A prediction method and device for logging obstruction and jam
Through the feedforward deep machine learning model, the logging encounters obstacles and blocks are predicted. The data such as the number of well logging, PCL logging, well diameter, well inclination and mud density are solved by using data such as the number of well logging, which is impossible to accurately predict well logging encounters in the existing technology. The problem of blockage reduces the construction risk and provides technical support for directional well drilling.
Patent Information
- Application Number
- CN202010684448.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-07-16
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2040-07-16
AI Technical Summary
The existing technology cannot accurately predict logging encounters obstacles and blockages, resulting in high construction risks and lack of systematic methods.
The feedforward deep machine learning model is used to train the machine learning model based on historical data, and the number of SLAM cables is used to predict obstacles and blocks when the number of data of SLAM cables are dropped, PCL, well diameter, well inclination and mud density.
Accurate prediction of obstacles and stuck problems during well logging construction is achieved, construction risks are reduced, and technical support is provided for directional well drilling.
Smart Images

Figure CN114016998B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of oil and gas exploration, and in particular to a method and a device for predicting obstructions and jams encountered in well logging. Background Art
[0002] Well logging technology is a related technology for obtaining various petroleum geological and engineering technical data after oil drilling to the designed well depth. It is a necessary means of original data for well completion and oil field development. According to geological and geophysical conditions, the rational selection of comprehensive logging methods can be used to study the geological profile of the borehole in detail, detect useful minerals, and provide detailed data necessary for calculating reserves, such as the effective thickness of the oil layer, porosity, oil and gas saturation and permeability, as well as study the drilling technology situation. At present, there are many reasons for logging obstructions and jams, and they cannot be accurately predicted. They often rely on manual experience and judgment, and there is no systematic method. How to accurately predict logging obstructions and jams in a specific area is of great significance for the smooth progress of logging in the area and reducing construction risks. Summary of the invention
[0003] In view of the problems in the prior art, the present invention provides a method and device for predicting obstruction and sticking in well logging, which can accurately predict obstruction and sticking problems and provide a reference for directional well drilling operations. In order to solve the above technical problems, the present invention provides the following technical solutions:
[0004] In a first aspect, the present invention provides a method for predicting well logging obstructions and jams, comprising:
[0005] Obtain the logging data of the target block to be predicted to encounter obstructions and jams, the logging data including: the number of SLAM cable runs, the number of PCL runs, the well diameter, the well deviation and the mud density;
[0006] The logging data and a pre-established machine learning model are used to predict logging obstructions and jams in the target block.
[0007] In one embodiment, the machine learning model is a feedforward deep machine learning model, and the steps of establishing the machine learning model include:
[0008] Collecting historical data of well logging data of the target block;
[0009] Generate a label for the feedforward deep machine learning model based on the historical obstruction and card encounter data and its corresponding historical data;
[0010] The initial model of the feedforward deep machine learning model is trained based on historical data and labels whose signal-to-noise ratio is greater than a preset value to generate the machine learning model.
[0011] In one embodiment, the initial model of the feedforward deep machine learning model is trained according to the historical data and labels with a signal-to-noise ratio greater than a preset value to generate the machine learning model, including:
[0012] According to the historical obstruction and card encounter data, the historical data corresponding to the historical obstruction and card encounter data, the visible layer bias and the hidden layer bias in the feedforward deep machine learning model, the connection weights of the visible layer bias and the connection weights of the hidden layer bias are generated to train the initial model.
[0013] In one embodiment, generating the connection weights of the visible layer bias and the connection weights of the hidden layer bias according to the historical obstruction and card encounter data, the historical data corresponding to the historical obstruction and card encounter data, the visible layer bias and the hidden layer bias in the feedforward deep machine learning model to train the initial model includes:
[0014] The initial value of the visible layer bias, the initial value of the hidden layer bias and the initial value of the connection weight are respectively set to three random numbers less than a preset value;
[0015] inputting the historical data into the initial model;
[0016] Perform the following iterative operation until the prediction error of the initial model is less than the preset error, thereby obtaining the deep belief network model;
[0017] The initial model is trained layer by layer from bottom to top to the n-1 layer, and the training result is input to the nth layer of the initial model to obtain the training result of the nth layer; wherein n is the total number of training layers of the feedforward deep machine learning model;
[0018] Calculate the prediction error according to the training result of the nth layer and the historical data;
[0019] When the prediction error is greater than a preset error, the initial to first layers are optimized layer by layer from top to bottom according to the prediction error.
[0020] In one embodiment, before establishing the machine learning model, the method further includes:
[0021] Filtering the historical data;
[0022] Normalizing the historical data;
[0023] Outliers in the historical data are removed.
[0024] In one embodiment, the historical obstruction and stuck data include: standpipe pressure, hook load, wellhead pressure, outlet flow, pump stroke, inlet flow, drilling time, bottom hole annular pressure, total hydrocarbon, C1 data and operating condition data;
[0025] The operating condition data include: overflow, leakage and normal.
[0026] In a second aspect, the present invention provides a prediction device for logging obstructions and jams, comprising:
[0027] A logging data acquisition unit is used to acquire the logging data of the target block to be predicted to encounter obstructions and jams, wherein the logging data includes: the number of SLAM cable runs, the number of PCL runs, the well diameter, the well deviation and the mud density;
[0028] The obstruction and jam prediction unit is used to predict the logging obstruction and jam of the target block by using the logging data and a pre-established machine learning model.
[0029] In one embodiment, the machine learning model is a feedforward deep machine learning model, and the prediction device for logging obstructions and jams further includes: a model building unit, which is used to build the machine learning model, and the model building unit includes:
[0030] A historical data collection module, used to collect historical data of well logging data of the target block;
[0031] A label generation module, used to generate labels for the feedforward deep machine learning model based on historical obstruction and card encounter data and its corresponding historical data;
[0032] The model training module is used to train the initial model of the feedforward deep machine learning model based on historical data and labels with a signal-to-noise ratio greater than a preset value to generate the machine learning model.
[0033] In one embodiment, the model training module includes:
[0034] The initial model training module is used to generate the connection weights of the visible layer bias and the connection weights of the hidden layer bias based on the historical obstruction and card encounter data, the historical data corresponding to the historical obstruction and card encounter data, the visible layer bias and the hidden layer bias in the feedforward deep machine learning model, so as to train the initial model.
[0035] In one embodiment, the initial model training module includes:
[0036] A random number setting module, used to set the initial value of the visible layer bias, the initial value of the hidden layer bias and the initial value of the connection weight to three random numbers less than a preset value;
[0037] A historical data input module, used for inputting the historical data into the initial model;
[0038] An iterative operation module, used to perform the following iterative operations until the prediction error of the initial model is less than a preset error, thereby obtaining the deep belief network model;
[0039] The initial model is trained layer by layer from bottom to top to the n-1 layer, and the training result is input to the nth layer of the initial model to obtain the training result of the nth layer; wherein n is the total number of training layers of the feedforward deep machine learning model;
[0040] Calculate the prediction error according to the training result of the nth layer and the historical data;
[0041] When the prediction error is greater than a preset error, the initial to first layers are optimized layer by layer from top to bottom according to the prediction error.
[0042] In one embodiment, the prediction device for logging obstructions and jams further includes:
[0043] A data filtering unit, used for filtering the historical data;
[0044] A data normalization unit, used for normalizing the historical data;
[0045] The outlier removal unit is used to remove outliers in the historical data.
[0046] In one embodiment, the historical obstruction and stuck data include: standpipe pressure, hook load, wellhead pressure, outlet flow, pump stroke, inlet flow, drilling time, bottom hole annular pressure, total hydrocarbon, C1 data and operating condition data;
[0047] The operating condition data include: overflow, leakage and normal.
[0048] In a third aspect, the present invention provides an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of a method for predicting obstructions and jams encountered in well logging are implemented.
[0049] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for predicting obstructions and jams encountered in well logging.
[0050] From the above description, it can be seen that the prediction method and device for logging obstruction and jamming provided by the embodiment of the present invention first obtains the logging data of the target block to be predicted to encounter obstruction and jamming, and the logging data includes: the number of SLAM cable downholes, the number of PCL downholes, the well diameter, the well deviation and the mud density; and uses the logging data and the pre-established machine learning model to predict the logging obstruction and jamming of the target block. The present invention can accurately predict the obstruction and jamming problems in the logging construction process, reduce complex risks, and provide technical support for directional well drilling construction operations. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0052] Figure 1 This is a first structural schematic diagram of a prediction system for logging obstructions and jams according to an embodiment of the present application;
[0053] Figure 2 This is a second structural schematic diagram of a prediction system for logging obstructions and jams according to an embodiment of the present application;
[0054] Figure 3 It is a schematic diagram of the first flow chart of the method for predicting well logging obstructions and jams in an embodiment of the present invention;
[0055] Figure 4 is a flow chart of step 300 in an embodiment of the present invention;
[0056] Figure 5 This is a flow chart of step 303 in an embodiment of the present invention;
[0057] Figure 6 This is a flow chart of step 3031 in an embodiment of the present invention;
[0058] Figure 7 It is a second flow chart of the method for predicting well logging obstruction and jamming in an embodiment of the present invention;
[0059] Figure 8 It is a flow chart of a method for predicting obstructions and jams in well logging in a specific application example of the present invention;
[0060] Fig. 9 It is a first structural block diagram of a device for predicting obstructions and jams in well logging in an embodiment of the present invention;
[0061] Fig.10It is a second structural block diagram of the device for predicting obstructions and jams in well logging in an embodiment of the present invention;
[0062] Fig.11 A schematic diagram of the structure of a model building unit in an embodiment of the present invention;
[0063] Fig.12 It is a structural diagram of a model training module in an embodiment of the present invention;
[0064] Fig.13 It is a structural schematic diagram of an initial model training module in an embodiment of the present invention;
[0065] Fig.14 It is a third structural block diagram of the prediction device for well logging encountering obstruction and jamming in the embodiment of the present invention;
[0066] Fig.15 It is a schematic diagram of the structure of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION
[0067] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0068] The present application provides a prediction system for logging obstructions and jams, the system comprising a prediction device for logging obstructions and jams, see Figure 1 The device may be a server A1, which may be connected to multiple logging data receiving terminals B1 in communication, and the server A1 may be connected to multiple databases in communication, or Figure 2 As shown, these databases can also be directly set in the server A1. Among them, the drilling data receiving terminal B1 is used to receive data such as standpipe pressure, hook load, wellhead pressure, outlet flow, pump stroke, inlet flow, drilling time, bottom hole annulus pressure, total hydrocarbon, C1 data and working conditions during the drilling process. After receiving the logging data, the server A1 performs real-time prediction on the logging data and displays the prediction results to the user through the client C1.
[0069] It is understandable that the logging data receiving end B1 may be a sensor, such as a pressure sensor, a flow sensor, a displacement sensor, a gas sensor, etc. The client C1 may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, smart watches, smart bracelets, etc.
[0070] In practical applications, the prediction of logging obstructions and jams can be performed on the server A1 side as described above, that is, Figure 1 or Figure 2 The architecture shown can also be completed in the client C1 device. The specific selection can be based on the processing capability of the client device and the limitations of the user's usage scenario. This application does not limit this. If all operations are completed in the client device, the client device may also include a processor for processing the drilling overflow and leakage condition prediction results and other operations.
[0071] The client C1 device mentioned above may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the prediction side of the logging blockage card, and other implementation scenarios may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the prediction server of the logging blockage card. The server may include a single computer device, or a server cluster consisting of multiple servers, or a server structure of a distributed device.
[0072] The server and the client device may communicate using any suitable network protocol, including network protocols that have not yet been developed on the date of filing this application. Network protocols may include, for example, TCP / IP, UDP / IP, HTTP, HTTPS, etc. Of course, network protocols may also include, for example, RPC (Remote Procedure Call Protocol) and REST (Representational State Transfer) protocols used on top of the above protocols.
[0073] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0074] The embodiment of the present invention provides a specific implementation method of a method for predicting well logging encountering obstructions and jams, see Figure 3 , the method specifically includes the following contents:
[0075] Step 100: Obtain logging data of the target block to be predicted to encounter obstructions or stuck.
[0076] It is understandable that the logging data in step 100 include: the number of SLAM cable runs, the number of PCL runs, the well diameter, the well deviation and the mud density. Specifically, the actual well logging operation statistics of the target block are collected to prepare for the main control factors of logging obstructions and jams. It mainly includes the depth of obstructions and jams, the mud density at the obstruction and jam location, the well deviation at the obstruction and jam location, the number of SLAM cable runs, the number of PCL runs, etc.
[0077] Step 200: Use the logging data and a pre-established machine learning model to predict logging obstructions and jams in the target block.
[0078] It can be understood that a machine learning model refers to the use of algorithms to parse data, learn from it, and then make decisions or predictions about something. It can improve the performance of specific algorithms through empirical learning. It is a computer algorithm that can be automatically improved through experience. Generally, data or past experience is used as the performance standard for optimizing the algorithm.
[0079] The obstruction and jamming during logging, which leads to well handling, has always been one of the important factors affecting the timeliness of logging. Due to the complex geological conditions of petroleum, small target area, large inclination, and large displacement directional well technology are often used for exploration and development. At the same time, in recent years, in order to improve exploration efficiency and shorten the engineering cycle, the drilling progress has been accelerated. If the wellbore and mud are not properly handled, it will also cause complex wellbore environments such as unstable wellbore wall and poor wellbore trajectory, which greatly increases the probability of obstruction and jamming during logging construction. Through the convenience of machine learning and based on a large amount of historical data, it is possible to more accurately predict the obstruction and jamming of logging in the target work area.
[0080] From the above description, it can be seen that the prediction method for logging obstruction and stuck provided by the embodiment of the present invention first obtains the logging data of the target block to be predicted to encounter obstruction and stuck, and the logging data includes: the number of SLAM cable downholes, the number of PCL downholes, the well diameter, the well deviation and the mud density; and uses the logging data and the pre-established machine learning model to predict the logging obstruction and stuck of the target block. The present invention can accurately predict the obstruction and stuck problems in the logging construction process, reduce complex risks, and provide technical support for directional well drilling construction operations.
[0081] In one embodiment, the machine learning model is a feedforward deep machine learning model, and the method for predicting logging obstructions and jams further includes: Step 300: establishing the machine learning model. Figure 4 , step 300 further comprises:
[0082] Step 301: Collect historical data of well logging data of the target block;
[0083] Specifically, the number of SLAM cable downhole trips, PCL downhole trips, well diameters, well inclinations, and mud density corresponding to the obstructions and jams that occurred in the target block in the past are collected.
[0084] Step 302: Generate a label for the feedforward deep machine learning model based on the historical obstruction and card encounter data and its corresponding historical data;
[0085] Step 303: Train the initial model of the feedforward deep machine learning model based on historical data and labels with a signal-to-noise ratio greater than a preset value to generate the machine learning model.
[0086] In step 301 to step 303, logging data with a high signal-to-noise ratio are selected as sample data, and the number of SLAM cable downhole times, PCL downhole times, well diameter, well inclination and mud density are calibrated as model training constraints. A feedforward deep learning network is used to train samples and labels to obtain a configured machine learning model.
[0087] In one embodiment, see Figure 5 , step 303 further comprises:
[0088] Step 3031: Generate the connection weights of the visible layer bias and the connection weights of the hidden layer bias according to the historical obstruction and card encounter data, the historical data corresponding to the historical obstruction and card encounter data, the visible layer bias and the hidden layer bias in the feedforward deep machine learning model to train the initial model.
[0089] In one embodiment, see Figure 6 , step 3031 further includes:
[0090] Step 30311: respectively setting the initial value of the visible layer bias, the initial value of the hidden layer bias and the initial value of the connection weight to three random numbers less than a preset value;
[0091] Step 30312: inputting the historical data into the initial model;
[0092] Step 30313: perform the following iterative operation until the prediction error of the initial model is less than the preset error, thereby obtaining the deep belief network model;
[0093] The initial model is trained layer by layer from bottom to top to the n-1 layer, and the training result is input to the nth layer of the initial model to obtain the training result of the nth layer; wherein n is the total number of training layers of the feedforward deep machine learning model;
[0094] Calculate the prediction error according to the training result of the nth layer and the historical data;
[0095] When the prediction error is greater than a preset error, the initial to first layers are optimized layer by layer from top to bottom according to the prediction error.
[0096] In step 30311 to step 30313, the initial value of the connection weight, the initial value of the visible layer bias and the hidden layer bias can be set to the same random number, or can be set to three different random numbers. The preprocessed training data set is substituted into the initial model, and trained layer by layer from bottom to top to obtain the parameters of each layer such as the connection weight, visible layer bias, hidden layer bias, etc., until the n-1 layer is trained, the n-1 layer training result is input into the n-th layer (top layer) machine learning model, and the n-th layer training result is compared with the working condition data (actual data) in the training data set to obtain the corresponding error. Then, the error is transmitted layer by layer from top to bottom, and the connection weight, visible layer bias, and hidden layer bias in each layer are fine-tuned according to the error of each layer until the first layer, and then the training data is used to train upward layer by layer, and the cycle is repeated... until the error is less than the preset error, and a deep belief network model is obtained.
[0097] In one embodiment, see Figure 7 Before step 300, the method for predicting logging obstructions and jams also includes:
[0098] Step 400: filtering the historical data;
[0099] Step 500: normalizing the historical data;
[0100] Step 600: Remove outliers from the historical data.
[0101] It is understandable that historical data needs to be preprocessed, including filtering, normalization, etc., to remove unexpected values, and normalize the historical data into the format required by the feedforward deep machine learning model.
[0102] In one embodiment, the historical obstruction and stuck data include: standpipe pressure, hook load, wellhead pressure, outlet flow, pump stroke, inlet flow, drilling time, bottom hole annular pressure, total hydrocarbon, C1 data and operating condition data;
[0103] The operating condition data include: overflow, leakage and normal.
[0104] To further illustrate the present solution, the present invention takes the K block as an example to provide a specific application example of the prediction method for logging encountering obstructions and jams, which specifically includes the following contents, see Figure 8 .
[0105] The R layer in the K block is under-compacted, and improper drilling parameters will cause wellbore collapse and a large wellbore expansion rate (commonly known as a "big belly", a wellbore expansion rate greater than 0.15 is a "big belly"); the S and L layers are mainly composed of mud shale, which swells when exposed to water, causing necking and drill sticking; the SS layer is mainly composed of sandstone and mudstone interlayers, and improper drilling parameters will form a "sugar gourd" wellbore. In addition, it was found during logging construction that directional well logging encountered serious problems of obstruction and sticking (52%).
[0106] S1: Building a feed-forward deep machine learning model.
[0107] It is understandable that the goal of a feedforward network is to approximate a function. For example, for a classifier, y = f*(x) maps the input x to a category y. The feedforward network defines a mapping y = f(x; θ) and learns the value of the parameter θ so that it can obtain the best function approximation. Inside the feedforward neural network, the parameters propagate unidirectionally from the input layer to the output layer. Usually, a feedforward deep learning network contains hidden layer bias, visible layer bias, convolution layer, pooling layer, and fully connected layer.
[0108] S2: Real-time acquisition of well logging data of the K block to be predicted to encounter obstructions and jams.
[0109] The logging data in step S2 include: the number of SLAM cable runs, the number of PCL runs, the well diameter, the well deviation, and the mud density.
[0110] S3: Preprocess the logging data in step S2.
[0111] In specific implementation, the preprocessing in S3 includes the following: filtering, normalizing, removing outliers and other operations on the training data set and the test data set into the format required by the feedforward deep machine learning model, such as data in the range of [0,1].
[0112] Preferably, the normalization method selected in step S3 is: dividing the data by the largest value in the group of data.
[0113]
[0114] x i ′——a certain data after normalization, x i ——Some data before normalization, x j ——All data before normalization.
[0115] S4: Use logging data and feed-forward deep machine learning models to predict logging obstructions and jams in Block K.
[0116] Specifically, the logging data preprocessed by step S3 is input into the feedforward deep machine learning model generated by step S1 to predict the logging obstruction and jamming situation of block K.
[0117] From the above description, it can be seen that the prediction method for logging obstruction and stuck provided by the embodiment of the present invention first obtains the logging data of the target block to be predicted to encounter obstruction and stuck, and the logging data includes: the number of SLAM cable downholes, the number of PCL downholes, the well diameter, the well deviation and the mud density; and uses the logging data and the pre-established machine learning model to predict the logging obstruction and stuck of the target block. The present invention can accurately predict the obstruction and stuck problems in the logging construction process, reduce complex risks, and provide technical support for directional well drilling construction operations.
[0118] Based on the same inventive concept, the embodiments of the present application also provide a prediction device for logging obstructions and jams, which can be used to implement the methods described in the above embodiments, such as the following embodiments. Since the principle of solving the problem by the prediction device for logging obstructions and jams is similar to the prediction method for logging obstructions and jams, the implementation of the prediction device for logging obstructions and jams can refer to the implementation of the prediction method for logging obstructions and jams, and the repeated parts will not be repeated. As used below, the terms "unit" or "module" can be a combination of software and / or hardware that implements predetermined functions. Although the system described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.
[0119] The embodiment of the present invention provides a specific implementation of a prediction device for well logging encountering obstruction and jamming, which can realize the prediction method for well logging encountering obstruction and jamming, see Fig. 9 The prediction device for logging obstruction and jamming specifically includes the following contents:
[0120] The logging data acquisition unit 10 is used to acquire the logging data of the target block to be predicted to encounter obstructions and jams, and the logging data includes: the number of SLAM cable running down the well, the number of PCL running down the well, the well diameter, the well deviation and the mud density;
[0121] The obstruction and stuck prediction unit 20 is used to predict the logging obstruction and stuck of the target block by using the logging data and a pre-established machine learning model.
[0122] In one embodiment, see Fig.10 , the machine learning model is a feedforward deep machine learning model, the prediction device for logging encountering obstruction and jamming also includes: a model building unit 30, which is used to build the machine learning model, see Fig.11 , the model building unit 30 includes:
[0123] A historical data collection module 301 is used to collect historical data of well logging data of the target block;
[0124] A label generation module 302, used to generate labels for the feedforward deep machine learning model based on historical obstruction and card encounter data and its corresponding historical data;
[0125] The model training module 303 is used to train the initial model of the feedforward deep machine learning model based on historical data and labels with a signal-to-noise ratio greater than a preset value to generate the machine learning model.
[0126] In one embodiment, see Fig.12 , the model training module 303 includes:
[0127] The initial model training module 3031 is used to generate the connection weights of the visible layer bias and the connection weights of the hidden layer bias based on the historical obstacle and card encounter data, the historical data corresponding to the historical obstacle and card encounter data, the visible layer bias and the hidden layer bias in the feedforward deep machine learning model, so as to train the initial model.
[0128] In one embodiment, see Fig.13 , the initial model training module 3031 includes:
[0129] A random number setting module 30311 is used to set the initial value of the visible layer bias, the initial value of the hidden layer bias and the initial value of the connection weight to three random numbers less than a preset value;
[0130] A historical data input module 30312, used for inputting the historical data into the initial model;
[0131] Iterative operation module 30313, used to perform the following iterative operation until the prediction error of the initial model is less than the preset error, thereby obtaining the deep belief network model;
[0132] The initial model is trained layer by layer from bottom to top to the n-1 layer, and the training result is input to the nth layer of the initial model to obtain the training result of the nth layer; wherein n is the total number of training layers of the feedforward deep machine learning model;
[0133] Calculate the prediction error according to the training result of the nth layer and the historical data;
[0134] When the prediction error is greater than a preset error, the initial to first layers are optimized layer by layer from top to bottom according to the prediction error.
[0135] In one embodiment, see Fig.14 The prediction device for logging obstructions and jams also includes:
[0136] A data filtering unit 40, configured to filter the historical data;
[0137] A data normalization unit 50, configured to normalize the historical data;
[0138] The outlier removal unit 60 is used to remove outliers in the historical data.
[0139] In one embodiment, the historical obstruction and stuck data include: standpipe pressure, hook load, wellhead pressure, outlet flow, pump stroke, inlet flow, drilling time, bottom hole annular pressure, total hydrocarbon, C1 data and operating condition data;
[0140] The operating condition data include: overflow, leakage and normal.
[0141] From the above description, it can be seen that the prediction device for logging obstruction and jamming provided by the embodiment of the present invention first obtains the logging data of the target block to be predicted to encounter obstruction and jamming, and the logging data includes: the number of SLAM cable downhole times, the number of PCL downhole times, the well diameter, the well deviation and the mud density; and uses the logging data and the pre-established machine learning model to predict the logging obstruction and jamming of the target block. The present invention can accurately predict the obstruction and jamming problems in the logging construction process, reduce complex risks, and provide technical support for directional well drilling construction operations.
[0142] The devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is an electronic device, and the electronic device may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
[0143] In a typical example, the electronic device specifically includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the above-mentioned method for predicting well logging obstructions and jams are implemented, and the steps include:
[0144] Step 100: Acquire the logging data of the target block to be predicted to encounter obstructions and jams, wherein the logging data includes: the number of SLAM cable runs, the number of PCL runs, the well diameter, the well deviation and the mud density;
[0145] Step 200: Use the logging data and a pre-established machine learning model to predict logging obstructions and jams in the target block.
[0146] Reference below Fig.15 , which shows a schematic structural diagram of an electronic device 600 suitable for implementing an embodiment of the present application.
[0147] like Fig.15As shown, electronic device 600 includes a central processing unit (CPU) 601, which can perform various appropriate operations and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 into a random access memory (RAM) 603. In RAM 603, various programs and data required for the operation of system 600 are also stored. CPU 601, ROM 602, and RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to bus 604.
[0148] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed as needed as the storage section 608.
[0149] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned method for predicting well logging encounters obstructions and jams are implemented, and the steps include:
[0150] Step 100: Acquire the logging data of the target block to be predicted to encounter obstructions and jams, wherein the logging data includes: the number of SLAM cable runs, the number of PCL runs, the well diameter, the well deviation and the mud density;
[0151] Step 200: Use the logging data and a pre-established machine learning model to predict logging obstructions and jams in the target block.
[0152] In such an embodiment, the computer program may be downloaded and installed from a network via the communication section 609 , and / or installed from the removable medium 611 .
[0153] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0154] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.
[0155] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0156] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0157] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0158] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, commodity or device including the element.
[0159] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.
[0160] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.
[0161] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.
[0162] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims
1. A method for predicting obstructions and jams in well logging, characterized in that: include: Obtain the logging data of the target block to be predicted to encounter obstructions and jams, the logging data including: the number of SLAM cable runs, the number of PCL runs, the well diameter, the well deviation and the mud density; Using the logging data and a pre-established machine learning model to predict logging obstructions and jams in the target block; The machine learning model is a feedforward deep machine learning model, and the steps of establishing the machine learning model include: Collecting historical data of well logging data of the target block; Generate a label for the feedforward deep machine learning model based on the historical obstruction and card encounter data and its corresponding historical data; Training the initial model of the feedforward deep machine learning model according to historical data and labels whose signal-to-noise ratio is greater than a preset value to generate the machine learning model; The initial model of the feedforward deep machine learning model is trained according to the historical data and labels whose signal-to-noise ratio is greater than a preset value to generate the machine learning model, including: Generate the connection weight of the visible layer bias and the connection weight of the hidden layer bias according to the historical obstruction and card encounter data, the historical data corresponding to the historical obstruction and card encounter data, the visible layer bias and the hidden layer bias in the feedforward deep machine learning model, so as to train the initial model; The generating the connection weight of the visible layer bias and the connection weight of the hidden layer bias according to the historical obstruction and card encounter data, the historical data corresponding to the historical obstruction and card encounter data, the visible layer bias and the hidden layer bias in the feedforward deep machine learning model to train the initial model includes: The initial value of the visible layer bias, the initial value of the hidden layer bias and the initial value of the connection weight are respectively set to three random numbers less than a preset value; inputting the historical data into the initial model; Perform the following iterative operation until the prediction error of the initial model is less than the preset error, thereby obtaining the machine learning model; The initial model is trained layer by layer from bottom to top to the n-1 layer, and the training result is input to the nth layer of the initial model to obtain the training result of the nth layer; wherein n is the total number of training layers of the feedforward deep machine learning model; Calculate the prediction error according to the training result of the nth layer and the historical data; When the prediction error is greater than a preset error, the initial model is optimized layer by layer from top to bottom to the first layer according to the prediction error.
2. The method for predicting logging obstructions and jams according to claim 1, characterized in that: Before establishing the machine learning model, it also includes: Filtering the historical data; Normalizing the historical data; Outliers in the historical data are removed.
3. The method for predicting logging obstructions and jams according to claim 1, characterized in that: The historical obstruction and stuck data include: standpipe pressure, hook load, wellhead pressure, outlet flow, pump stroke, inlet flow, drilling time, bottom hole annular pressure, total hydrocarbon, C1 data and operating condition data; The operating condition data include: overflow, leakage and normal.
4. A prediction device for logging obstructions and jams, characterized in that: include: A logging data acquisition unit is used to acquire the logging data of the target block to be predicted to encounter obstructions and jams, wherein the logging data includes: the number of SLAM cable runs, the number of PCL runs, the well diameter, the well deviation and the mud density; An obstruction and jam prediction unit, used to predict the logging obstruction and jam of the target block by using the logging data and a pre-established machine learning model; The machine learning model is a feedforward deep machine learning model, and further includes: a model building unit, which is used to build the machine learning model, and the model building unit includes: A historical data collection module, used to collect historical data of well logging data of the target block; A label generation module, used to generate labels for the feedforward deep machine learning model based on historical obstruction and card encounter data and its corresponding historical data; A model training module, used to train the initial model of the feedforward deep machine learning model according to historical data and labels with a signal-to-noise ratio greater than a preset value, so as to generate the machine learning model; The model training module includes: An initial model training module, used to generate a connection weight of the visible layer bias and a connection weight of the hidden layer bias according to historical obstruction and card encounter data, historical data corresponding to the historical obstruction and card encounter data, and a visible layer bias and a hidden layer bias in the feedforward deep machine learning model, so as to train the initial model; The initial model training module includes: A random number setting module, used to set the initial value of the visible layer bias, the initial value of the hidden layer bias and the initial value of the connection weight to three random numbers less than a preset value; A historical data input module, used for inputting the historical data into the initial model; An iterative operation module, used to perform the following iterative operations until the prediction error of the initial model is less than a preset error, thereby obtaining the machine learning model; The initial model is trained layer by layer from bottom to top to the n-1 layer, and the training result is input to the nth layer of the initial model to obtain the training result of the nth layer; wherein n is the total number of training layers of the feedforward deep machine learning model; Calculate the prediction error according to the training result of the nth layer and the historical data; When the prediction error is greater than a preset error, the initial model is optimized layer by layer from top to bottom to the first layer according to the prediction error.
5. The prediction device for logging obstructions and jams according to claim 4, characterized in that: Also includes: A data filtering unit, used for filtering the historical data; A data normalization unit, used for normalizing the historical data; The outlier removal unit is used to remove outliers in the historical data.
6. The prediction device for logging obstructions and jams according to claim 4, characterized in that: The historical obstruction and stuck data include: standpipe pressure, hook load, wellhead pressure, outlet flow, pump stroke, inlet flow, drilling time, bottom hole annular pressure, total hydrocarbon, C1 data and operating condition data; The operating condition data include: overflow, leakage and normal.
7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method for predicting logging obstructions and jams as described in any one of claims 1 to 3 are implemented.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for predicting logging obstructions and jams as described in any one of claims 1 to 3 are implemented.
Citation Information
Patent Citations
Method and system for predicting drill string stuck pipe event
CN104024572A
Fault diagnosis method for rolling bearing based on deep learning and SVM (Support Vector Machine)
CN104616033A
Logging facies identification and analysis method based on fuzzy depth learning in big data environment
CN106529667A
A fault diagnosis method of high voltage circuit breaker based on depth belief network
CN109086817A
Automatic identification method and device of electrical-imaging logging facies
CN109212617A
