Ocean drilling accident prediction method, system and device and storage medium

Abstract

The invention discloses an ocean drilling accident prediction method, system and device and a storage medium, and the method comprises the steps: obtaining the monitoring data of ocean drilling, recognizing the abnormal data in the monitoring data, assigning a preset accident to a current to-be-classified accident, carrying out the binary classification of the current to-be-classified accident through a trained support vector machine model and the abnormal data, and obtaining the current to-be-classified accident. And determining whether the current prediction accident satisfies a condition, if the current prediction accident does not satisfy the condition, assigning the current prediction accident to the current to-be-classified accident, and returning to perform binary classification on the current to-be-classified accident by using the trained support vector machine model and the abnormal data to obtain the current prediction accident. And outputting the current predicted accident meeting the condition as a final predicted drilling accident until the current predicted accident meeting the condition is obtained. The method can improve the information perception capability and safety of ocean drilling operation, and can be widely applied to the field of intelligent control.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control, and in particular to a prediction method, system, device and storage medium for ocean drilling accidents. Background Art

[0002] Ocean drilling ships need to operate in the deep sea far away from the land. Different from general land drilling, due to the complex and changeable operating environment, drilling ships need to complete drilling work under complex conditions such as wind, waves, and currents, which puts extremely high demands on the safety and reliability of drilling operations.

[0003] Existing drilling accident prediction methods lack information perception capabilities for ocean drilling operations and have low safety. Summary of the invention

[0004] In order to solve one of the above problems, the purpose of the present invention is to provide a prediction method, system, device and storage medium for ocean drilling accidents, which can improve the information perception capability and safety of ocean drilling operations.

[0005] In one aspect, the present invention provides a method for predicting an ocean drilling accident, comprising the following steps:

[0006] Acquiring ocean drilling monitoring data and identifying abnormal data in the monitoring data; the monitoring data includes drilling data, ship attitude data, remote controlled unmanned submersible data, underwater blowout preventer data or dynamic positioning data;

[0007] Assigning a preset accident value to the current accident to be classified, and using the trained support vector machine model and the abnormal data to perform binary classification on the current accident to be classified to obtain the current predicted accident;

[0008] Determine whether the current predicted accident meets the conditions. If the current predicted accident does not meet the conditions, assign the current predicted accident to the current accident to be classified, return to use the trained support vector machine model and the abnormal data to perform binary classification on the current accident to be classified, and obtain the current predicted accident until the current predicted accident that meets the conditions is obtained, and output the current predicted accident that meets the conditions as the final predicted drilling accident.

[0009] Optionally, the identifying abnormal data in the monitoring data specifically includes:

[0010] The monitoring data is input into a preset abnormality recognition model, and the monitoring data that deviates from the normal data range in the preset abnormality recognition model is identified as abnormal data; the preset abnormality recognition model includes a drilling model, a ship attitude model, a remote-controlled unmanned submersible model, underwater blowout preventer data and a dynamic positioning model.

[0011] Optionally, the trained support vector machine model is obtained by the following method:

[0012] Get the original data set and assign it to the current training data set;

[0013] The parameters of the support vector machine model are updated using the current training data set and the particle swarm algorithm to obtain the trained support vector machine model.

[0014] Optionally, the method further includes updating the parameters of the support vector machine model according to the final predicted drilling accident, wherein the updating the parameters of the support vector machine model according to the final predicted drilling accident specifically includes:

[0015] Determine whether the final predicted drilling accident is correct. If the final predicted drilling accident is incorrect, correct the final predicted drilling accident, obtain an original data set, and add the corrected drilling accident and the monitoring data to the original data set to obtain a current training data set;

[0016] The parameters of the support vector machine model are updated using the current training data set and the particle swarm algorithm.

[0017] Optionally, the updating of the parameters of the support vector machine model using the current training data set and the particle swarm algorithm specifically includes:

[0018] Randomly initializing the current particle position and current particle velocity of the particle swarm algorithm; the particle position represents a solution of the parameters of the support vector machine model;

[0019] Calculating the fitness of each current particle position according to the current particle position and the current training data set;

[0020] Comparing the fitness of the current particle position with the fitness of the historical local optimal position and the fitness of the historical global optimal position, respectively, to obtain the current local optimal position and the current global optimal position;

[0021] Determine whether the current global optimal position meets a preset condition, and if not, update the current particle position according to the current local optimal position, the current global optimal position and the current particle velocity, and update the current particle velocity;

[0022] The fitness of the current local optimal position is assigned to the fitness of the historical local optimal position, and the fitness of the current global optimal position is assigned to the fitness of the historical global optimal position, and the process returns to execute the fitness calculation of each current particle position according to the current particle position and the current training data set until the preset condition is met, and the current global optimal position that meets the preset condition is the updated parameter of the support vector machine model.

[0023] Optionally, the calculating the fitness of each current particle position according to the current particle position and the current training data set specifically includes:

[0024] Assigning each current particle position to the parameters of the support vector machine model respectively, inputting the current training data set into the assigned support vector machine model, and obtaining N samples to predict drilling accidents; wherein N is a natural number greater than 2;

[0025] The number of samples correctly predicted among the N samples for predicting drilling accidents is determined, and the fitness of each current particle position is obtained according to the number of samples correctly predicted and the total number of samples in the current training data set.

[0026] Optionally, the step of comparing the fitness of the current particle position with the fitness of a historical local optimal position and the fitness of a historical global optimal position to obtain the current local optimal position and the current global optimal position specifically includes:

[0027] Compare the fitness of each current particle position with the fitness of each historical local optimal position corresponding to the current particle position; if the fitness of the current particle position is better than the fitness of the historical local optimal position corresponding to the current particle position, assign the current particle position to the historical local optimal position corresponding to the current particle position;

[0028] The fitness of all current particle positions is compared with the fitness of the historical global optimal position. If there is any current particle position whose fitness is better than the fitness of the historical global optimal position, the current particle position corresponding to the fitness better than the fitness of the historical global optimal position is assigned to the current global optimal position.

[0029] In another aspect, the present invention provides a method and apparatus for predicting ocean drilling accidents, comprising:

[0030] at least one processor;

[0031] at least one memory for storing at least one program;

[0032] When the at least one program is executed by the at least one processor, the at least one processor implements the method described above.

[0033] In another aspect, the present invention provides a computer-readable storage medium, which stores a program executable by a processor. When the program executable by the processor is executed by the processor, it is used to perform the above-mentioned method.

[0034] On the other hand, the present invention provides a prediction system for ocean drilling accidents, comprising a data layer, a decision layer and a user layer, wherein:

[0035] The data layer is used to collect monitoring data of ocean drilling and upload the monitoring data to the decision layer;

[0036] The decision layer is used to implement the above method and push the obtained data to the user layer;

[0037] The user layer is used to display the monitoring data and / or the data pushed by the decision layer.

[0038] The implementation of the present invention includes the following beneficial effects: the present invention obtains monitoring data of ocean drilling and identifies abnormal data in the monitoring data; the monitoring data includes drilling data, ship attitude data, remote control unmanned submersible data, underwater blowout preventer data or dynamic positioning data, assigns a preset accident to the current accident to be classified, uses the trained support vector machine model and the abnormal data to perform binary classification on the current accident to be classified, obtains the current predicted accident, determines whether the current predicted accident meets the conditions, and if the current predicted accident does not meet the conditions, assigns the current predicted accident to the current accident to be classified, and returns the trained support vector machine model. The model and abnormal data are used to classify the current accident to be classified into two categories to obtain the current predicted accident, until the current predicted accident that meets the conditions is obtained, and the current predicted accident that meets the conditions is output as the final predicted drilling accident; by monitoring all aspects of data, all-round monitoring is achieved, which can improve the safety of drilling operations, by first identifying the acquired data, identifying abnormal data, and then using the abnormal data to classify drilling accidents, and using the support vector machine model to gradually refine the classification of drilling accidents until the predicted drilling accident is obtained, the processing capability of monitoring data can be improved, thereby improving the information perception capability and safety of ocean drilling operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of the steps of a method for predicting ocean drilling accidents provided by the present invention;

[0040] Figure 2 It is a flow chart of steps for obtaining a trained support vector machine model provided by the present invention;

[0041] Figure 3 It is a flow chart of steps for updating parameters of a support vector machine model provided by the present invention;

[0042] Figure 4 It is a flow chart of steps for calculating the fitness of particle position provided by the present invention;

[0043] Figure 5It is a flowchart of the steps of updating the local optimal position and the global optimal position provided by the present invention;

[0044] Figure 6 It is an iterative flow chart of a particle swarm algorithm provided by the present invention;

[0045] Figure 7 It is a flow chart of the steps of updating the support vector machine model according to the final predicted drilling accident provided by the present invention;

[0046] Figure 8 It is a structural schematic diagram of a prediction system for ocean drilling accidents provided by the present invention;

[0047] Fig. 9 It is a structural schematic diagram of a method and device for predicting ocean drilling accidents provided by the present invention;

[0048] Fig.10 It is a structural schematic diagram of another ocean drilling accident prediction system provided by the present invention. DETAILED DESCRIPTION

[0049] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only provided for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0050] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the embodiments of the present application. The singular forms of "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0051] When the following description relates to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. In the description of the present application, it is necessary to understand that the terms "first / S100", "second / S200", "third / S300", etc. are only used to distinguish similar objects, and do not have to be used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0052] In order to better understand the technical solution of the present invention, the following technical terms are explained:

[0053] Support Vector Machines (SVM) model: It is a binary classification model. Its classification method is to find a hyperplane to classify the sample set given a sample set including two categories.

[0054] like Figure 1 As shown, Figure 1 The present invention is a flow chart of the steps of a method for predicting an ocean drilling accident. The present invention provides a method for predicting an ocean drilling accident, comprising the following steps:

[0055] S100: Acquire ocean drilling monitoring data, and identify abnormal data in the monitoring data.

[0056] Among them, the monitoring data includes drilling data, ship attitude data, remote-controlled unmanned submersible data, underwater blowout preventer data or dynamic positioning data or dynamic positioning data.

[0057] The drilling data include but are not limited to drill bit pressure (Kpa), mechanical speed (r / min), rotary table torque (N*m), hook load (maximum weight it can bear), footage (drilling length in a period of time), well depth, well diameter, tripping and drilling time, downhole torque (N*m), cement slurry density (kg / m 3 ), porosity (reflects the pore condition of the matrix. A large porosity indicates that the matrix is ​​lighter and looser, while a small porosity indicates that the matrix is ​​heavier and more solid), resistivity (reflects the conductivity of the conductor), pump pressure, pump displacement, standpipe pressure, casing pressure, injection velocity, annulus flow rate, pump strokes, drilling fluid density, inlet flow rate, outlet flow rate, drilling fluid tripping volume, CO2 content, H2S content, drilling fluid temperature, and drilling fluid level.

[0058] The ship attitude data includes but is not limited to information including ship coordinates, platform / ship speed, heading, roll, pitch, bow roll, sway, sway, heave, draft, etc. Generally, a rectangular coordinate system with the ship's center of gravity as the origin is taken, the X-axis is parallel to the hull baseline and points to the bow, the Z-axis is perpendicular to the hull baseline and points upward, and the Y-axis points to the left side of the hull. The roll, pitch, and bow roll are reciprocating rotations around the X, Y, and Z axes respectively, and sway, sway, and heave are one or more reciprocating rotations along the X, Y, and Z axes respectively.

[0059] Remotely operated unmanned submersible data includes, but is not limited to, one or more of water entry conditions, depth, water temperature, salinity, pressure, speed, heading, and attitude.

[0060] The subsea blowout preventer data includes, but is not limited to, one or more of the positions of the gates, the pressure of the hydraulic system, the temperature, the valve opening pressure, and the valve closing pressure.

[0061] The dynamic positioning data includes, but is not limited to, one or more of the voltage, current, frequency, cooling water temperature, lubricating oil pressure, nacelle azimuth, rotation speed, lubricating oil pressure, hydraulic oil pressure and temperature, positioning time, emergency release time, and the relationship information and displacement information between the dynamic positioning system and the pipe string.

[0062] Specifically, the monitoring data of ocean drilling is obtained, and the monitoring data is classified to obtain various types of data such as drilling data, ship attitude data, remote control unmanned submersible data, underwater blowout preventer data or dynamic positioning data, and the monitoring data are respectively input into the corresponding models. The specific method is as follows:

[0063] The monitoring data is input into a preset abnormality recognition model, and the monitoring data that deviates from the normal data range in the preset abnormality recognition model is identified as abnormal data, wherein the preset abnormality recognition model includes a drilling model, a ship attitude model, a remote-controlled unmanned submersible model, an underwater blowout preventer model, and a dynamic positioning model. For example, the drilling data is input into the drilling model, and abnormal data in the drilling data is identified.

[0064] S200, assigning a preset accident value to the current accident to be classified, and using the trained support vector machine model and the abnormal data to perform binary classification on the current accident to be classified to obtain the current predicted accident.

[0065] Specifically, possible drilling accidents are taken as preset accidents. When classifying, the preset accidents are assigned to the current accident to be classified. The trained support vector machine model and abnormal data are used to classify the current accident to be classified into two categories to obtain the current predicted accident. Among them, the current accident to be classified represents the drilling accident before this classification; the current predicted accident represents the drilling accident predicted by the support vector machine model in this classification.

[0066] S300, determine whether the current predicted accident meets the conditions. If the current predicted accident does not meet the conditions, assign the current predicted accident to the current accident to be classified, return to use the trained support vector machine model and the abnormal data to perform binary classification on the current accident to be classified to obtain the current predicted accident, until the current predicted accident that meets the conditions is obtained, and output the current predicted accident that meets the conditions as the final predicted drilling accident.

[0067] The condition may be, but is not limited to, whether the currently predicted accident will cause the drilling to generate the aforementioned abnormal data.

[0068] Specifically, determine the correlation between the current predicted accident (the drilling accident predicted by this classification) and the abnormal data to determine whether it is the drilling accident corresponding to the abnormal data. If so, end the classification and use the current predicted accident obtained by this classification as the final predicted drilling accident. If not, continue to use the support vector machine to perform binary classification on the current predicted accident obtained by this classification until it is finally determined that the current predicted accident meets the conditions, and output the current predicted accident that meets the conditions as the final drilling accident.

[0069] A classification method of steps S200-300 may be:

[0070] The abnormal data is input into the support vector machine model. During the first classification, the support vector machine model uses the abnormal data to classify the preset accidents into two categories: drilling accidents related to the abnormal data and drilling accidents unrelated to the abnormal data (i.e., drilling accidents will not cause the aforementioned abnormal data in drilling). The drilling accidents unrelated to the abnormal data are excluded, and the drilling accidents related to the abnormal data are used as the current predicted accidents for the first classification.

[0071] During the second classification, one abnormal data or a group of abnormal data can be used as the abnormal data for the second classification. The support vector machine model uses the abnormal data for the second classification to perform a second classification on the current predicted accidents obtained from the first classification, and eliminates drilling accidents that are not related to the abnormal data for the second classification to obtain the current predicted accidents for the second classification.

[0072] During the third classification, the abnormal data of the second classification is excluded, and the same operation as the second classification is continued, and so on, until the current predicted accident that meets the conditions is obtained.

[0073] For example, as shown in Table 1, Table 1 is a diagram of the relationship between some drilling accidents and monitoring data. If there are a total of twenty types of drilling accidents that may occur in ocean drilling, a set of data such as drilling data, ship attitude data, ROV data, underwater blowout preventer data, and dynamic positioning data are input into each abnormal recognition model. The abnormal recognition model recognizes abnormal information such as the turntable torque in the drilling data is greater than the normal value and the drilling temperature rises. First, the support vector machine is used to classify the twenty types of drilling accidents into two categories, and the drilling accident types that are not related to the abnormal data are excluded. After exclusion, the remaining types of drilling accidents may be stuck drill, stuck drill, stuck gear, and slipping drill. Eight drilling accidents, such as stuck drill, emptying, and blocked water holes, are related to abnormal data. The support vector machine model continues to identify the abnormal data one by one to obtain the abnormal data of this classification. For example, the first time it is identified that the riser pressure is less than the normal value, the support vector machine model uses the abnormal data of this classification (such as the riser pressure is less than the normal value) to continue to recursively subdivide the remaining eight drilling accidents. At this time, there are four drilling accidents left, namely, stuck drill, stuck cone, dropped water holes, and blocked water holes. In this way, the scope of drilling accidents is narrowed down according to abnormal data until one or more drilling accidents are finally identified, and the results are output to complete the accident prediction.

[0074]

[0075] In some embodiments, Figure 2 As shown, Figure 2 is a flowchart of steps for obtaining a trained support vector machine model, wherein the trained support vector machine model is obtained by the following method:

[0076] S210: Obtain an original data set, and assign the original data set to the current training data set.

[0077] Among them, the original data set includes multiple sets of drilling data, ship attitude data, remote-controlled unmanned submersible data, underwater blowout preventer data, dynamic positioning data and drilling accidents, and their specific meanings are the same as the above explanations.

[0078] S220: Update the parameters of the support vector machine model using the current training data set and the particle swarm algorithm to obtain the trained support vector machine model.

[0079] Specifically, the particle swarm algorithm is used to obtain the parameters of the support vector machine model, the parameters and the current training data set are substituted into the support vector machine model, the support vector machine model predicts the current training data set under the parameters, and obtains the drilling accidents under the parameters. The parameters are evaluated according to the drilling accidents obtained by the prediction. If the prediction effect is good, the support vector machine model is updated with the parameters. If the prediction effect is not good, the particle swarm algorithm is used again to obtain the parameters of the support vector machine model until the parameters that meet the prediction effect are obtained. The parameters that meet the prediction effect are saved in the support vector machine model to obtain a trained support vector machine model.

[0080] like Figure 3 As shown, Figure 3 The invention is a flowchart of the steps of updating the parameters of the support vector machine model. Step S220 uses the current training data set and the particle swarm algorithm to update the parameters of the support vector machine model, which specifically includes:

[0081] S221, randomly initializing the current particle position and current particle velocity of the particle swarm algorithm.

[0082] The particle positions represent the solutions of the parameters of the support vector machine model.

[0083] Specifically, determine the particle size m, the maximum number of iterations t max , inertia weight ω and learning factor, and randomly initialize the particle position of each particle and particle speed Among them, i = 1, 2, ..., m, the penalty factor c, the error ε, and the width coefficient σ are three parameters to be optimized of the SVM model.

[0084] S222. Calculate the fitness of each current particle position according to the current particle position and the current training data set.

[0085] Specifically, Figure 4 As shown, Figure 4 It is a flow chart of steps for calculating the fitness of a particle position, and S222 specifically includes the following steps:

[0086] S222a, respectively assigning each current particle position to the parameters of the support vector machine model, inputting the current training data set into the assigned support vector machine model, and obtaining N samples to predict drilling accidents.

[0087] Wherein, N is a natural number greater than 2.

[0088] Specifically, each current particle position Substitute the parameters of the support vector machine model, input the current training data set into the assigned support vector machine model, and obtain Next, the support vector machine model predicts drilling accidents for the N samples predicted by the current training set.

[0089] S222b, determining the number of samples that are correctly predicted among the N samples predicting drilling accidents, and obtaining the fitness of each current particle position according to the number of samples that are correctly predicted and the total number of samples in the current training data set.

[0090] Specifically, determine the number of samples n that are correctly predicted among the N samples predicting drilling accidents. p , substitute the number of correctly predicted samples and the total number of samples in the current training data set into formula (1) to obtain the fitness of each current particle position:

[0091]

[0092] Among them, β i is the current particle position The fitness of .

[0093] For example, there are 100 groups of accident type samples in the current training data set. If the support vector machine model correctly determines that the number of drilling accidents corresponding to the data samples is 95 groups, then the current particle position Fitness

[0094] S223, comparing the fitness of the current particle position with the fitness of the historical local optimal position and the fitness of the historical global optimal position, respectively, to obtain the current local optimal position and the current global optimal position.

[0095] Specifically, in the first iteration, the historical local optimal position p of each particle is ibest The current particle position obtained by initialization The historical global optimal position G of all particles best is the position of the particle with the best fitness among the m particles.

[0096] like Figure 5 As shown, Figure 5 It is a flowchart of the steps of updating the local optimal position and the global optimal position, and step S223 specifically includes the following steps:

[0097] S223a, respectively comparing the fitness of each current particle position with the fitness of each historical local optimal position corresponding to the current particle position; if the fitness of the current particle position is better than the fitness of the historical local optimal position corresponding to the current particle position, assigning the current particle position to the historical local optimal position corresponding to the current particle position.

[0098] Specifically, each current particle position The fitness value β i With its historical local optimal position p ibest =(p ic ,p iε ,p iσ )’s fitness value β ibest For comparison, if β i Better than β ibest , for example, i Greater than β ibest ,but

[0099] S223b, comparing the fitness of all current particle positions with the fitness of the historical global optimal position; if there is any current particle position whose fitness is better than the fitness of the historical global optimal position, assigning the current particle position corresponding to the fitness better than the fitness of the historical global optimal position to the current global optimal position.

[0100] Specifically, each current particle position is The fitness of i and the global optimal position g best =(g c ,g ε ,g σ )’s fitness β best For comparison, if β i Better than β best ,but It is worth noting that in the comparison, g best is obtained after the last comparison update. For example, assuming that the last update was During this comparison, the current particle position The fitness of i With the global optimal position The fitness of best When making comparisons.

[0101] S224, determining whether the current global optimal position meets a preset condition, and if not, updating the current particle position and the current particle velocity according to the current local optimal position, the current global optimal position and the current particle velocity.

[0102] The preset condition may be that a preset number of iterations is met, or that the fitness of the current global optimal position is greater than a preset value.

[0103] Specifically, if the current global optimal position does not meet the preset conditions, the current particle position is updated according to the current local optimal position, the current global optimal position and the current particle velocity, and the current particle velocity is updated. The updating method is the same as that in the prior art and will not be described in detail here. If the current global optimal position meets the preset conditions, the current global optimal position is output as the updated parameter of the support vector machine model.

[0104] S225, assigning the fitness of the current local optimal position to the fitness of the historical local optimal position, assigning the fitness of the current global optimal position to the fitness of the historical global optimal position, returning to execute the fitness calculation of each current particle position according to the current particle position and the current training data set until the preset condition is met, and the current global optimal position that meets the preset condition is the updated parameter of the support vector machine model.

[0105] Specifically, Figure 6 As shown, Figure 6 It is an iterative flow chart of a particle swarm algorithm. This step is an iterative process. The updated particle position is assigned to the current particle position, the fitness of the current local optimal position is assigned to the fitness of the historical local optimal position, and the fitness of the current global optimal position is assigned to the fitness of the historical global optimal position. Return to execute steps S222-S223 until the current global optimal position that meets the preset conditions is obtained, and the iteration ends.

[0106] like Figure 7 As shown, Figure 7 The method further comprises S400, updating the parameters of the support vector machine model according to the final predicted drilling accident. The specific process is as follows:

[0107] S410, determining whether the final predicted drilling accident is correct. If the final predicted drilling accident is incorrect, correcting the final predicted drilling accident, acquiring an original data set, and adding the corrected drilling accident and the monitoring data to the original data set to obtain a current training data set.

[0108] Specifically, a method for determining whether the final predicted drilling accident is correct may be to receive feedback instructions from the client on the final predicted drilling accident; if the feedback instructions determine that the final predicted drilling accident is incorrect, the incorrect drilling accident is modified to a drilling accident corresponding to the monitoring data, and the corrected drilling accident and monitoring data are added to the original data set to obtain the current training data set.

[0109] S420: Update the parameters of the support vector machine model using the current training data set and the particle swarm algorithm.

[0110] Specifically, this step is the same as the aforementioned steps S221-S225 and will not be repeated here.

[0111] It is understandable that the ocean drilling accident prediction method provided in the embodiment of the present invention can be applied to any computer device with data processing and computing capabilities, and this computer device can be various terminals or servers. When the computer device in the embodiment is a server, the server is an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Optionally, the terminal is a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited to this.

[0112] The implementation of the present invention includes the following beneficial effects: the present invention can improve the safety of drilling operations by monitoring all aspects of data to achieve all-round monitoring, by first identifying the acquired data to identify abnormal data, and then using the abnormal data to classify drilling accidents, and using a support vector machine model to gradually refine the classification of drilling accidents until a predicted drilling accident is obtained, thereby improving the processing capability of monitoring data and thereby improving the information perception capability and safety of ocean drilling operations.

[0113] The present invention can also optimize the support vector machine model. When an error occurs in the predicted drilling accident, the erroneous drilling accident and its corresponding monitoring data are added to the training data set, and the parameters of the support vector machine model are re-optimized to improve the accuracy of subsequent predictions.

[0114] The present invention optimizes the parameters of the support vector machine model by using the particle swarm algorithm, and takes the probability of the model's correct prediction on the data set as the fitness of the particle swarm algorithm, so as to obtain the optimal parameters of the support vector machine model and improve the accuracy of the model prediction.

[0115] like Figure 8 As shown, Figure 8 The present invention is a structural diagram of a prediction system for ocean drilling accidents. The present invention also provides a prediction system for ocean drilling accidents, including a data layer, a decision layer and a user layer, wherein:

[0116] The data layer is used to collect monitoring data of ocean drilling and upload the monitoring data to the decision layer;

[0117] The decision layer is used to implement the above method and push the obtained data to the user layer;

[0118] The user layer is used to display the monitoring data and / or the data pushed by the decision layer.

[0119] Specifically, the data layer may include but is not limited to a platform information system, a drilling monitoring system, a meteorological information system, a dynamic positioning data system, a riser system, a lifting system and an underwater equipment system, etc., which are used to obtain monitoring data and meteorological data.

[0120] The decision layer may include but is not limited to a watertight pipe model, a dynamic positioning model, a drilling and completion knowledge base, a drilling fluid expert system, a remote expert support system, and a reasoning mechanism and algorithm. The reasoning mechanism and algorithm are configured with the aforementioned support vector machine model and anomaly recognition model to implement the aforementioned prediction method, input the monitoring data into the preset anomaly recognition model, and identify the monitoring data that deviates from the normal data range in the preset anomaly recognition model as abnormal data; the preset anomaly recognition model includes a drilling model, a ship attitude model, a remote-controlled unmanned submersible model, an underwater blowout preventer model, and a dynamic positioning model. The watertight pipe model is used to identify abnormal data of the watertight pipe, the drilling and completion knowledge base is used to store monitoring data, the drilling fluid expert system is used to provide expert support, and the remote expert system is used to provide remote support, including but not limited to remote identification of abnormal data and remote identification of drilling accidents.

[0121] The user layer may include but is not limited to the central control room integrated large screen, captain's client, drilling supervision client, positioning operation client, mud worker client and shore-based command hall client, etc.

[0122] After the decision layer completes the prediction, it can push the monitoring data, abnormal data or predicted drilling accidents to one or more clients in the user layer. The client can be, but is not limited to, a mobile phone, a computer or a tablet. Specifically, the reasoning mechanism and algorithm system are used to assign the preset accident to the current accident to be classified, and use the trained support vector machine model and the abnormal data to classify the current accident to be classified into two categories to obtain the current predicted accident; determine whether the current predicted accident meets the conditions, if the current predicted accident does not meet the conditions, assign the current predicted accident to the current accident to be classified, return to use the trained support vector machine model and the abnormal data to classify the current accident to be classified into two categories to obtain the current predicted accident, until the current predicted accident that meets the conditions is obtained, and the current predicted accident that meets the conditions is output as the final predicted drilling accident.

[0123] The reasoning mechanism and algorithm system are also specifically used to obtain the original data set, assign the original data set to the current training data set; use the current training data set and the particle swarm algorithm to update the parameters of the support vector machine model to obtain a trained support vector machine model.

[0124] The specific process is the same as that of the aforementioned method embodiment and will not be repeated here.

[0125] The reasoning mechanism and algorithm system are also specifically used to update the parameters of the support vector machine model according to the final predicted drilling accident, determine whether the final predicted drilling accident is correct, and if the final predicted drilling accident is incorrect, correct the final predicted drilling accident, and obtain the original data set, add the corrected drilling accident and monitoring data to the original data set to obtain the current training data set, and use the current training data set and the particle swarm algorithm to update the parameters of the support vector machine model.

[0126] The specific process is the same as that of the aforementioned method embodiment and will not be repeated here.

[0127] The reasoning mechanism and algorithm system are also specifically used to randomly initialize the current particle position and current particle velocity of the particle swarm algorithm; the particle position represents the solution of the parameters of the support vector machine model; the fitness of each current particle position is calculated according to the current particle position and the current training data set; the fitness of the current particle position is compared with the fitness of the historical local optimal position and the fitness of the historical global optimal position respectively to obtain the current local optimal position and the current global optimal position; determine whether the current global optimal position meets the preset conditions, if not, update the current particle position according to the current local optimal position, the current global optimal position and the current particle velocity, and update the current particle velocity; assign the updated particle position to the current particle position, assign the fitness of the current local optimal position to the fitness of the historical local optimal position, assign the fitness of the current global optimal position to the fitness of the historical global optimal position, return to execute the fitness of each current particle position calculated according to the current particle position and the current training data set until the preset conditions are met, and output the current global optimal position that meets the preset conditions as the updated parameters of the support vector machine model.

[0128] The specific process is the same as that of the aforementioned method embodiment and will not be repeated here.

[0129] The reasoning mechanism and algorithm system are also specifically used to assign each current particle position to the parameters of the support vector machine model respectively, input the current training data set into the assigned support vector machine model, and obtain N samples to predict drilling accidents; N is a natural number greater than 2; determine the number of correctly predicted samples in the N samples to predict drilling accidents, and obtain the fitness of each current particle position based on the number of correctly predicted samples and the total number of samples in the current training data set.

[0130] The specific process is the same as that of the aforementioned method embodiment and will not be repeated here.

[0131] The reasoning mechanism and algorithm system are also specifically used to compare the fitness of each current particle position with the fitness of the historical local optimal position corresponding to each current particle position. If the fitness of the current particle position is better than the fitness of the historical local optimal position corresponding to the current particle position, the current particle position is assigned to the historical local optimal position corresponding to the current particle position; the fitness of all current particle positions is compared with the fitness of the historical global optimal position. If there is any current particle position whose fitness is better than the fitness of the historical global optimal position, the current particle position corresponding to the fitness better than the fitness of the historical global optimal position is assigned to the current global optimal position.

[0132] The specific process is the same as that of the aforementioned method embodiment and will not be repeated here.

[0133] It can be seen that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0134] like Fig. 9 As shown, Fig. 9 The present invention is a schematic diagram of a method and device for predicting an ocean drilling accident. The present invention also provides a method and device for predicting an ocean drilling accident, comprising:

[0135] at least one processor;

[0136] at least one memory for storing at least one program;

[0137] When the at least one program is executed by the at least one processor, the at least one processor implements the prediction method steps described in the above method embodiment.

[0138] Among them, the memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. The memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a remote memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0139] It can be seen that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0140] In addition, the embodiment of the present application also discloses a computer program product or a computer program, and the computer program product or the computer program is stored in a computer-readable storage medium. The processor of the computer device can read the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device performs the above method. Similarly, the contents in the above method embodiment are all applicable to the storage medium embodiment, and the functions specifically implemented by the storage medium embodiment are the same as those in the above method embodiment, and the beneficial effects achieved are also the same as those achieved by the above method embodiment.

[0141] The present invention also provides a computer-readable storage medium, which stores a program executable by a processor. The program executable by the processor is used to implement the above method when executed by the processor.

[0142] It is understood that all or some steps and systems in the disclosed method above can be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, and the computer-readable medium can include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, disk storage or other magnetic storage device, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0143] like Fig.10 As shown, Fig.10 It is a structural diagram of another prediction system for ocean drilling accidents, including a data acquisition device and a computer device connected to the data acquisition device; wherein,

[0144] The data acquisition device is used to collect monitoring data of ocean drilling and upload the monitoring data to the computer device;

[0145] The computer device comprises:

[0146] at least one processor;

[0147] at least one memory for storing at least one program;

[0148] When the at least one program is executed by the at least one processor, the at least one processor implements the above-mentioned method.

[0149] Specifically, for data acquisition equipment, it is mainly implemented through sensors, and it may specifically include at least one sensor; and for computer equipment, it may be different types of electronic devices, including but not limited to desktop computers, laptops and other terminals.

[0150] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0151] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application. Those skilled in the art explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

Claims

1. A method for predicting ocean drilling accidents, characterized in that: The following steps are involved: Acquiring ocean drilling monitoring data and identifying abnormal data in the monitoring data; the monitoring data includes drilling data, ship attitude data, remote controlled unmanned submersible data, underwater blowout preventer data or dynamic positioning data; Assigning a preset accident value to the current accident to be classified, and using the trained support vector machine model and the abnormal data to perform binary classification on the current accident to be classified to obtain the current predicted accident; Determine whether the current predicted accident meets the conditions. If the current predicted accident does not meet the conditions, assign the current predicted accident to the current accident to be classified, return to use the trained support vector machine model and the abnormal data to perform binary classification on the current accident to be classified, and obtain the current predicted accident until the current predicted accident that meets the conditions is obtained, and output the current predicted accident that meets the conditions as the final predicted drilling accident.

2. The method according to claim 1, characterized in that The identifying abnormal data in the monitoring data specifically includes: The monitoring data is input into a preset abnormality recognition model, and the monitoring data that deviates from the normal data range in the preset abnormality recognition model is identified as abnormal data; the preset abnormality recognition model includes a drilling model, a ship attitude model, a remote-controlled unmanned submersible model, underwater blowout preventer data and a dynamic positioning model.

3. The method according to claim 1, characterized in that The trained support vector machine model is obtained by the following method: Get the original data set and assign it to the current training data set; The parameters of the support vector machine model are updated using the current training data set and the particle swarm algorithm to obtain the trained support vector machine model.

4. The method according to claim 1, characterized in that: The method further includes updating the parameters of the support vector machine model according to the final predicted drilling accident, wherein updating the parameters of the support vector machine model according to the final predicted drilling accident specifically includes: Determine whether the final predicted drilling accident is correct. If the final predicted drilling accident is incorrect, correct the final predicted drilling accident, obtain an original data set, and add the corrected drilling accident and the monitoring data to the original data set to obtain a current training data set; The parameters of the support vector machine model are updated using the current training data set and the particle swarm algorithm.

5. The method according to any one of claims 3-4, characterized in that: The updating of the parameters of the support vector machine model using the current training data set and the particle swarm algorithm specifically includes: Randomly initializing the current particle position and current particle velocity of the particle swarm algorithm; the particle position represents a solution of the parameters of the support vector machine model; Calculating the fitness of each current particle position according to the current particle position and the current training data set; Comparing the fitness of the current particle position with the fitness of the historical local optimal position and the fitness of the historical global optimal position, respectively, to obtain the current local optimal position and the current global optimal position; Determine whether the current global optimal position meets a preset condition, and if not, update the current particle position according to the current local optimal position, the current global optimal position and the current particle velocity, and update the current particle velocity; The fitness of the current local optimal position is assigned to the fitness of the historical local optimal position, and the fitness of the current global optimal position is assigned to the fitness of the historical global optimal position, and the process returns to execute the fitness calculation of each current particle position according to the current particle position and the current training data set until the preset condition is met, and the current global optimal position that meets the preset condition is the parameter of the support vector machine model.

6. The method according to claim 5, characterized in that The step of calculating the fitness of each current particle position according to the current particle position and the current training data set specifically includes: Assigning each current particle position to the parameters of the support vector machine model respectively, inputting the current training data set into the assigned support vector machine model, and obtaining N samples to predict drilling accidents; wherein N is a natural number greater than 2; The number of samples correctly predicted among the N samples for predicting drilling accidents is determined, and the fitness of each current particle position is obtained according to the number of samples correctly predicted and the total number of samples in the current training data set.

7. The method according to claim 5, characterized in that The step of comparing the fitness of the current particle position with the fitness of the historical local optimal position and the fitness of the historical global optimal position to obtain the current local optimal position and the current global optimal position specifically includes: Compare the fitness of each current particle position with the fitness of each historical local optimal position corresponding to the current particle position; if the fitness of the current particle position is better than the fitness of the historical local optimal position corresponding to the current particle position, assign the current particle position to the historical local optimal position corresponding to the current particle position; The fitness of all current particle positions is compared with the fitness of the historical global optimal position. If there is any current particle position whose fitness is better than the fitness of the historical global optimal position, the current particle position corresponding to the fitness better than the fitness of the historical global optimal position is assigned to the current global optimal position.

8. A method and device for predicting ocean drilling accidents, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 7.

9. A computer-readable storage medium storing a program executable by a processor, characterized in that: The processor-executable program is used to perform the method according to any one of claims 1 to 7 when executed by the processor.

10. A prediction system for ocean drilling accidents, characterized in that: It includes data layer, decision layer and user layer, among which: The data layer is used to collect monitoring data of ocean drilling and upload the monitoring data to the decision layer; The decision layer is used to implement the method described in any one of claims 1 to 7, and push the obtained data to the user layer; The user layer is used to display the monitoring data and / or the data pushed by the decision layer.

Images