave reflection time distribution of the underground of a to-be-explored region; the to-be-explored region is a region planned to be explored for oil; the magnetic field data is the magnetic field intensity distribution of the ground of the to-be-explored region; and the gravity field data is the gravity distribution of the ground of the to-be-explored region; S2: analyzing the geological information data and constructing a three-dimensional geological model; the step of constructing the three-dimensional geological model comprises: S201: preprocessing the geological information data; S202: performing time-depth conversion on the preprocessed seismic data to obtain depth data; S203: constructing a three-dimensional geological model according to the depth data; S204: performing seismic inversion on the preprocessed seismic data to obtain velocity data; S205: performing magnetic field inversion on the preprocessed magnetic field data to obtain magnetization data; S206: performing gravity inversion on the preprocessed gravity field data to obtain density data; S207: predicting geological data according to the velocity data, the magnetization data and the density data, and updating the three-dimensional geological model according to the geological data; the geological data comprises rock layer types, rock layer depths, oil reservoir regions and oil reservoir depths; S3: planning an exploration path based on the three-dimensional geological model; the step of planning the exploration path comprises: step S301: presetting a grid area, and dividing the to-be-explored region into grids according to the grid area; taking all the grids as nodes, and taking the oil reservoir regions as oil nodes; step S302: randomly selecting an oil node and marking it as an initial node; exploring all adjacent nodes of the initial node, and selecting one node from all the adjacent nodes as a successor node according to a preset screening rule, and marking the successor node as a screened node; step S303: exploring all adjacent nodes of the successor node, and selecting one node from all the adjacent nodes as an updated node according to the preset screening rule, updating the successor node to the updated node, and marking the updated node as the screened node; step S304: repeating step S303 until all adjacent nodes corresponding to the successor node are marked as the screened node, and then the cycle is ended, and an initial path is obtained; step S305: determining whether the initial path comprises all the oil nodes; if yes, the initial path is reserved, and if no, the initial path is deleted; Step S306: Circulating step S302-step S305, obtaining the initial path, is an integer greater than 1, the initial paths are all different. Step S307: Calculate the cumulative depth gradient of each initial path, and select out the initial path with the maximum cumulative depth gradient as the intermediate path. Step S308: Calculate the cumulative depth gradient of each initial path, and select out the initial path with the maximum cumulative depth gradient as the intermediate path. Step S309: Mark each initial path with the maximum cumulative depth gradient as the intermediate path. Step S310: Calculate the cumulative depth gradient of each initial path, and select out the initial path with the maximum cumulative step S308: counting the path distances of each intermediate path, and screening out a best path, and taking the best path as the exploration path; S4: controlling an exploration robot to perform drilling work according to the exploration path, and collecting drilling data; the method for controlling the exploration robot to perform drilling work comprises: Mark the oil reservoir area where drilling operation is to be performed as a current area, determine a maximum drilling depth of the current area according to a depth of the oil reservoir of the current area; acquire a current drilling depth in real time, the current drilling depth being a vertical distance of a drill bit relative to a ground to be explored; mark depths of strata corresponding to the current area as comparison depths, sort all the comparison depths and the current drilling depth from small to large, acquire a comparison depth ranked in a first position before the current drilling depth according to a positive order, and mark the comparison depth as a determined depth, mark a strata type corresponding to the determined depth as a determined type; control drilling parameters of the exploration robot in real time according to the current drilling depth and the determined type; when the current drilling depth reaches the maximum drilling depth, control the exploration robot to stop drilling operation; S5: analyze the drilling data and evaluate the oil resources.

2. The oil exploration method employing the intelligent oil exploration robot system according to claim 1, characterized in that, In the step S201, the method for pre-processing the seismic data includes: Construct a rectangular coordinate system for the area to be explored and mark it as an area coordinate system; acquire coordinates of each grid point and seismic point according to the area coordinate system, the grid point being a center point of each grid, and the seismic point being a position point in the area to be explored corresponding to each reflection time in the seismic data; calculate Euclidean distances between each grid point and each seismic point according to the coordinates of each grid point and seismic point, and take the Euclidean distances as influence distances; A preset influence factor is used. Based on the influence factor and influence distance, a weighting coefficient between each grid point and each seismic point is calculated. The expression for the weighting coefficient is: In the formula, For the first The grid point and the first Weighting coefficients between earthquake points For the first The grid point and the first The influence distance between earthquake points As the impact factor, , , For the number of grid points, The number of seismic points; based on the weighting coefficient between each grid point and each seismic point, the reflection time of each seismic point is weighted and summed to calculate the reflection time corresponding to each grid point, and this summation is marked as the weighted time; the expression for the weighted time is: In the formula, For the first Weighted time for each grid point For the first Reflection time at each earthquake point; The method for pre-processing the magnetic field data includes: Count the number of magnetic field strengths in the magnetic field data and mark it as a sampling number; add each magnetic field strength in the magnetic field data in turn, and then divide the sum by the sampling number to obtain a background strength; subtract the background strength from each magnetic field strength in the magnetic field data to obtain magnetic anomaly data corresponding to each magnetic field point, the magnetic field point being a position point in the area to be explored corresponding to each magnetic field strength in the magnetic field data; The method for pre-processing the gravity field data includes: Obtain the exploration altitude, and calculate the terrain influence value; subtract each gravity in the gravity field data from the terrain influence value respectively to obtain the gravity correction value corresponding to each gravity point, and the gravity point is the position point in the to-be-explored area corresponding to each gravity in the gravity field data; the expression of the terrain influence value is: ; in the formula, the terrain influence value, the crust density, the standard gravity, the exploration altitude.

3. The oil exploration method employing the intelligent oil exploration robot system according to claim 2, characterized in that, In the step S202, the method for acquiring the depth data includes: Acquire preliminary velocity data, the preliminary velocity data being an estimated wave velocity distribution of the area to be explored underground, acquire an estimated wave velocity of each grid point according to the preliminary velocity data; multiply the estimated wave velocity of each grid point by a corresponding weighting time, and then divide the product by 2 to obtain a depth value of each grid point, and take the depth value of each grid point as the depth data; In the step S203, the method for constructing the three-dimensional geological model includes: Construct a three-dimensional coordinate of each grid point according to the coordinates and the depth value of each grid point; convert the three-dimensional coordinate of each grid point into a three-dimensional geological model by using a three-dimensional grid reconstruction algorithm; In the step S207, the method for updating the three-dimensional geological model includes: The speed data, magnetization data and density data are taken as analysis data, the analysis data is input into the trained geological analysis model, and corresponding geological labels are predicted; the geological labels are digital labels corresponding to the geological data, and the geological labels corresponding to different geological data are all different; wherein, the rock layer type is the type of each rock layer in the area to be explored; the rock layer depth is the vertical distance from the ground of the area to be explored to the top of each rock layer; the oil reservoir area is the grid in the area to be explored where oil exists; the oil reservoir depth is the vertical distance from the ground of the area to be explored to the bottom of each oil-bearing rock layer; according to the geological labels, corresponding geological data is obtained, and the three-dimensional geological model is updated; the training process of the geological analysis model comprises: Pre-collection Group analysis data, for Each set of analytical data is assigned a corresponding geological label, and the analytical data and the corresponding geological labels are converted into a set of feature vectors. Each set of feature vectors is used as input to the geological analysis model, which outputs a set of predicted geological labels corresponding to each set of analytical data and uses the actual geological labels corresponding to each set of analytical data as the prediction target. The actual geological labels are the pre-set geological labels corresponding to the analytical data. The training objective is to minimize the sum of prediction errors of all analytical data. The geological analysis model is trained until the sum of prediction errors converges, at which point training stops. The geological analysis model is a deep neural network model.

4. The method of oil exploration using intelligent oil exploration robot system as claimed in claim 3 wherein, In the step S302, the screening rule is: if there is an oil node in the adjacent nodes, randomly select one of the oil nodes as the successor node, and if there is no oil node in the adjacent nodes, randomly select one of the nodes as the successor node; In the step S307, the expression of the accumulated depth gradient is: ; wherein, is the accumulated depth gradient of the initial path of the i-th oil node, , is the oil reservoir depth of the i-th oil node in the initial path of the i-th oil node, is the oil reservoir depth of the i-th oil node in the initial path of the i-th oil node, is the oil reservoir depth of the i-th oil node in the initial path of the i-th oil node, , is the number of oil nodes;​​​​ Screening out The method of strip initial path includes: Sort the accumulated depth gradients of each initial path from smallest to largest, and select the path with the highest gradient in ascending order. The cumulative depth gradient of each is calculated and marked as the filtering depth; from Filtering from the initial path The initial path corresponding to each filtering depth; In the step S308, the path distance of the intermediate path is the number of all nodes passed through by the intermediate path; The method for screening the best path comprises: The path distance of each intermediate path is sorted from small to large, the path distance ranked first is selected according to the positive order, and is marked as the shortest distance; and The intermediate path corresponding to the shortest distance is selected from the intermediate paths.

5. The method of oil exploration using intelligent oil exploration robot system as claimed in claim 4 wherein, The step of real-time regulating and controlling the drilling parameters of the exploration robot comprises: Step S401: constructing N parameter sets, setting different digital labels for the N parameter sets, and marking as parameter labels; Step S402: Constructing a population , the population includes n individuals, generate the position of each individual, the individual position is one-to-one corresponding to the parameter label, and set the iteration number t as 0; Step S403: determining the attraction degree function and the iteration threshold T; Step S404: Calculate the population the attraction degree of each individual, and update the position of each individual; Step S405: judging whether the iteration number t is less than the iteration threshold T, if yes, then letting and returning to step S404, if not, then entering step S406; Step S406: calculating the population The attraction degree corresponding to each individual is calculated, the parameter label corresponding to the individual with the maximum attraction degree is obtained, the corresponding parameter set is obtained according to the obtained parameter label, and the drilling parameters of the exploration robot are controlled in real time according to the obtained parameter set.

6. The method of oil exploration using intelligent oil exploration robot system as claimed in claim 5 wherein, In the step S401, the method for constructing the N parameter sets is: obtaining a parameter range, the parameter range comprising the range of each parameter in the drilling parameters; the drilling parameters comprising drilling speed, drill bit rotation speed, drill bit pressure and drill bit temperature; randomly selecting a value from each range in the parameter range, and constructing a parameter set, a total of N parameter sets, the N parameter sets being all different; In the step S402, the expression of each individual position is: wherein, is the position of the i-th individual, is the random coefficient of the i-th individual, , ; In the step S403, the expression of the attraction degree function is: ; wherein, is the attraction degree, is the drilling quality; the method for obtaining the drilling quality is: setting different numerical labels for different rock types and marking them as rock labels; obtaining a corresponding parameter set according to the parameter label corresponding to the individual position; taking the current drilling depth, the rock label corresponding to the determined type and the parameter set corresponding to the individual position as evaluation data, inputting the evaluation data into the trained quality evaluation model, evaluating the corresponding drilling quality, and the drilling quality is a comprehensive index of drilling efficiency, drilling safety and drilling cost; the training process of the quality evaluation model is consistent with the training process of the geological analysis model, and both are deep neural network models; In the S404, the method for updating the position of each individual comprises: ； wherein, is the position of the i-th individual after update, is the maximum attraction, is the attenuation factor, is the natural constant, is the distance between the i-th individual and the best individual, is the position of the best individual, is the position of the i-th individual before update, is the perturbation factor, is the random number between 0 and 1, is the position of the best individual in the population with the maximum attraction.

7. The method of oil exploration using intelligent oil exploration robot system as claimed in claim 6 wherein, The drilling data comprises lithology data, porosity data and permeability data; the lithology data is the actual type of the oil-bearing rock layer; the porosity data is the ratio of the pore volume to the total volume in the oil-bearing rock layer; the permeability data is the ability of the rock in the oil-bearing rock layer to allow fluid to pass through; and the oil resources comprise oil reserves and extractability; The method for evaluating the oil resources comprises: According to the grid area, the three-dimensional geological model is divided into grids, the divided grids are marked as unit cells; the drilling data is added to the corresponding unit cells, a simulation geological model is obtained; a numerical simulation calculation method is used to dynamically simulate the simulation geological model, and oil reserves and extractability are obtained.

Images