Drilling process optimization method and system based on digital twinning

Abstract

The application discloses a drilling process optimization method and system based on digital twinning, relates to the technical field of digital twinning, establishes respective infinite data chains corresponding to respective sensors, deploys multistage data transmission relay nodes on the paths from the respective sensors to data receiving ends, performs preprocessing and buffering on the sensor data that arrives in advance at respective transmission relay nodes, waits for preservation in the case of incomplete data, stores the sensor data of future time nodes received during the preservation waiting period into a secondary cache queue according to time nodes, performs compression preservation on overflow data based on the cache residual amount and the waiting time length, and aligns and injects data according to the time axis of the drilling digital twinning model when it is detected that the downhole sensor data of the same time node all arrive at the transmission relay nodes. The application realizes accurate synchronization and alignment of multichannel drilling data on the time axis of the digital twinning model.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology, specifically to a drilling process optimization method and system based on digital twins. Background Technology

[0002] Drilling is a complex systems engineering project, often requiring continuous operation in extreme environments. With the development of sensor technology, the Internet of Things (IoT), and digital technologies, the drilling process deploys numerous sensors to monitor various parameters downhole and on the surface in real time, such as drilling pressure, rotational speed, torque, mud flow rate, bottom hole pressure, and temperature. The real-time data generated by these sensors is transmitted to the surface control center via a transmission network to monitor the drilling status and guide operations. In recent years, the application of digital twin technology in the drilling field has attracted significant attention. A digital twin is a virtual model created in virtual space that corresponds to the physical drilling system. By continuously injecting real-time sensor data, the virtual model can synchronously and accurately reflect the dynamic evolution of the physical drilling process.

[0003] However, to fully leverage the capabilities of drilling digital twins, high-quality real-time data input is essential. This requires data from multiple downhole sensors to be synchronously delivered to the digital twin model on a unified time reference. In reality, the different locations, sampling frequencies, data transmission paths, and media of downhole sensors, coupled with the complex and variable downhole environment, often lead to inconsistencies in the arrival times of data from different sensors at the surface data receiver. In other words, even if all sensors physically collect data simultaneously, factors such as network latency and data rate differences may cause some sensor data to arrive at the surface earlier, while others may lag behind. This asynchronous data arrival problem severely impacts the real-time performance and accuracy of the digital twin model. If the asynchronous data is used directly to drive the digital twin without processing, some parameters in the virtual model may be updated to the current moment, while others remain at a previous moment, causing a misalignment between the digital twin state and the actual drilling state, thus weakening the decision support effect.

[0004] Therefore, in order to address the above problems, there is an urgent need for drilling process optimization methods and systems based on digital twins. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a drilling process optimization method and system based on digital twins, which solves the problem of multi-sensor data synchronization and enhances the system's real-time data acquisition and processing capabilities in complex downhole environments.

[0006] To achieve the above objectives, this invention provides the following technical solution: a drilling process optimization method based on digital twins, comprising the following steps: establishing a unified drilling data timeline, forming an infinite data chain corresponding to each sensor, wherein the infinite data chain is composed of multiple time nodes connected sequentially, and the infinite data chains of each sensor are connected according to the same time nodes; deploying multi-level data transmission relay nodes on the path from each sensor to the data receiving end, configuring differentiated relay buffer spaces for different sensors, forming a hierarchical transmission link; and performing pre-processing on the sensor data arriving in advance at each transmission relay node. The system processes and caches data, while simultaneously monitoring the arrival of data at the same time point on other sensor data chains. For data that is not fully received, it performs a "fresh" waiting period. For sensor data received at future time points during this waiting period, it categorizes and stores the data in a secondary cache queue according to time point, and compresses and preserves overflowing data based on cache availability and waiting time. When all downhole sensor data at the same time point arrives at the transmission relay node, it extracts and integrates the data from each sensor in the cache space for that time point, sends the integrated synchronization data packet to the data receiver, and then aligns and injects the data according to the timeline of the drilling digital twin model.

[0007] Furthermore, the step of establishing a unified drilling data timeline is as follows: attach a collection timestamp to the data collected by each downhole sensor, map the data from different sensors to the same time node sequence according to the collection timestamp, and mark the corresponding data index of each sensor to achieve synchronous alignment of data from different sensors under the same time reference, thereby forming a unified drilling data timeline.

[0008] Furthermore, the specific analysis of the corresponding connection of the infinite data chains of each sensor according to the same time node is as follows: For the raw data collected in real time by each downhole sensor, it is divided into periods based on a preset time period, with each time period as a fixed time window. All collected data within the time window is aggregated to form a one-to-one corresponding virtual data unit, which serves as the time node in the infinite data chain. The time window is uniformly set across all sensors, so that the virtual data units formed by all sensors within the same time period have the same time identifier on the time axis. The infinite data chain formed by each sensor is composed of multiple virtual data units with consistent time connected sequentially. The corresponding nodes between each infinite data chain have a consistent time identifier, realizing a one-to-one correspondence between data from different sensors in logical time. The corresponding connection allows time nodes with the same time identifier to be synchronously extracted from the infinite data chains of multiple sensors for transmission at any time node.

[0009] Furthermore, the multi-level data transmission relay node includes relay devices arranged in at least two levels sequentially. The analysis of configuring differentiated relay buffer space for different sensors is as follows: based on fuzzy reasoning based on the acquisition frequency and transmission delay, the downhole sensors are divided into a first type of sensor, a second type of sensor, and a third type of sensor; the relay buffer space type is determined based on the relay buffer space size, including a first type of relay buffer space, a second type of relay buffer space, and a third type of relay buffer space; a first type of relay buffer space is allocated to the first type of sensor, a second type of relay buffer space is allocated to the second type of sensor, and a third type of relay buffer space is allocated to the third type of sensor.

[0010] Furthermore, the preprocessing and caching specifically involve: performing preliminary feature extraction on the sensor data that arrives early at each transmission relay node, and temporarily storing the data after preliminary feature extraction in the transmission relay node cache; the freshness waiting specifically involves refreshing or heartbeat monitoring when a preset refresh cycle is reached.

[0011] Furthermore, the secondary cache queue is temporarily stored using a circular buffer or a sliding window.

[0012] Further, the specific analysis of compressing and preserving overflow data based on cache space and waiting time is as follows: During the preservation waiting period, for data from future time nodes continuously generated and arriving by the sensor, they are sequentially categorized and stored in a preset secondary cache queue according to their corresponding time nodes; when it is detected that the occupied space of the cache queue reaches a preset space threshold, and there is cached data waiting time exceeding a preset time threshold, the compression and preservation mechanism is activated to perform fuzzy inference filtering on the overflow data. The compression and preservation mechanism includes the following steps: collecting the cache waiting time and current cache space space for each data frame; constructing a fuzzy rule base based on the cache waiting time and cache space space as input factors, using a fuzzy inference engine to perform fuzzy judgment on each data frame to be processed, and outputting the retention strength level of the data frame; comparing the fuzzy inference output result with the set retention strength threshold, and performing compression operation on data frames with retention strength levels lower than the retention strength threshold, the compression operation including removing redundant fields, retaining only key feature quantities, or storing representative information in summary form; and reordering and adapting data frames with retention strength levels greater than or equal to the retention strength threshold into the remaining cache.

[0013] The drilling process optimization system based on digital twins, applying the aforementioned drilling process optimization method based on digital twins, includes: an infinite data chain generation module, used to establish a unified drilling data timeline, forming an infinite data chain corresponding to each sensor, wherein the infinite data chain is composed of multiple time nodes connected sequentially, and the infinite data chains of each sensor are connected according to the same time nodes; a hierarchical transmission link formation module, used to deploy multi-level data transmission relay nodes on the path from each sensor to the data receiver, configuring differentiated relay buffer spaces for different sensors to form a hierarchical transmission link; and a freshness buffer module, used to store early-arriving sensor data at each transmission relay node. The system performs preprocessing and caching, while simultaneously monitoring the arrival of data at the same time point on other sensor data chains. For data not yet fully received, it implements a preservation waiting period. A dynamic buffer module categorizes sensor data received during the preservation waiting period into a secondary cache queue based on time point, and compresses and preserves overflowing data based on cache availability and waiting time. A synchronous packet transmission module extracts and integrates the data from each sensor at that time point from the cache space when all downhole sensor data at the same time point arrives at the transmission relay node. This integrated synchronous data packet is then sent to the data receiving end for alignment and data injection according to the timeline of the drilling digital twin model.

[0014] The present invention has the following beneficial effects:

[0015] This drilling process optimization method and system based on digital twins ensures strict temporal alignment of data from multiple downhole sensors through a unified timeline and a freshness preservation mechanism. Each update of the digital twin model uses data corresponding to the same moment in the physical system, avoiding virtual model deviations caused by data asynchrony in traditional solutions and improving the synchronization accuracy of the digital twin with the real drilling process. Through multi-level caching and dynamic buffering strategies, data loss is minimized. A compression and freshness preservation mechanism is employed; when cache overload risks occur, the system does not simply discard delayed data but compresses and simplifies it, retaining key information and saving it in summary form. This ensures that even if some original data cannot be fully retained due to excessive waiting time, the digital twin model can still obtain data for that time period. The summary information ensures the integrity and continuity of the data link, enhancing the system's robustness. Differentiated cache allocation and fuzzy inference scheduling enable the system to accommodate the data transmission needs of both high-speed and low-speed sensors. For fast sensor data, sufficient cache space and preprocessing are provided so that it can wait for slow data without failing. For slow sensor data, mechanisms such as heartbeat monitoring prevent indefinite waiting, thus ensuring synchronization while avoiding unnecessary delays. The introduction of fuzzy inference algorithms enables intelligent decision-making for cache management and data compression, allowing the system to adaptively adjust its strategies based on the current operating state. Compared to traditional methods with fixed rules or thresholds, intelligent control improves system resource utilization efficiency, reduces manual intervention, and increases the system's automation level.

[0016] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0017] Figure 1 This is a framework diagram of the drilling process optimization method based on digital twins of the present invention.

[0018] Figure 2 This is a schematic diagram of the drilling process optimization method based on digital twins according to the present invention.

[0019] Figure 3 This is a schematic diagram of data alignment and injection for the drilling digital twin model of the present invention.

[0020] Figure 4 This is a schematic diagram illustrating the synchronous alignment and reception of multi-sensor drilling data according to the present invention.

[0021] Figure 5 This is a structural diagram of the drilling process optimization system based on digital twins according to the present invention. Detailed Implementation

[0022] This application embodiment achieves precise synchronous alignment of multi-sensor drilling data on the time axis of the digital twin model through a drilling process optimization method and system based on digital twins, while taking into account both real-time transmission efficiency and data integrity.

[0023] The problem addressed in this application's embodiments can be summarized as follows:

[0024] By establishing a unified timeline to integrate multi-sensor data, deploying multi-level relay nodes with differentiated configuration caching, preprocessing and caching data that arrives early and waiting for data from the same time node, compressing and preserving overflowing data using fuzzy inference based on cache availability and waiting time, and integrating and transmitting data after all data from the same time node is complete, the digital twin model data alignment and injection is achieved.

[0025] Please see Figure 1 , Figure 2 , Figure 3 , Figure 4 This invention provides a technical solution: a drilling process optimization method based on digital twins, comprising the following steps: establishing a unified drilling data timeline to form an infinite data chain corresponding to each sensor, wherein the infinite data chain is composed of multiple time nodes connected sequentially, and the infinite data chains of each sensor are connected according to the same time nodes; deploying multi-level data transmission relay nodes on the path from each sensor to the data receiving end, configuring differentiated relay buffer spaces for different sensors to form a hierarchical transmission link; and performing preprocessing and buffering on the sensor data that arrives early at each transmission relay node. The system stores data and monitors the arrival of data at the same time point on other sensor data chains. For data that is not fully received, it performs a preservation wait. For sensor data received at future time points during the preservation wait period, it is categorized by time point and stored in a secondary cache queue. Overflowing data is compressed and preserved based on cache remaining capacity and waiting time. When all downhole sensor data at the same time point are detected to have arrived at the transmission relay node, the system extracts the data of that time point from each sensor in the cache space, integrates them, and sends the integrated synchronization data packet to the data receiving end. Then, it aligns and injects the data according to the timeline of the drilling digital twin model.

[0026] Specifically, the analysis of establishing a unified drilling data timeline is as follows: A timestamp is added to each data point collected by each downhole sensor. Data from different sensors are mapped to a unified time node sequence based on the timestamps. The sensor identifier and sequence number from which each data point originates are recorded, thus forming an "infinite data chain" corresponding to each sensor. This infinite data chain refers to the time sequence formed by the continuous generation of data by each sensor. After being mapped to the unified timeline, the time interval between adjacent data points is fixed and consistent with the sampling interval. Theoretically, it can be extended to an infinitely long sequence. By establishing a unified timeline, a unified time reference can be provided for data from different sensors, making it possible to achieve cross-sensor data synchronization and alignment in subsequent steps.

[0027] The specific analysis of connecting the infinite data chains of various sensors according to the same time node is as follows: For the raw data collected in real time by each downhole sensor, it is divided into periods based on a preset time period. Each time period is a fixed time window. All collected data within the time window is aggregated to form a one-to-one virtual data unit. The virtual data unit serves as the time node in the infinite data chain. The time window is uniformly set across all sensors, so that the virtual data units formed by all sensors within the same time period have the same time identifier on the time axis. The infinite data chain formed by each sensor is composed of multiple virtual data units with consistent time connected sequentially. The corresponding nodes between the infinite data chains have the same time identifier, realizing a one-to-one correspondence between the data of different sensors in logical time. The corresponding connection allows time nodes with the same time identifier to be synchronously extracted from the infinite data chains of multiple sensors for transmission at any time node.

[0028] Specifically, the analysis of forming a hierarchical transmission link is as follows: deploy multi-level data transmission relay nodes on the transmission path from each sensor to the ground data receiving end to construct a hierarchical transmission link. The transmission link includes at least two levels of relay devices, such as a first-level relay near the drill bit downhole, an intermediate relay in the wellbore, and a higher-level relay at the wellhead on the ground. Data is transmitted sequentially level by level. These relay nodes can be gateways or data servers with storage and processing capabilities, used for temporary storage, processing, and forwarding of sensor data. Different relay buffer space sizes are configured for different sensors to adapt to the differences in data generation rate and transmission delay of each sensor.

[0029] The specific analysis of configuring differentiated relay buffer spaces for different sensors is as follows: Fuzzy inference is used to determine the buffer configuration for different sensors. The specific steps are as follows: Based on the sensor's acquisition frequency and transmission delay, a fuzzy rule base is established to classify downhole sensors into three categories: Type I sensors, Type II sensors, and Type III sensors. Simultaneously, the buffer space provided by relay nodes is divided into Type I, Type II, and Type III buffer spaces. The fuzzy inference engine infers and classifies the sampling frequency and delay of each sensor: for example, sensors with high acquisition frequency and low transmission delay may be classified as Type I (high-speed sensors), those with low frequency and high delay as Type III (low-speed or slow-speed sensors), and those with other moderate characteristics as Type II. Type I sensors are allocated Type I relay buffer space (larger buffer space), Type II sensors are allocated Type II relay buffer space (medium-sized buffer), and Type III sensors are allocated Type III relay buffer space (smaller buffer space). This differentiated allocation ensures that the large amount of data generated by high-speed sensors can be temporarily cached, preventing overflow and loss while waiting for slow sensors, and also avoids resource waste caused by allocating too much buffer space to low-speed sensors.

[0030] In this implementation scheme, a fuzzy rule base is established based on two factors: sensor acquisition frequency and transmission delay. The specific analysis for classifying downhole sensors into three categories—Type 1, Type 2, and Type 3—is as follows: Based on acquisition frequency and transmission delay as input factors, sensors are fuzzily classified into multiple semantic levels. For example, acquisition frequency can be categorized as "low," "medium," and "high," and transmission delay as "short," "medium," and "long." A two-dimensional fuzzy rule base is constructed accordingly, where each rule corresponds to a sensor type judgment condition. A fuzzy inference engine is used to perform fuzzy calculations on the acquisition frequency and transmission delay of each sensor, and the sensor's classification level is output based on the preset rule base. Based on the fuzzy inference results, all sensors are classified into three preset types: Type 1 sensors are those with high acquisition frequency and short transmission delay; Type 2 sensors are those with both acquisition frequency and transmission delay at intermediate levels; and Type 3 sensors are those with low acquisition frequency and long transmission delay.

[0031] The specific analysis of dividing the cache space provided by relay nodes into three types of cache space is as follows: Obtain the available cache resource parameters of each relay node during the initialization phase or runtime, including total capacity, currently allocable space, read / write rate, etc., as the basis for division; based on the overall system cache capacity distribution, set a cache space division benchmark to form cache capacity ranges. For example, a total capacity higher than a first preset threshold is defined as a "large capacity cache area," between the first and second thresholds is a "medium capacity cache area," and below the second threshold is a "small capacity cache area"; these capacity ranges are mapped to three types of relay cache space: the first type corresponds to a large capacity cache area, suitable for high-frequency, high-flow sensors; the second type corresponds to a medium capacity cache area, suitable for medium-frequency sensors; and the third type corresponds to a small capacity cache area, suitable for low-frequency, low-flow sensors.

[0032] Specifically, the analysis of preprocessing, caching, and freshness-preservation waiting is as follows: A freshness-preservation caching mechanism is set up at each relay node of the hierarchical transmission link. When data from a certain downhole sensor arrives at a relay node before other sensors, the relay node immediately preprocesses the early-arriving data and temporarily stores it in the cache space. Preprocessing can include preliminary feature extraction, format conversion, or noise filtering of the original data as needed, thereby reducing storage usage while maintaining key data features. This allows the data to receive preliminary "freshness-preservation" during the cache waiting period, preventing its value from decreasing due to long waiting times. Simultaneously, the relay node monitors the arrival status of data from other sensors at the same time point. If it finds that the data from other sensors corresponding to that time point has not yet arrived in full, then... Entering a freshness-preservation waiting state means that the received data is not sent upwards temporarily, but is kept fresh in the buffer after preprocessing. The relay node has a refresh cycle and a heartbeat monitoring mechanism: during the waiting process, if the preset refresh time cycle is reached, the relay node can perform a data status refresh or check. For example, it can send a heartbeat signal to query the sensors or lower-level relays that have not yet received data to confirm their working status, or simply record the types of data that are still missing and update the waiting time to prevent indefinite waiting for data from a certain sensor. If the heartbeat monitoring finds that a sensor may be offline due to failure, it decides to skip the data of that sensor to continue the process, for example, by using the data from the previous moment or setting it to null, and issuing an alarm when necessary, thereby ensuring the continuity of the overall data flow.

[0033] In this implementation plan, the preliminary feature extraction analysis is as follows: Based on the sensor data type, extract the core observation fields from its structure, such as displacement, amplitude, voltage, current, and torque; within a unit time window or fixed sampling segment, calculate the statistical characteristics of the raw data, including but not limited to mean, extreme values, standard deviation, rate of change, slope, skewness, and kurtosis, to compress and express the data change trend; perform necessary normalization on the characteristics to remove the influence of dimensions, and combine dynamic range control technology to compress the characteristic values ​​to a set accuracy, further reducing storage space requirements; and uniformly convert the preprocessed feature data into the system-defined standard data format to adapt to subsequent caching and transmission module calls.

[0034] To prevent relay nodes from indefinitely waiting for data from sensors that have not yet arrived during the data preservation waiting process, the system employs a refresh mechanism and a heartbeat monitoring mechanism to dynamically assess the data completeness status and sensor operating status. The refresh mechanism involves setting an adjustable refresh cycle in the relay node. Upon reaching the cycle, a cache status check is triggered, including updating the current target time node, recording whether data from each sensor has been received at that time node, and reassessing whether the data completeness conditions are met. If met, the data packaging and transmission process is triggered; otherwise, the waiting timer is reset. The heartbeat monitoring mechanism involves the relay node actively sending an inquiry request to the sensor whose data has not yet arrived or its upstream relay node to confirm whether the data source is still online or experiencing a communication failure. This includes sending a handshake request to the target device, setting a response timeout, and recording its status as "online but delayed" if a response is received within the timeout period. If no response is received after multiple consecutive refresh cycles, the device is determined to be "disconnected" or "offline."

[0035] Specifically, the analysis of secondary cache queuing and compression preservation is as follows: During the preservation waiting period, data from future time nodes that arrive ahead of schedule will continue to arrive at the relay node. To manage this data from future time moments, it is categorized according to its time tag and stored in the secondary cache queue within the node. The secondary cache queue is used to temporarily store subsequent time node data that has not yet been sent synchronously. In order to effectively utilize memory and prevent the cache from growing indefinitely during the waiting process, the secondary cache queue preferably adopts a circular buffer or sliding window structure. On the one hand, when the circular buffer is full, new data will overwrite the oldest data, and the sliding window can discard outdated data as time moves forward. This structure can limit the maximum capacity of the cache and prevent the cache from expanding indefinitely. On the other hand, it also facilitates the management of data in chronological order.

[0036] As waiting time increases and new data continues to enter the secondary cache queue, the cache space may gradually become full. To prevent data loss due to buffer overflow, a "compression and preservation" strategy is introduced in the relay node. When the space occupied by the secondary buffer queue reaches a preset threshold, and some data has been waiting for more than the preset time threshold without being synchronized, the compression and preservation mechanism is activated. The core of the compression and preservation mechanism is to compress or simplify data frames with long waiting times and large space occupation while preserving key information as much as possible, so as to free up buffer space and ensure that the overall data chain is not broken. The specific process of activating the compression and preservation mechanism is as follows: Collect the current buffer waiting time of each data frame to be processed in the secondary buffer queue, as well as the current buffer space remaining of the system. Use the waiting time and buffer space remaining as input factors for fuzzy inference to construct a corresponding fuzzy rule base. For example, the waiting time can be divided into three fuzzy sets: "short", "medium", and "long", and the buffer space remaining can be divided into three fuzzy sets: "sufficient", "average", and "insufficient". Predefine several fuzzy rules, such as: "If the waiting time is long and the buffer space remaining is insufficient, then the retention strength level of the data frame is low". Use the fuzzy inference engine to infer and calculate the retention strength level of each data frame in the secondary buffer queue. The retention level is an evaluation value reflecting the necessity or importance of retaining a data frame. It can be discretely categorized into, for example, three levels (high, medium, and low) or a continuous numerical value. The retention strength level of each data frame is compared with a preset retention strength threshold. Data frames with evaluation results below the threshold are compressed, while those above or equal to the threshold are retained. Compression involves compressing and simplifying the content of the data frame to reduce its space usage while preserving as much information as possible that represents the key state of the data. The compression method can be determined in various ways depending on the data type and application requirements. For example, it may involve removing redundant or infrequently changing fields and retaining only the key features of the data frame, or converting the original detailed data into summary information. Through compression, the cache space occupied by the data frame is significantly reduced. The compression preservation mechanism reorders the compressed data and the uncompressed retained data frames according to time order and re-adapts them to fill the remaining space in the cache queue. The cache queue frees up space to accept new data writes while preserving the data characteristics of each time point in the waiting queue as much as possible, avoiding the simple discarding of entire frames due to overflow, thus maintaining the integrity and continuity of the data chain. The packet loss preservation process ensures that even during long waits for some slow sensors, the continuous flow of data from fast sensors does not cause cache crashes. The data is either compressed and preserved, or in extreme cases, it exists in the form of a summary, thus ensuring that the digital twin model can still obtain approximate information from these periods later.

[0037] Specifically, the analysis of sending the integrated synchronization data packet to the data receiving end, and then aligning and injecting data according to the timeline of the drilling digital twin model, is as follows: When a relay node detects that all downhole sensor data for the same time node has arrived, it immediately ends the waiting process for that time node; the relay node extracts the data corresponding to each sensor for that time node from the cache space. For data that arrived early and was temporarily stored, its preprocessed version is used; for data that arrived at the exact moment, its original value or a value that has undergone rapid preprocessing is directly used. These data from different sensors are integrated and packaged into a synchronization data packet. The synchronization data packet gathers the observation values ​​of all sensors at that time node and can be organized according to the required format, such as forming a comprehensive data record with a timestamp, which includes the sensor identifier and its... At that time, the relay node sends the integrated synchronization data packet to the data receiver. For a multi-level relay architecture, after completing the data packet synchronization, the lower-level relay node will pass the synchronization data packet to the upper-level relay or directly to the final data receiver. After receiving the synchronization data packets from each lower-level node or each sensor, the upper-level node can also perform a similar synchronization check and repackage them until all data is finally synchronized and injected at the highest level. Data from all sensors at the same time point arrives at the ground data receiver through layer-by-layer transmission, such as the ground drilling data center or digital twin master server. At the data receiver, according to the time axis used by the drilling digital twin model, the received synchronization data packets are time-aligned and injected, that is, the data is fed into the corresponding time step of the digital twin model according to the timestamp it carries. The digital twin system can update the virtual downhole status in a timely and accurate manner, thereby providing reliable support for monitoring and optimization decisions. It realizes the synchronization and efficient transmission of downhole multi-sensor data driven by digital twin. Under the premise of ensuring that key data is not missed or out of sync, it effectively coordinates the asynchronous data streams of each sensor into a synchronized data frame sequence and inputs it into the digital twin model, which significantly improves the real-time performance and reliability of the digital twin in the drilling process.

[0038] Please see Figure 2The drilling process optimization system based on digital twins, applying the aforementioned drilling process optimization method based on digital twins, includes: an infinite data chain generation module, used to establish a unified drilling data timeline, forming an infinite data chain corresponding to each sensor, wherein the infinite data chain is composed of multiple time nodes connected sequentially, and the infinite data chains of each sensor are connected according to the same time nodes; a hierarchical transmission link formation module, used to deploy multi-level data transmission relay nodes on the path from each sensor to the data receiver, configuring differentiated relay buffer spaces for different sensors to form a hierarchical transmission link; and a freshness buffer module, used to store early-arriving sensor data at each transmission relay node. The system performs preprocessing and caching, while simultaneously monitoring the arrival of data at the same time point on other sensor data chains. For data not yet fully received, it implements a preservation waiting period. A dynamic buffer module categorizes sensor data received during the preservation waiting period into a secondary cache queue based on time point, and compresses and preserves overflowing data based on cache availability and waiting time. A synchronous packet transmission module extracts and integrates the data from each sensor at that time point from the cache space when all downhole sensor data at the same time point arrives at the transmission relay node. This integrated synchronous data packet is then sent to the data receiving end for alignment and data injection according to the timeline of the drilling digital twin model.

[0039] In summary, this application has at least the following effects:

[0040] A unified timeline ensures consistent data time references, laying the foundation for accurate mapping of digital twin models; multi-level relays and differentiated caching improve data transmission efficiency and adapt to different sensor characteristics; data preprocessing and preservation ensure data integrity at the same time point; fuzzy inference compresses overflowing data, balancing data storage and transmission needs within limited cache space; ultimately, efficient data synchronization and integration are achieved, providing high-quality data support for the digital twin model of the drilling process, helping to optimize decision-making, and improving drilling efficiency and safety.

[0041] Those skilled in the art will understand that embodiments of the present invention can be provided as methods or systems. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0042] This invention is described with reference to flowchart illustrations and structural diagrams of methods and systems according to embodiments of the invention. It should be understood that the combination of each process and module in the flowchart and structural diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the process. Figure 1 One or more processes and structures Figure 1 A device for a function specified in one or more modules.

[0043] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and structures Figure 1 The function specified in one or more modules.

[0044] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and structures Figure 1 The steps of a specified function in one or more modules.

[0045] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0046] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A drilling process optimization method based on digital twins, characterized in that, Includes the following steps: Establish a unified drilling data timeline to form an infinite data chain corresponding to each sensor. The infinite data chain is composed of multiple time nodes connected sequentially. The infinite data chains of each sensor are connected according to the same time nodes. Multi-level data transmission relay nodes are deployed along the path from each sensor to the data receiver, and differentiated relay buffer spaces are configured for different sensors to form a hierarchical transmission link. At each transmission relay node, preprocessing and caching are performed on the sensor data that arrives early, while monitoring the arrival of data at the same time node on other sensor infinite data chains, and keeping the data fresh for cases where all data has not arrived. For sensor data received at future time points during the preservation waiting period, the data is categorized according to time point and stored in a secondary cache queue. The overflow data is then compressed and preserved based on the remaining cache space and the waiting time. The specific analysis of performing compression and preservation of overflow data based on cache remaining capacity and waiting time is as follows: During the preservation waiting period, the data generated and arriving at future time nodes by the sensor are sequentially categorized and stored in a preset secondary cache queue according to the corresponding time nodes. When the cache queue's occupied space reaches a preset space threshold, and the cached data's waiting time exceeds a preset duration threshold, a compression and preservation mechanism is activated to perform fuzzy inference filtering on the overflowing data. The compression and preservation mechanism includes the following steps: Collect the buffer wait time and current buffer space remaining for each data frame; Based on cache wait time and cache space remaining as input factors, a fuzzy rule base is constructed, and a fuzzy inference engine is used to perform fuzzy judgment on each data frame to be processed, and output the retention strength level of the data frame. The fuzzy inference output is compared with a set retention strength threshold. For retention strength levels lower than the retention strength threshold, a compression operation is performed. The compression operation includes removing redundant fields, retaining only key features, or storing representative information in summary form. For data frames with a retention strength level greater than or equal to the retention strength threshold, reorder them and fit them into the remaining buffer; When all downhole sensor data at the same time point are detected to have arrived at the transmission relay node, the data of each sensor at that time point is extracted from the buffer space and integrated. The integrated synchronization data packet is then sent to the data receiving end, and then aligned and injected according to the time axis of the drilling digital twin model.

2. The drilling process optimization method based on digital twins according to claim 1, characterized in that, The steps for establishing a unified drilling data timeline are as follows: A timestamp is added to the data collected by each downhole sensor. Data from different sensors are mapped to the same time node sequence according to the timestamps. The corresponding data index of each sensor is marked to achieve synchronous alignment of data from different sensors under the same time reference, forming a unified drilling data timeline.

3. The drilling process optimization method based on digital twins according to claim 1, characterized in that, The specific analysis of the connection of the infinite data chains of each sensor according to the same time node is as follows: For the raw data collected in real time by each downhole sensor, the data is divided into periods based on a preset time period. Each time period is a fixed time window. All the collected data within the time window is aggregated to form a one-to-one virtual data unit. The virtual data unit serves as a time node in an infinite data chain. The time window is set uniformly across all sensors, so that the virtual data units formed by all sensors within the same time period have the same time identifier on the time axis. Each sensor forms an infinite data chain consisting of multiple time-consistent virtual data units connected sequentially. The corresponding nodes between each infinite data chain have a consistent time identifier, realizing a one-to-one correspondence between data from different sensors in logical time. The corresponding connection enables the synchronous extraction and transmission of time nodes with the same time identifier from an infinite data chain of multiple sensors at any given time point.

4. The drilling process optimization method based on digital twins according to claim 1, characterized in that, The multi-level data transmission relay nodes include at least two levels of relay devices arranged sequentially. The analysis of configuring differentiated relay buffer space for different sensors is as follows: Based on fuzzy reasoning using acquisition frequency and transmission delay, downhole sensors are classified into three types: Type I, Type II, and Type III sensors. The type of relay cache space is determined based on the size of the relay cache space, including the first type of relay cache space, the second type of relay cache space, and the third type of relay cache space; Allocate first-type relay buffer space for first-type sensors, second-type relay buffer space for second-type sensors, and third-type relay buffer space for third-type sensors.

5. The drilling process optimization method based on digital twins according to claim 1, characterized in that, The preprocessing and caching specifically involve: performing preliminary feature extraction on the sensor data that arrives in advance at each transmission relay node, and temporarily storing the data after preliminary feature extraction in the transmission relay node cache; The preservation waiting period is specifically achieved by refreshing or monitoring the heartbeat when a preset refresh cycle is reached.

6. The drilling process optimization method based on digital twins according to claim 1, characterized in that, The secondary cache queue is temporarily stored using a circular buffer or a sliding window.

7. A drilling process optimization system based on digital twins, employing the drilling process optimization method based on digital twins as described in any one of claims 1-6, characterized in that, include: The infinite data chain generation module is used to establish a unified drilling data timeline and form an infinite data chain corresponding to each sensor. The infinite data chain is composed of multiple time nodes connected in sequence, and the infinite data chains of each sensor are connected in accordance with the same time nodes. The hierarchical transmission link forming module is used to deploy multi-level data transmission relay nodes on the path from each sensor to the data receiver, configure differentiated relay buffer space for different sensors, and form a hierarchical transmission link. The freshness-preserving and caching module is used to preprocess and cache sensor data that arrives early at each transmission relay node, while monitoring the arrival status of data at the same time node on other sensor infinite data chains, and preserving and waiting for data that has not arrived in full. The dynamic buffer module is used to classify and store sensor data received at future time points during the preservation waiting period into a secondary buffer queue according to the time point, and to compress and preserve the overflow data based on the buffer remaining amount and the waiting time. The synchronous packet transmission module is used to extract and integrate the data of each sensor at that time point from the buffer space when all downhole sensor data at the same time point are detected to have arrived at the transmission relay node. The integrated synchronous data packet is then sent to the data receiving end for alignment and data injection according to the time axis of the drilling digital twin model.

Images