Deepwater drilling and complex drilling multiphase flow simulation system and method

Abstract

The invention provides a deepwater drilling complex drilling multiphase flow simulation system and method, and the system comprises a drilling multiphase flow simulation database which is used for storing and managing multi-source heterogeneous data in a drilling process; the data service architecture module is used for constructing a data service architecture of drilling multiphase flow analogue simulation; the data standardization and cleaning fusion module is used for processing multi-source heterogeneous data in the drilling process; and the analogue simulation and real-time optimization module is used for carrying out analogue simulation and real-time optimization on drilling multiphase flow. According to the technical scheme, real-time simulation and optimization of multiphase flow in deepwater drilling engineering are achieved.

Description

Technical Field

[0001] The present application relates to the field of drilling engineering technology, and in particular to a system and method for simulating complex multiphase flow in deepwater drilling. Background Art

[0002] In deepwater drilling projects, multiphase flow characteristics are complex, involving the interaction of multiple media such as drilling fluid, gas, and cuttings. Traditional simulation methods have difficulty accurately simulating transient behavior, resulting in large deviations in mechanical drilling rate predictions and difficulties in wellbore pressure control. In the existing technology, although some multiphase flow simulation tools exist (such as SLB's Olga simulator), they are not adaptable enough to the complex formation conditions in deepwater and lack deep integration with real-time data services. In addition, deepwater drilling data suffers from multi-source heterogeneity and noise interference, and there is an urgent need to establish a standardized data flow mechanism and cleaning and fusion algorithm. Summary of the Invention

[0003] The present application provides a deepwater drilling complex drilling multiphase flow simulation system and method for achieving real-time simulation and optimization of multiphase flow in deepwater drilling projects.

[0004] In a first aspect, a deepwater drilling complex drilling multiphase flow simulation system is provided, comprising:

[0005] Drilling multiphase flow simulation database, used to store and manage multi-source heterogeneous data during the drilling process;

[0006] Data service architecture module, used to build the data service architecture for drilling multiphase flow simulation;

[0007] Data standardization, cleaning and fusion module, used to process multi-source heterogeneous data during the drilling process;

[0008] The simulation and real-time optimization module is used to simulate and optimize drilling multiphase flow in real time.

[0009] In the above technical solution, a drilling multiphase flow simulation database is set up to store and manage multi-source heterogeneous data in the drilling process; a data service architecture module is used to build a data service architecture for drilling multiphase flow simulation; a data standardization and cleaning fusion module is used to process multi-source heterogeneous data in the drilling process; a simulation and real-time optimization module is used to simulate and optimize drilling multiphase flow in real time; thus, real-time simulation and optimization of multiphase flow in deepwater drilling projects are realized.

[0010] In a specific implementation scheme, the drilling multiphase flow simulation database adopts a distributed storage architecture.

[0011] In a specific implementation scheme, the data service architecture module adopts a three-tier B / S architecture.

[0012] In a specific implementation scheme, the data standardization and cleaning fusion module includes:

[0013] Missing value processing submodule, used to process missing values ​​of multi-source heterogeneous data during drilling;

[0014] Noise data identification submodule, used to identify noise from multi-source heterogeneous data during drilling;

[0015] The data fusion submodule is used to fuse multi-source heterogeneous data during the drilling process.

[0016] In a specific implementation scheme, the simulation and real-time optimization module adopts a dynamic simulation calculation model.

[0017] In a specific implementation scheme, the missing value processing submodule adopts a prediction filling algorithm based on machine learning.

[0018] In a specific implementation scheme, the noise data identification submodule uses a box plot.

[0019] In a specific implementation scheme, the data fusion submodule adopts an entity recognition algorithm.

[0020] In a specific implementation scheme, the three-tier B / S architecture includes a presentation layer, a business logic layer, and a data access layer.

[0021] In a second aspect, a method for simulating complex multiphase flow in deepwater drilling is provided, comprising the following steps:

[0022] Use the drilling multiphase flow simulation database to store and manage multi-source heterogeneous data during the drilling process;

[0023] Use the data service architecture module to build a data service architecture for drilling multiphase flow simulation;

[0024] Use data standardization and cleaning fusion modules to process multi-source heterogeneous data during drilling;

[0025] The simulation and real-time optimization modules are used to simulate and optimize the multiphase flow in drilling.

[0026] In the above technical solution, the multi-source heterogeneous data in the drilling process are stored and managed by utilizing the drilling multiphase flow simulation database; the data service architecture module is used to construct the data service architecture for drilling multiphase flow simulation; the data standardization and cleaning fusion module is used to process the multi-source heterogeneous data in the drilling process; the simulation and real-time optimization module is used to simulate and optimize the drilling multiphase flow in real time; and the real-time simulation and optimization of multiphase flow in deepwater drilling projects are realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a structural block diagram of a deepwater drilling complex drilling multiphase flow simulation system provided in an embodiment of the present application;

[0028] Figure 2 This is a flowchart of the deepwater drilling complex drilling multiphase flow simulation method provided in the embodiment of the present application. DETAILED DESCRIPTION

[0029] The present application will be further described in detail below with reference to the accompanying drawings and examples, through which the features and advantages of the present application will become more clearly understood.

[0030] The word "exemplary" is used exclusively herein to mean "serving as an example, example, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise indicated.

[0031] In addition, the technical features involved in different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0032] To facilitate understanding of the deepwater drilling complex drilling multiphase flow simulation system and method provided in the embodiment of the present application, its application scenario is first explained. The deepwater drilling complex drilling multiphase flow simulation system and method provided in the embodiment of the present application are used to realize real-time simulation and optimization of multiphase flow in deepwater drilling projects. In deepwater drilling projects, multiphase flow characteristics are complex, involving the interaction of multiple media such as drilling fluid, gas, and cuttings. Traditional simulation methods are difficult to accurately simulate transient behavior, resulting in large deviations in mechanical drilling speed prediction, difficulty in wellbore pressure control and other problems. In the prior art, although there are some multiphase flow simulation tools (such as SLB's Olga simulator), they are not adaptable enough under deepwater complex formation conditions and lack deep integration with real-time data services. In addition, deepwater drilling data has problems such as multi-source heterogeneity and noise interference, and it is urgent to establish a standardized data flow mechanism and cleaning fusion algorithm. For this reason, the embodiment of the present application provides a deepwater drilling complex drilling multiphase flow simulation system and method to realize real-time simulation and optimization of multiphase flow in deepwater drilling projects. The following describes it in detail with reference to specific drawings and embodiments.

[0033] refer to Figure 1 and Figure 2 , Figure 1 This is a structural block diagram of a deepwater drilling complex drilling multiphase flow simulation system provided in an embodiment of the present application; Figure 2 This is a flowchart of the deepwater drilling complex drilling multiphase flow simulation method provided in the embodiment of the present application.

[0034] exist Figure 1 In the embodiment of the present application, a deepwater drilling complex drilling multiphase flow simulation system is provided, comprising:

[0035] Drilling multiphase flow simulation database, used to store and manage multi-source heterogeneous data during the drilling process;

[0036] Data service architecture module, used to build the data service architecture for drilling multiphase flow simulation;

[0037] Data standardization, cleaning and fusion module, used to process multi-source heterogeneous data during the drilling process;

[0038] The simulation and real-time optimization module is used to simulate and optimize drilling multiphase flow in real time.

[0039] In the above technical solution, a drilling multiphase flow simulation database is set up to store and manage multi-source heterogeneous data in the drilling process; a data service architecture module is used to build a data service architecture for drilling multiphase flow simulation; a data standardization and cleaning fusion module is used to process multi-source heterogeneous data in the drilling process; a simulation and real-time optimization module is used to simulate and optimize drilling multiphase flow in real time; thus, real-time simulation and optimization of multiphase flow in deepwater drilling projects are realized.

[0040] Specifically, the beneficial effects include:

[0041] Improve drilling efficiency: Through real-time simulation and optimization of drilling multiphase flow, drilling parameters can be more accurately predicted and adjusted, reducing unnecessary drilling operations and waiting time, thereby significantly improving the overall efficiency of the drilling project.

[0042] Reduce drilling risks: The system can monitor and analyze the multiphase flow during the drilling process in real time, promptly identify potential safety hazards and abnormal conditions, and take corresponding optimization measures to intervene, effectively reducing the probability of drilling accidents and ensuring the safety of drilling operations.

[0043] Optimize resource utilization: Through comprehensive processing and analysis of multi-source heterogeneous data, the system can more reasonably allocate and utilize drilling resources, such as drilling fluid and drill bits, to avoid waste and overconsumption of resources and maximize resource utilization.

[0044] Improve scientific decision-making: The simulation and real-time optimization results provided by the system provide drilling engineers with a scientific and accurate basis for decision-making, helping them make more reasonable and effective drilling decisions and improve the success rate and economic benefits of drilling projects.

[0045] Promoting technological innovation: The research and development and application of this system have promoted the innovation and development of multiphase flow simulation technology in deepwater drilling projects, providing strong support for technological progress and industrial upgrading in the drilling industry.

[0046] Enhanced data management capabilities: The setting of the drilling multiphase flow simulation database and data service architecture module enhances the system's storage, management and service capabilities for multi-source heterogeneous data, providing a solid foundation for subsequent data analysis and mining.

[0047] Improve data quality: The data standardization and cleaning fusion module processes multi-source heterogeneous data, improving data quality and consistency, and providing more reliable data support for simulation and real-time optimization.

[0048] In a specific implementation scheme, the drilling multiphase flow simulation database adopts a distributed storage architecture.

[0049] Specifically, the drilling multiphase flow simulation database adopts a distributed storage architecture, which has the following benefits:

[0050] Improved data storage capacity and scalability: A distributed storage architecture allows the database to store data across multiple nodes, significantly increasing data storage capacity. When storage capacity needs to be increased, simply add new storage nodes, eliminating the need for large-scale modifications to the existing system. This provides excellent scalability.

[0051] Enhanced data reliability and fault tolerance: Data is redundantly stored across multiple nodes. Even if a node fails, data is not lost, ensuring data reliability. Distributed storage systems typically include automatic data recovery mechanisms that can quickly restore data after a node failure, ensuring continuous system operation.

[0052] Improved data access speed and performance: Distributed storage architectures can store data across multiple nodes. Users access data from the nearest node, reducing data transmission latency and increasing data access speed. Through parallel processing technology, multiple nodes can process data requests simultaneously, significantly improving overall system performance.

[0053] Optimize resource utilization and cost-effectiveness: Distributed storage architectures can dynamically allocate storage resources based on actual demand, avoiding resource waste. Compared with centralized storage, distributed storage architectures typically have lower hardware and maintenance costs, improving cost-effectiveness.

[0054] Support for big data processing and analysis: Deepwater drilling projects generate massive amounts of multi-source, heterogeneous data. Distributed storage architectures enable efficient storage and management of this data. Combined with big data processing technologies, this allows for in-depth mining and analysis of data stored in distributed databases, providing strong support for drilling project optimization and decision-making.

[0055] Enhanced system flexibility and adaptability: The distributed storage architecture can be customized and optimized based on the specific needs of drilling projects, meeting the data storage and management requirements in different scenarios. As drilling technology or the environment changes, the distributed storage architecture can be easily adjusted and upgraded to meet new requirements.

[0056] In a specific implementation scheme, the data service architecture module adopts a three-tier B / S architecture.

[0057] Specifically, the benefits of the three-tier B / S architecture adopted by the data service architecture module include:

[0058] 1. Improve system maintainability and upgrade convenience

[0059] Centralized maintenance management

[0060] The three-tier B / S architecture centralizes business logic on the server side, with the client side responsible only for data display and simple interaction. When system maintenance or upgrades are needed, they can be performed solely on the server side, eliminating the need to update client software individually. For example, in a deepwater drilling multiphase flow simulation system, optimizing the data calculation logic requires simply modifying the corresponding code on the server side. All clients will automatically receive the updated service the next time they access it, significantly reducing maintenance costs and workload.

[0061] Reduce the difficulty of upgrading

[0062] Clients don't need to install complex software; they simply access the system through a browser. When system features are upgraded, users can simply refresh the page to use the new functionality without having to manually download and install the update package. This avoids compatibility issues caused by inconsistent client software versions and improves the overall stability and usability of the system.

[0063] 2. Enhance system scalability and flexibility

[0064] Easy to expand business functions

[0065] The three-tier architecture organizes different functional modules into layers, with each layer communicating via interfaces. When new business functionality is needed, the corresponding business logic layer module can be added on the server side without requiring major changes to the client and database layers. For example, adding a new drilling parameter analysis function to a deepwater drilling simulation system simply requires developing a new analysis algorithm module in the business logic layer and adding the corresponding display page in the presentation layer.

[0066] Adapt to different client devices

[0067] Clients only need a browser to access the system, regardless of operating system or hardware. Whether using a desktop, laptop, or mobile device, anyone with a browser can easily use the system. This allows deepwater drilling site personnel to access the simulation system anytime, anywhere, on any device, obtaining the required data and analysis results, thus improving work efficiency.

[0068] 3. Improve system security and data confidentiality

[0069] Centralized data storage and management

[0070] Data is stored in a server-side database. Clients cannot access the database directly and can only obtain data through interfaces provided by the server. This centralized data storage method facilitates unified data management and security control, effectively preventing data leaks and unauthorized access. For example, deepwater drilling projects involve a large amount of sensitive data, such as drilling parameters and geological information. A three-tier B / S architecture ensures the security of this data.

[0071] Fine-grained control of user permissions

[0072] The server can perform unified identity authentication and permission management for users, assigning different operational permissions based on their roles and responsibilities. For example, ordinary staff members might only be allowed to view and query data, while senior management might be authorized to perform operations such as data modification and system configuration. This refined permission control mechanism effectively prevents users from exceeding their authorized permissions and ensures the secure operation of the system.

[0073] 4. Optimize user experience and ease of operation

[0074] Simplify client operations

[0075] Clients can access the system using only a browser, eliminating the need for complex software installation and configuration. The user interface is simple and intuitive, and the operational process is easy to understand, reducing the user learning curve. For example, in the deepwater drilling simulation system, workers can easily input drilling parameters, view simulation results, and receive real-time optimization recommendations through a browser, without requiring specialized software skills.

[0076] Support remote access and collaboration

[0077] Thanks to its B / S architecture, users can access the system anytime, anywhere via the internet, enabling remote work and collaboration. Workers in different locations can log in simultaneously to simulate and analyze multiphase flow in drilling and discuss optimization solutions. This breaks down geographical limitations and improves work efficiency and team collaboration.

[0078] 5. Reduce system deployment and usage costs

[0079] Reduce client hardware requirements

[0080] The client only needs basic browser functionality and does not require high-performance computer hardware. This reduces the procurement and maintenance costs of client equipment, making it particularly suitable for locations with harsh environmental conditions such as deepwater drilling sites, where simple and durable devices can be used as clients.

[0081] Reduce software licensing and training costs

[0082] Clients don't need to install specialized software or purchase software licenses, reducing software procurement costs. Furthermore, because the system is easy to operate and understand, user training costs are relatively low. Enterprises can invest more resources in core business and technology research and development, improving their competitiveness.

[0083] In a specific implementation scheme, the data standardization and cleaning fusion module includes:

[0084] Missing value processing submodule, used to process missing values ​​of multi-source heterogeneous data during drilling;

[0085] Noise data identification submodule, used to identify noise from multi-source heterogeneous data during drilling;

[0086] The data fusion submodule is used to fuse multi-source heterogeneous data during the drilling process.

[0087] Specifically, the benefits of the data standardization and cleaning fusion module include:

[0088] 1. Improve data quality and provide a reliable basis for subsequent analysis

[0089] Missing value processing submodule

[0090] Avoiding Data Bias: If missing values ​​in heterogeneous, multi-source drilling data are not addressed, subsequent analysis may lead to biased conclusions based on an incomplete dataset. For example, when analyzing the relationship between drilling speed and drill bit pressure, if drill bit pressure data for some samples is missing, direct analysis may lead to incorrect judgments about the relationship. The Missing Value Handling submodule fills missing values ​​using appropriate methods (such as mean filling, median filling, and regression prediction filling), making the dataset more complete and reducing bias caused by missing data.

[0091] Improved data availability: Many data analysis algorithms and models require data integrity. Missing values ​​can cause algorithms to malfunction or produce inaccurate results. By addressing missing values, data can meet the input requirements of more algorithms, improving data availability and enabling subsequent simulations and real-time optimization to be conducted based on more complete datasets.

[0092] Noise data identification submodule

[0093] Ensuring Data Accuracy: Noise data refers to abnormal data that deviates significantly from the actual data and is caused by measurement errors, data transmission errors, and other factors. During the drilling process, sensors may be subject to interference, generating noise data. If this noise is not identified and eliminated, it can seriously affect the accuracy of data analysis. For example, when measuring drilling fluid flow, electromagnetic interference may cause an abnormal increase in flow data at a certain moment. If this noise data is not identified and processed, it will affect the proper judgment of the drilling fluid circulation system. The noise data identification submodule can accurately identify this abnormal data and ensure data accuracy.

[0094] Enhanced model stability: Using datasets containing noisy data to train models can lead to overfitting or instability, reducing the model's generalization ability. By identifying and removing noisy data, the training data can be made purer, thereby improving model stability and prediction accuracy, providing more reliable support for complex multiphase flow simulations in deepwater drilling.

[0095] 2. Realize data integration and tap into the potential value of data

[0096] Data fusion submodule

[0097] Integrating Multi-Source Information: The drilling process involves a variety of data sources, such as sensor data, geological exploration data, and drilling logs. These data have varying formats, structures, and semantics. The data fusion submodule integrates data from these diverse sources, eliminating redundancy and inconsistencies and creating a unified, complete data view. For example, fusing drilling parameter data collected in real time by sensors with formation information from geological exploration provides a more comprehensive understanding of the geological conditions and drilling status during the drilling process.

[0098] Uncovering Potential Relationships: Data fusion can uncover potential relationships and patterns between different data sources, providing more valuable information for drilling project decision-making. For example, fusing vibration data from drilling with drill bit wear data may reveal correlations between vibration characteristics and drill bit wear. This can help predict drill bit replacement time in advance, optimize drilling operations, and improve drilling efficiency and quality.

[0099] 3. Improving overall system performance and scientific decision-making

[0100] Positive impact on system performance

[0101] Reduced computing resource consumption: After data standardization, cleaning, and fusion, data volume is more reasonable and of higher quality. This reduces the computational burden on subsequent simulation and real-time optimization modules, improving system efficiency and reducing computing resource requirements. For example, during simulation, processing high-quality data without missing values ​​or noise can reduce the number of iterations and speed up simulations.

[0102] Enhanced system stability: High-quality data can prevent system failures and errors caused by data issues, improving system stability and reliability. In safety-critical deepwater drilling projects, system stability is crucial. The data standardization and cleansing fusion modules ensure stable system operation.

[0103] Improve the scientific nature of decision-making

[0104] Providing Accurate Basis: Based on processed, high-quality data, simulation and real-time optimization modules can produce more accurate results, providing drilling engineers with a more reliable basis for decision-making. For example, when formulating drilling parameter adjustment plans, accurate data analysis and simulation results can help engineers make more reasonable decisions, thereby improving drilling success rates and economic benefits.

[0105] Supporting real-time decision-making: During the drilling process, real-time decisions must be made based on data. The data standardization and cleaning fusion module can quickly process data, providing timely and accurate data support for the real-time optimization module, enabling engineers to adjust drilling strategies and address various issues that arise during the drilling process.

[0106] In a specific implementation scheme, the simulation and real-time optimization module adopts a dynamic simulation calculation model.

[0107] Specifically, the benefits of using a dynamic simulation calculation model in the simulation and real-time optimization module include:

[0108] 1. Improve simulation accuracy to better match actual drilling conditions

[0109] Accurately capture dynamic changes in parameters

[0110] During deepwater drilling, drilling parameters (such as drilling fluid flow, drilling pressure, and rotational speed) and geological conditions (such as formation pressure and permeability) change continuously over time and with drilling depth. Dynamic simulation models can track these dynamic changes in parameters in real time and incorporate them into simulation calculations. For example, as drilling depth increases and formation pressure gradually rises, dynamic models can promptly adjust the simulation of relevant parameters such as drilling fluid density, making the simulation results more similar to the multiphase flow state during actual drilling, and providing more accurate simulation results than static models.

[0111] Considering the coupling effects of multiple factors: During the drilling process, multiple factors interact and couple with each other. For example, the interaction between drilling fluid properties and formation characteristics can affect drilling efficiency. Dynamic simulation models can comprehensively consider the dynamic coupling effects of these factors, more realistically reflecting the complex multiphase flow process in drilling. For example, changes in drilling fluid viscosity not only affect the fluid's rock-carrying capacity but also interact with formation pore pressure. Dynamic models can accurately simulate this complex coupling relationship, improving simulation accuracy.

[0112] 2. Realize real-time optimization to improve drilling efficiency and safety

[0113] Quickly respond to changes in drilling status

[0114] The dynamic simulation calculation model can receive real-time data from the drilling site and rapidly perform simulation calculations and optimization analysis based on this data. When anomalies (such as kicks, lost circulation, etc.) occur during drilling or drilling parameters deviate from normal ranges, the model can immediately respond and provide corresponding optimization suggestions. For example, if an abnormal decrease in drilling fluid return is detected, the dynamic model can quickly simulate the impact of different treatment measures (such as adjusting drilling fluid flow rate and changing drilling fluid properties) on the downhole situation, providing drilling engineers with real-time optimization solutions to promptly resolve drilling issues and prevent further accidents.

[0115] Optimizing Drilling Parameter Combinations: By simulating the drilling effects of different drilling parameter combinations in real time, the dynamic simulation model can find the optimal drilling parameter combination, thereby improving drilling efficiency. For example, during the drilling process, the model can simulate the drilling speed and drill bit wear under different combinations of drilling pressure, rotation speed, and drilling fluid flow in real time, providing engineers with optimal parameter adjustment suggestions to achieve the highest efficiency while ensuring safety.

[0116] 3. Enhance system adaptability and flexibility

[0117] Adapt to different drilling conditions

[0118] Deepwater drilling projects involve a variety of drilling conditions, including different well types (vertical, directional, and horizontal) and different formation types (sandstone, shale, and carbonate). Dynamic simulation models offer strong adaptability, allowing for flexible adjustment and optimization based on these diverse drilling conditions. For example, when drilling horizontal wells, the complex wellbore trajectory creates significantly different drilling fluid flow characteristics compared to vertical wells. Dynamic models can simulate and optimize these characteristics accordingly, providing accurate technical support for horizontal well drilling.

[0119] Supporting the application of new technologies and methods: With the continuous advancement of drilling technology, new drilling processes, tools, and materials are constantly emerging. Dynamic simulation models can easily integrate these new technologies and methods, simulating and evaluating them. For example, when adopting a new drilling fluid system or drill bit, the model can quickly simulate its performance during the drilling process, providing a scientific basis for the promotion and application of new technologies.

[0120] 4. Reduce costs and improve economic benefits

[0121] Reduce trial and error costs

[0122] In the actual drilling process, finding the optimal drilling plan through trial and error often leads to increased costs and extended project times. Dynamic simulation models can simulate and optimize different drilling plans in a virtual environment, identifying potential problems and implementing adjustments in advance. This reduces the number of trial and error steps involved in the actual drilling process and reduces drilling costs. For example, by simulating the drilling costs and benefits under different drilling parameter combinations, the model can select the drilling plan with the lowest cost and highest benefit, avoiding unnecessary waste of resources.

[0123] Extending equipment life: Appropriate drilling parameters and operating methods can reduce wear and damage to drilling equipment, extending its service life. Dynamic simulation models can optimize drilling parameters in real time to prevent equipment from operating under overload or adverse conditions, thereby reducing equipment repair and replacement costs and improving the economic benefits of drilling projects.

[0124] In a specific implementation scheme, the missing value processing submodule adopts a prediction filling algorithm based on machine learning.

[0125] Specifically, the missing value processing submodule uses a machine learning-based predictive filling algorithm, which has the following beneficial effects:

[0126] 1. Improve data integrity and ensure the foundation for analysis

[0127] Maximize the retention of data samples

[0128] In multi-source heterogeneous drilling data, the integrity of data samples is crucial for subsequent analysis. Traditional simple missing value processing methods (such as directly deleting samples with missing values) may lead to the loss of a large amount of valid data, especially when the distribution of missing values ​​is relatively random and the proportion is high. The predictive filling algorithm based on machine learning can use the correlation between other features in the data to predict and fill in missing values, thereby maximizing the retention of data samples. For example, in a drilling parameter data set, if the drilling speed value of a record is missing, by analyzing the potential relationship between other relevant parameters in the record (such as drilling pressure, drill bit type, formation characteristics, etc.) and the drilling speed, the predictive filling algorithm can more accurately estimate the missing drilling speed value, thereby avoiding the loss of other valuable information due to deleting the record.

[0129] Reduce data bias

[0130] Directly deleting samples with missing values ​​or using simple mean or median filling methods may introduce data bias and affect the accuracy of subsequent analysis results. The machine learning-based predictive filling algorithm takes into account the complex relationships between data and can more reasonably estimate missing values, reducing the bias caused by improper handling of missing values. For example, when analyzing the relationship between drilling efficiency and multiple parameters, simply filling the missing drilling time values ​​with the mean may mask the true differences in drilling time under different drilling conditions. The predictive filling algorithm can predict a drilling time that is more in line with the actual situation based on other relevant parameters (such as drilling depth, rock hardness, etc.), thereby making the analysis results more accurate.

[0131] 2. Improve data quality and enhance model performance

[0132] Provide high-quality input data

[0133] Machine learning models and data analysis algorithms require high-quality input data. The presence of missing values ​​can affect model training and prediction accuracy. Machine learning-based prediction-filling algorithms can generate data closer to the true values, providing high-quality input for the model. For example, when using a regression model to predict drilling costs, if the dataset contains a large number of missing drilling material prices and these missing values ​​are not properly handled, the model may learn incorrect data patterns, resulting in inaccurate predictions. Filling in missing drilling material prices with a prediction-filling algorithm allows the model to be trained on more complete and accurate data, improving its predictive performance.

[0134] Mining potential relationships in data

[0135] When predicting missing values, the algorithm automatically discovers potential relationships and patterns between various features in the data. These potential relationships and patterns not only help fill missing values ​​but also provide valuable information for subsequent data analysis and modeling. For example, during the drilling process, the predictive filling algorithm discovered a certain nonlinear relationship between drilling fluid density and formation permeability. This relationship plays an important role in subsequent drilling fluid performance optimization and formation evaluation, helping engineers better understand the physical phenomena during the drilling process.

[0136] 3. Adapt to complex data characteristics and have stronger versatility

[0137] Processing multi-source heterogeneous data

[0138] Drilling projects involve a variety of data sources, such as sensor data, geological exploration data, and drilling logs. These data have different formats, structures, and semantics, and may exhibit complex missing value patterns. Machine learning-based predictive filling algorithms are highly adaptable and can handle missing value issues in heterogeneous multi-source data. For example, the missing value patterns of drilling parameter data from different sensors may be affected by factors such as sensor failure and data transmission interruptions. Predictive filling algorithms can predict missing values ​​using appropriate models based on the data characteristics and interrelationships of different sensors, ensuring that all types of data are effectively processed.

[0139] Coping with different missing mechanisms

[0140] Missing data can be caused by a variety of mechanisms, such as missing completely at random (MCAR), missing at random (MAR), and missing not at random (MNAR). Machine learning-based predictive imputation algorithms can adapt to different missing mechanisms to a certain extent by analyzing data characteristics and missing patterns. For example, even when data are missing not at random, the algorithm can improve the accuracy of missing value prediction by introducing relevant auxiliary variables or adopting specific model structures, making it more versatile than traditional simple imputation methods.

[0141] 4. Support real-time optimization and decision-making to improve drilling efficiency

[0142] Meet real-time data processing needs

[0143] Deepwater drilling requires real-time data acquisition and processing to facilitate timely decision-making and adjustments to drilling strategies. Machine learning-based predictive filling algorithms can rapidly predict and fill missing values ​​by optimizing model structure and calculation methods, meeting the requirements of real-time data processing. For example, when sensor data at a drilling site is missing, the algorithm can quickly predict the missing value and transmit the complete data to the simulation and real-time optimization modules, providing timely decision support for drilling engineers.

[0144] Optimize drilling parameters and processes

[0145] After accurately filling missing values, simulation and real-time optimization based on high-quality data can more accurately predict various drilling conditions and optimize drilling parameters and processes. For example, by accurately simulating and optimizing parameters such as drilling pressure and flow rate, accidents such as well kicks and lost circulation can be avoided, improving drilling safety and efficiency and reducing drilling costs.

[0146] In a specific implementation scheme, the noise data identification submodule uses a box plot.

[0147] Specifically, the beneficial effects of using boxplots in the noise data identification submodule include:

[0148] 1. Visually present data distribution and quickly locate anomalies

[0149] Clearly display data range and central trends

[0150] Boxplots depict data distribution using five key statistics: minimum value, first quartile (Q1), median (Q2), third quartile (Q3), and maximum value. For heterogeneous drilling data from multiple sources, such as drilling pressure data, boxplots can intuitively display the range in which most drilling pressure values ​​fall (the interquartile range (IQR) interval determined by Q1 and Q3), as well as the median position of the data. This helps engineers quickly understand the overall distribution and central tendency of the data, providing a foundation for subsequent noise data identification.

[0151] Quickly identify potential outliers

[0152] According to the outlier identification rules of the boxplot, data points less than Q1-1.5×IQR or greater than Q3+1.5×IQR are generally considered potential outliers. In drilling data, if a drilling speed value is significantly lower than Q1-1.5×IQR or higher than Q3+1.5×IQR, the boxplot can immediately mark it, allowing engineers to quickly identify possible noise data. For example, during normal drilling, the drilling speed typically fluctuates within a certain range, but the boxplot shows that the drilling speed at a certain moment is far outside the normal range. This may mean that the data point is affected by measurement error or other abnormal factors and is therefore noise data.

[0153] 2. The algorithm is simple and efficient, adapting to real-time data processing needs

[0154] Low computational complexity

[0155] The boxplot calculation primarily involves sorting and simple quartile calculations, resulting in a relatively low algorithmic complexity. During the drilling process, large amounts of sensor data must be processed in real time to quickly identify noisy data. The boxplot's low computational complexity enables rapid processing of newly acquired drilling data, promptly identifying noisy data without delays caused by excessive computational effort. For example, in scenarios where drilling parameter data is collected multiple times per second, the boxplot can quickly determine the noise level of each data point, meeting the requirements of real-time data processing.

[0156] Easy to implement and integrate

[0157] The boxplot algorithm is relatively simple to implement, requiring no complex mathematical models or extensive parameter adjustments. This makes it easy to integrate into drilling data processing systems and collaborate with other modules (such as data acquisition and simulation modules). For example, within a drilling data acquisition system, the boxplot algorithm can be directly invoked to perform real-time noise detection on the collected data, and the detection results can be passed to subsequent processing modules, improving the efficiency and reliability of the entire system.

[0158] 3. Loose assumptions about data distribution and wide applicability

[0159] Not dependent on a specific data distribution

[0160] Unlike some noise identification methods that assume a normal distribution, boxplots do not require data to follow a specific distribution. In drilling engineering, drilling data often exhibit non-normal distribution characteristics due to the influence of multiple complex factors (such as geological conditions, drilling technology, and equipment performance). Boxplots can adapt to the uncertainty of this data distribution and still effectively identify noisy data. For example, drilling fluid density data may exhibit a skewed distribution due to variations in formation lithology, but a boxplot can still accurately identify noisy data based on quartiles and outlier identification rules.

[0161] Applicable to different types of data

[0162] Boxplots can be applied to various types of drilling data, including continuous data (such as drilling pressure, temperature, and flow rate) and discrete data (such as drill bit wear levels). For continuous data, boxplots can be used to identify outliers directly based on numerical values. For discrete data, the data can be converted to an appropriate numerical form before the boxplot is applied to identify noise. This wide range of applicability makes boxplots highly practical in drilling data processing.

[0163] 4. Provide visual results to facilitate engineers’ understanding and decision-making

[0164] Intuitive visualization of noisy data

[0165] Boxplots graphically display data distribution and outliers, allowing engineers to intuitively see the location and distribution of noise data within a dataset. Compared to simple lists of values, visual results are easier for engineers to understand and accept. For example, in a set of boxplots containing multiple drilling parameters, engineers can quickly compare the noise data of different parameters, identify which parameters are most susceptible to noise interference, and take targeted measures to address them.

[0166] Assisted decision making

[0167] Noise data identified through boxplots can provide decision-making support for drilling engineers. Based on the type and severity of the noise data, engineers can decide whether to exclude it, perform corrections, or further investigate the cause of the noise. For example, if a sensor's data contains a significant amount of noise, engineers can determine that the sensor may be faulty and arrange for repair or replacement to ensure the accuracy and reliability of drilling data.

[0168] In a specific implementation scheme, the data fusion submodule adopts an entity recognition algorithm.

[0169] Specifically, the beneficial effects of the entity recognition algorithm used in the data fusion submodule include:

[0170] 1. Accurately match and integrate multi-source data entities

[0171] Eliminate data redundancy and ambiguity

[0172] In drilling projects, data comes from a wide range of sources, including sensor data, geological exploration reports, drilling logs, and more. Different data sources may have different descriptions of the same entity (such as a drill bit, formation, drilling equipment, etc.). For example, a drill bit may have different numbers or names in different data records. Entity recognition algorithms can accurately identify the actual entities corresponding to these descriptions, integrating information about the same entity scattered across different data sources to eliminate data redundancy and ambiguity. For example, accurately matching "stratum A" mentioned in a geological exploration report with "stratum in the same depth interval" recorded in the drilling log can avoid data confusion caused by inconsistent entity descriptions.

[0173] Build a complete entity information view

[0174] Through entity recognition, the data fusion submodule builds a comprehensive information view for each entity. For example, for a drill bit, the algorithm can integrate information from multiple sources, including sensor data (such as real-time parameters such as drill bit speed, torque, and wear), equipment maintenance records (such as drill bit replacement time and repair history), and drilling process parameters (such as the drilling parameter settings used when using the drill bit). This enables engineers to fully understand the status and characteristics of each entity from multiple dimensions, providing a richer basis for drilling decisions.

[0175] 2. Improving the Accuracy and Consistency of Data Fusion

[0176] Accurately associate data relationships

[0177] Entity recognition algorithms can accurately identify relationships between data, ensuring the accuracy of data fusion. During the drilling process, complex logical relationships exist between different data. For example, there is a correlation between the composition of the drilling fluid and the characteristics of the formation, and there is also a certain connection between the operating parameters of the drilling equipment and drilling efficiency. Through entity recognition, the algorithm can accurately identify these relationships and correctly fuse the relevant data. For example, it can identify the association between a specific formation type and the corresponding drilling fluid formula. This allows the formation data and the corresponding drilling fluid data to be effectively integrated during data fusion, providing accurate data support for subsequent drilling process optimization.

[0178] Maintaining data consistency

[0179] Maintaining data consistency is crucial during multi-source data fusion. Entity recognition algorithms can unify the descriptions and metrics for the same entity across different data sources, avoiding data inconsistencies caused by differences in data format, units, or semantics. For example, drilling pressure data collected by different sensors may use different units (e.g., MPa, psi). Entity recognition algorithms can convert this data to the same units before fusion, ensuring semantic and numerical consistency in the fused data, improving data quality and usability.

[0180] 3. Enhance data analysis and mining capabilities

[0181] Mining deep data patterns

[0182] After data fusion based on entity recognition algorithms, data analysts can more easily conduct in-depth research on the integrated data. Because the data is organized and associated by entity, analysts can more easily discover the underlying patterns and interactions between different entities. For example, by analyzing the relationship between drill bit wear and drilling parameters under different formation conditions, the underlying patterns of drill bit wear can be discovered, providing a scientific basis for drill bit selection and use, extending drill bit life, and reducing drilling costs.

[0183] Support complex data query and retrieval

[0184] The fused data is organized around entities, making complex data queries and retrieval more efficient and accurate. Engineers can quickly retrieve all relevant data based on a specific entity (such as a specific piece of drilling equipment or a specific formation). For example, when needing to understand the operating status, maintenance records, and related drilling parameters of a piece of drilling equipment over a specific time period, the data structure constructed using the entity recognition algorithm can quickly locate and return relevant information, greatly improving data query efficiency and providing timely support for engineering decision-making.

[0185] 4. Adapting to Complex Drilling Scenarios and Data Changes

[0186] Processing dynamically changing data

[0187] Drilling is a dynamic process, and data constantly changes as drilling progresses. Entity recognition algorithms adapt to this dynamic change, identifying and updating entity information in the data in real time. For example, during the drilling process, drill bits wear and may need to be replaced. Entity recognition algorithms can promptly identify the newly replaced drill bit and integrate its related information with the drilling data, ensuring that the data always reflects the current drilling status.

[0188] Coping with complex drilling scenarios

[0189] In complex drilling scenarios, such as deepwater drilling, data sources are increasingly diverse and relationships are more complex. Entity recognition algorithms, with their strong adaptability and flexibility, are capable of handling data fusion challenges in these complex scenarios. They can identify complex relationships between different types of entities, such as the relationships between various equipment on deepwater drilling platforms, marine environmental parameters, and drilling operations, providing comprehensive data support for decision-making in these complex drilling scenarios.

[0190] In a specific implementation scheme, the three-tier B / S architecture includes a presentation layer, a business logic layer, and a data access layer.

[0191] Specifically, the three-tier B / S architecture includes a presentation layer, a business logic layer, and a data access layer. Its benefits include:

[0192] 1. Presentation Layer

[0193] Improve user experience

[0194] User-friendly interface: The presentation layer is directly user-friendly, featuring an intuitive, concise interface design. Users can easily operate the system without requiring specialized technical knowledge. For example, in a deepwater drilling data management system, the presentation layer displays data and functions using clear charts, intuitive buttons, and concise menus. Drilling engineers can quickly find the information and operations they need, improving work efficiency.

[0195] Multi-device compatibility: Based on the presentation layer of a B / S architecture, users can access the system through browsers on a variety of terminal devices (such as computers, tablets, and mobile phones), regardless of device type or operating system. Whether engineers are analyzing data on their office computers or viewing real-time data on tablets at the drilling site, they can enjoy a consistent user experience.

[0196] Reduce maintenance costs

[0197] Centralized update and maintenance: The presentation layer code and resources are deployed on the server side. When the interface needs to be updated or vulnerabilities need to be fixed, only one operation needs to be performed on the server side. All users can automatically obtain the latest version when accessing through a browser. There is no need for separate installation and configuration on each client device, which greatly reduces maintenance costs and workload.

[0198] 2. Business Logic Layer

[0199] Enhance system flexibility and scalability

[0200] Modular design: The business logic layer divides the system's business functions into modules, with each module responsible for specific business logic processing. For example, in a drilling data processing system, business logic such as data cleaning, data fusion, and simulation can be encapsulated into independent modules. When adding new business functions or modifying existing business logic, only the corresponding module needs to be developed and adjusted without affecting the normal operation of other modules, thereby improving the system's flexibility and scalability.

[0201] Adapting to Business Changes: Drilling engineering operations may adapt to changing technology and project requirements. The modular design of the business logic layer enables the system to quickly adapt to these changes. For example, when new drilling techniques or equipment are introduced, corresponding processing modules can be added to the business logic layer to process and analyze the new data without requiring a large-scale overhaul of the entire system.

[0202] Improve system performance and stability

[0203] Rational resource allocation: The business logic layer can rationally allocate computing resources based on system load. For example, when processing large amounts of drilling data, the business logic layer can distribute computing tasks to multiple server nodes for parallel processing, improving data processing speed and efficiency. Furthermore, by optimizing algorithms and logical processes, unnecessary computing and resource usage are reduced, improving overall system performance.

[0204] Error handling and fault tolerance: The business logic layer has comprehensive error handling and fault tolerance mechanisms, enabling timely detection and resolution of errors that occur during business logic execution. For example, when the data access layer returns abnormal data or a logical error occurs during business logic processing, the business logic layer can provide appropriate error notifications and recovery operations, preventing system crashes or data loss and ensuring stable system operation.

[0205] 3. Data Access Layer

[0206] Ensure data security and consistency

[0207] Data access control: The data access layer strictly controls data access, limiting access and operations to different data based on user roles and permissions. For example, in a drilling data management system, ordinary engineers can only view and manipulate drilling data related to their work, while senior managers can access and manage all data. This effectively prevents unauthorized access and leakage of data, ensuring data security.

[0208] Transaction processing mechanism: The data access layer supports transaction processing, ensuring the atomicity, consistency, isolation, and durability (ACID properties) of data operations. For example, when updating drilling data, the data access layer can process multiple related data operations as a single transaction. If any operation fails, the entire transaction will be rolled back to ensure data consistency.

[0209] Optimizing data access efficiency

[0210] Data caching technology: The data access layer can use data caching technology to store frequently accessed data in memory, reducing the number of direct database accesses. For example, frequently used drilling parameter data can be cached. When users request this data again, they can directly obtain it from the cache, greatly improving data access speed and reducing the burden on the database.

[0211] Database connection management: The data access layer is responsible for managing database connections, properly allocating and releasing database connection resources, and avoiding connection leaks and excessive connections. By optimizing the configuration of the database connection pool, the database's concurrent processing capabilities are improved, ensuring efficient data access in high-concurrency scenarios.

[0212] 4. Overall Architecture Advantages

[0213] Facilitates team collaborative development

[0214] Layered and Clear Responsibilities: A three-tier B / S architecture divides system functionality into distinct layers, with each layer having clearly defined responsibilities and functions. Development teams can assign development tasks to different layers based on their members' expertise and specialties, improving development efficiency and quality. For example, the presentation layer can be managed by front-end developers, the business logic layer by back-end developers, and the data access layer optimized by database experts. Clear interfaces enable interaction between these layers, reducing coupling during the development process.

[0215] Parallel development and testing: Because each layer is relatively independent, developers can carry out development and testing work at different levels in parallel. For example, while designing the interface and developing interactions at the presentation layer, the business logic layer and data access layer can also be developing corresponding functions and conducting unit testing, shortening the system development cycle.

[0216] Reduce system upgrade and migration costs

[0217] Independent Layer Upgrades: When system upgrades or migrations are required, the layered design of the three-tier B / S architecture allows for more flexible upgrades. For example, if a database upgrade is needed, only adjustments and testing are required at the data access layer, eliminating the need for large-scale modifications to the presentation and business logic layers. Similarly, if the front-end technology framework needs to be replaced, only the presentation layer needs to be restructured, without impacting the normal operation of the business logic and data access layers, thus reducing the cost and risk of system upgrades and migrations.

[0218] exist Figure 2 In the embodiment of the present application, a method for simulating complex multiphase flow in deepwater drilling is provided, comprising the following steps:

[0219] Use the drilling multiphase flow simulation database to store and manage multi-source heterogeneous data during the drilling process;

[0220] Use the data service architecture module to build a data service architecture for drilling multiphase flow simulation;

[0221] Use data standardization and cleaning fusion modules to process multi-source heterogeneous data during drilling;

[0222] The simulation and real-time optimization modules are used to simulate and optimize the multiphase flow in drilling.

[0223] In the above technical solution, the multi-source heterogeneous data in the drilling process are stored and managed by utilizing the drilling multiphase flow simulation database; the data service architecture module is used to construct the data service architecture for drilling multiphase flow simulation; the data standardization and cleaning fusion module is used to process the multi-source heterogeneous data in the drilling process; the simulation and real-time optimization module is used to simulate and optimize the drilling multiphase flow in real time; and the real-time simulation and optimization of multiphase flow in deepwater drilling projects are realized.

[0224] Those skilled in the art will appreciate that the present application may be implemented as a system, method, or computer program product.

[0225] Therefore, the present disclosure may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, the present disclosure may be implemented in the form of a computer program product embodied in one or more computer-readable media, wherein the computer-readable media contains computer-readable program code.

[0226] Any combination of one or more computer-readable media can be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device.

[0227] Although the embodiments of the present application have been shown and described above, it should be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application. Various substitutions and improvements may be made to the present application on this basis, all of which fall within the scope of protection of the present application.

Claims

1. A deepwater drilling complex drilling multiphase flow simulation system, characterized in that: include: Drilling multiphase flow simulation database, used to store and manage multi-source heterogeneous data during the drilling process; Data service architecture module, used to build the data service architecture for drilling multiphase flow simulation; Data standardization, cleaning and fusion module, used to process multi-source heterogeneous data during the drilling process; The simulation and real-time optimization module is used to simulate and optimize drilling multiphase flow in real time.

2. The deepwater drilling complex drilling multiphase flow simulation system according to claim 1 is characterized in that: The drilling multiphase flow simulation database adopts a distributed storage architecture.

3. The deepwater drilling complex drilling multiphase flow simulation system according to claim 2 is characterized in that: The data service architecture module adopts a three-tier B / S architecture.

4. The deepwater drilling complex drilling multiphase flow simulation system according to claim 3 is characterized in that: The data standardization and cleaning fusion module includes: Missing value processing submodule, used to process missing values ​​of multi-source heterogeneous data during drilling; Noise data identification submodule, used to identify noise from multi-source heterogeneous data during drilling; The data fusion submodule is used to fuse multi-source heterogeneous data during the drilling process.

5. The deepwater drilling complex drilling multiphase flow simulation system according to claim 4 is characterized in that: The simulation and real-time optimization module adopts a dynamic simulation calculation model.

6. The deepwater drilling complex drilling multiphase flow simulation system according to claim 5 is characterized in that: The missing value processing submodule adopts a prediction filling algorithm based on machine learning.

7. The deepwater drilling complex drilling multiphase flow simulation system according to claim 6 is characterized in that: The noise data identification submodule uses a box plot.

8. The deepwater drilling complex drilling multiphase flow simulation system according to claim 7 is characterized in that: The data fusion submodule adopts an entity recognition algorithm.

9. The deepwater drilling complex drilling multiphase flow simulation system according to claim 8, characterized in that: The three-tier B / S architecture includes a presentation layer, a business logic layer and a data access layer.

10. A method for simulating multiphase flow in complex deepwater drilling, characterized in that: The following steps are involved: Use the drilling multiphase flow simulation database to store and manage multi-source heterogeneous data during the drilling process; Use the data service architecture module to build a data service architecture for drilling multiphase flow simulation; Use data standardization and cleaning fusion modules to process multi-source heterogeneous data during drilling; The simulation and real-time optimization modules are used to simulate and optimize the multiphase flow in drilling.

Images