A method and system for real-time preprocessing of neutron detection data based on edge computing

Abstract

The present application relates to the technical field of data transmission control, and particularly relates to a neutron detection data real-time preprocessing method and system based on edge computing. The method comprises the following steps: obtaining original energy spectrum data packets through a high-speed acquisition circuit of downhole equipment and temporarily storing the original energy spectrum data packets into a first annular buffer area; calculating a geological disturbance index of a current frame of data according to a count difference of the current frame and a previous frame in a characteristic energy window and full spectrum information entropy of the current frame; reading an occupancy rate of a communication sending cache queue in real time, combining the geological disturbance index, and calculating a dynamic transmission admission probability of the current frame of data; and dynamically selecting a compression strategy and a transmission priority of the current frame of data according to a comparison result of the dynamic transmission admission probability and a preset threshold. The present application solves the problem of key geological information loss caused by limited channel bandwidth in deep well logging, realizes preferential and faithful transmission of high-value data through a coupling mechanism of geology and a channel, and eliminates the risk of cache overflow.

Description

Technical Field

[0001] This invention relates to the field of data transmission control technology. More specifically, this invention relates to a real-time preprocessing method and system for neutron detection data based on edge computing. Background Technology

[0002] Pulsed neutron logging tools are key high-end equipment in oil exploration for evaluating remaining oil saturation and identifying complex lithologies. These instruments are typically equipped with a neutron generator and multiple scintillator detectors, and during operation, they need to simultaneously acquire high-resolution gamma-ray spectra and nanosecond-level time spectra. In deep well logging operations, data transmission faces severe physical bottlenecks. Downhole instruments communicate with the surface system via single-core armored cables that are thousands of meters long. Due to limitations in channel bandwidth and signal-to-noise ratio, the remote transmission rate is usually strictly limited, making it difficult to meet the real-time upload requirements of massive amounts of data.

[0003] Existing technologies typically employ fixed-time-slot transmission protocols, meaning the same transmission bandwidth is allocated regardless of whether the instrument traverses homogeneous mudstone layers or complex oil and gas formations. This average allocation method results in a significant waste of bandwidth in non-target layers during long well runs, while in critical thin oil or gas layers, the bandwidth lock prevents the increase of sampling density, necessitating substantial dimensionality reduction or thinning of the data. This leads to the smoothing or loss of crucial geological details, failing to meet the demands of high-precision exploration.

[0004] Furthermore, the generation rate of downhole data often exceeds the transmission rate, relying on buffering. Existing flow control mechanisms are passive and cannot dynamically adjust the data compression ratio based on the current buffer capacity. In the event of a sudden high count rate formation, the buffer is prone to overflow, causing unpredictable data frame loss and severely impacting the integrity and accuracy of logging data. Therefore, a method that can dynamically adjust the transmission strategy based on geological value and channel conditions is urgently needed. Summary of the Invention

[0005] The purpose of this invention is to propose a real-time preprocessing method and system for neutron detection data based on edge computing, in order to solve the problems of key information loss caused by fixed bandwidth allocation and data overflow caused by passive caching mechanisms in the prior art; to this end, this invention provides solutions in the following two aspects.

[0006] In a first aspect, the present invention provides a real-time preprocessing method for neutron detection data based on edge computing, comprising:

[0007] The downhole main control unit receives pulse signals from the detector in parallel, converts them into discrete energy spectrum data, generates the raw energy spectrum data packet for the current moment, and temporarily stores it. Based on the preset characteristic energy window weights, the count changes within the characteristic energy window, and the full-spectrum information entropy of the current frame, the geological disturbance index of the current frame data is calculated to assess the severity of formation changes. The occupancy rate of the communication transmission buffer queue is monitored in real time, and the dynamic transmission admission probability of the current frame data is calculated based on the geological disturbance index and the occupancy rate to assess the carrying capacity of the current communication link. The dynamic transmission admission probability is compared with a preset threshold, and based on the comparison result, the compression encoding method and transmission queue priority of the current frame data are dynamically determined, and data transmission is executed.

[0008] In this way, by cleaning and evaluating the value of raw data at the edge, the geological importance of the data can be identified before it enters the transmission queue, providing a basis for decision-making for subsequent differentiated transmission and avoiding invalid data from occupying valuable cache resources.

[0009] Preferably, the expression for the geological disturbance index of the current frame data is:

[0010]

[0011] In the formula, This is the geological disturbance index. For the first Geological weight coefficients of each feature window. For the current moment In the The gamma count rate within each feature window For the previous moment at the first The gamma count rate within each feature window It is a smoothing constant. The information entropy of the full spectrum of the current frame. This is the theoretical maximum entropy value.

[0012] Thus, by combining the differential variation of the characteristic energy window with the full-spectrum information entropy, we can keenly capture the abrupt change interface of the strata. In particular, the introduction of the square root term in the denominator simulates the standard deviation characteristics of the Poisson distribution, making the calculation results statistically significant. Furthermore, by using information entropy as a gain factor, we can ensure that complex strata are marked as high-value, thereby scientifically assessing the degree of geological disturbance in the data.

[0013] Preferably, the information entropy of the current frame's full spectrum is calculated as follows:

[0014]

[0015] In the formula, The information entropy of the full spectrum of the current frame. It is the first The proportion of the count of each energy channel to the total count; the theoretical maximum entropy value. For a 256-channel energy spectrum, the constant is given. .

[0016] Preferably, the geological weight coefficient is preset according to the importance of each element to oil and gas evaluation, and the smoothing constant is set to 1.0 to simulate the characteristics of Poisson distribution and prevent the denominator from being zero.

[0017] Preferably, the expression for the dynamic transmission admission probability of the current frame data is:

[0018]

[0019] In the formula, To dynamically transmit admission probabilities, The occupancy rate of the current communication send buffer queue. This is the geological disturbance index. Let be the steepness factor of the decision function. This is the coefficient by which geological value offsets the pressure on the cache.

[0020] Thus, by introducing cache occupancy rate as a negative feedback adjustment variable and weighting it in conjunction with geological value, the system can automatically suppress the transmission of low-value data when the cache is congested, while forcibly ensuring the access of high-value data, thereby achieving intelligent flow control.

[0021] Preferably, the step of dynamically deciding the compression encoding method and transmission queue priority of the current frame data based on the comparison result includes:

[0022] If the dynamic transmission admission probability is greater than a preset threshold, it is determined to enter the high-fidelity mode. The energy spectrum data is subjected to a lossless compression algorithm to retain all energy channel data and place it into a high-priority queue.

[0023] If the dynamic transmission admission probability is less than or equal to a preset threshold, it is determined that the feature mode is entered, lossy compression is performed on the energy spectrum data, the principal component analysis coefficients and feature window count rates of the energy spectrum are extracted, and they are placed into the normal queue.

[0024] In this way, a hierarchical transmission strategy is adopted to achieve fine-grained allocation of bandwidth resources, significantly saving bandwidth at non-destination layers and releasing bandwidth in a concentrated manner at the destination layer, thereby achieving high-precision evaluation of critical layers under limited transmission rates.

[0025] Preferably, the lossy compression in the feature mode includes: performing principal component analysis on the current frame energy spectrum data, extracting and transmitting the first 5 principal component analysis coefficients of the energy spectrum, and the count rates of 5 feature energy windows of carbon, oxygen, silicon, calcium and iron.

[0026] Preferably, the feature window includes address ranges corresponding to the elements carbon, oxygen, silicon, calcium, and iron.

[0027] Preferably, the data transmission includes: the remote transmission interface reading data according to the principle of high-priority queue first dequeue; when in high-fidelity mode, inserting a keyframe flag bit in the data frame header, and encapsulating and sending the data using LZW or Huffman encoding.

[0028] In the second aspect, a real-time preprocessing system for neutron detection data based on edge computing includes:

[0029] processor;

[0030] The memory stores computer instructions for real-time preprocessing of neutron detection data based on edge computing. When the computer instructions are executed by the processor, the system performs the aforementioned real-time preprocessing method for neutron detection data based on edge computing.

[0031] The beneficial effects of this invention are as follows: This invention breaks away from the traditional model of evenly allocating bandwidth in logging instruments, and introduces a dynamic transmission mechanism coupled with geological channels. By calculating the geological disturbance index in real time at the downhole edge, it accurately identifies high-value data segments such as oil and gas layers, and dynamically adjusts the data compression strategy in conjunction with the real-time pressure of the communication buffer. In the mudstone section, which occupies most of the well section, data is significantly compressed through feature patterns to save bandwidth; while in the key oil and gas layer sections, the saved bandwidth is used to achieve full-spectrum lossless transmission, which significantly improves the actual data density of the target layer and enables detailed detection of oil.

[0032] Furthermore, this invention introduces a negative feedback adjustment mechanism based on cache occupancy rate. When communication links fluctuate or data backlog occurs, the algorithm automatically raises the entry threshold, prioritizing the discarding or compression of low-value data. This mechanism fundamentally eliminates the risk of cache overflow, ensuring that high-value geological data is never lost due to insufficient cache, thus guaranteeing the integrity of core information in well logging data. Simultaneously, all calculations are performed within the downhole FPGA, achieving millisecond-level mode switching response, avoiding the long delay issues of surface command control, and ensuring that no thin oil layers are missed. Attached Figure Description

[0033] Figure 1 The flowchart illustrating the steps of the real-time preprocessing method for neutron detection data based on edge computing in this invention is shown in the schematic diagram.

[0034] Figure 2 The diagram illustrates the real-time response curve of dynamic bandwidth allocation based on geological significance in this invention.

[0035] Figure 3The diagram illustrates a comparison of energy spectrum reconstruction quality under bandwidth-limited conditions in this invention. Detailed Implementation

[0036] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0037] like Figure 1 As shown in this embodiment, a real-time preprocessing method for neutron detection data based on edge computing includes the following steps:

[0038] Step S1: The downhole main control unit receives the pulse signal from the detector in parallel, converts it into discrete energy spectrum data, generates the original energy spectrum data packet for the current moment, and temporarily stores it.

[0039] Specifically, the system first acquires raw data through the high-speed acquisition circuit of the downhole instrument, and the FPGA main control unit receives pulse signals from three detectors—near, far, and ultra-far—in parallel. The internal multichannel pulse amplitude analysis module (MCA) converts the continuous analog pulses into discrete energy spectrum data.

[0040] Further, a raw energy spectrum data packet for the current moment is generated, containing count values ​​for 256 energy channels. The generated raw data packet is temporarily stored in a primary circular buffer, awaiting processing by the edge computing module. At this point, the data has not yet entered the transmission queue and is in a state of pending cleaning.

[0041] In this way, by initially acquiring and caching the raw data at the edge, a data foundation is provided for subsequent real-time computing, ensuring the continuity and real-time nature of data processing.

[0042] Step S2: Calculate the geological disturbance index of the current frame data based on the preset feature window weights, the count changes within the feature window, and the full-spectrum information entropy of the current frame, in order to assess the severity of stratigraphic changes.

[0043] Specifically, when evaluating the value of the current frame data at the edge, changes in geological strata inevitably cause abrupt changes in the energy spectrum morphology and an increase in information content. To quantify this change, a composite index, namely the geological disturbance index, is constructed, which combines the severity of feature window changes with the full-spectrum information entropy.

[0044] The expression for the geological disturbance index of the current frame data is:

[0045]

[0046] In the formula, This is the geological disturbance index. For the first The geological weight coefficients of each feature window are pre-stored in the downhole ROM. For the current moment In the The gamma count rate within each feature window For the previous moment at the first The gamma count rate within each feature window It is a smoothing constant. The information entropy of the full spectrum of the current frame. This represents the theoretical maximum entropy value. For example, the smoothing constant is set to 1.0 to prevent the denominator from being zero and to simulate the characteristics of a Poisson distribution.

[0047] The information entropy of the current frame's full spectrum is calculated as follows:

[0048]

[0049] In the formula, The information entropy of the full spectrum of the current frame. It is the first The proportion of the count of each energy channel to the total count; the theoretical maximum entropy value. For a 256-channel energy spectrum, the constant is given. .

[0050] In this way, this step can keenly capture abrupt changes in strata through weighted difference. The square root term in the denominator introduces the concept of standard deviation in statistics, which enables the index to distinguish between real count changes and statistical fluctuations. It also uses a logarithmic function to compress the numerical range and uses information entropy as a gain factor to ensure that complex strata are marked as high value, thereby accurately assessing the geological importance of the data.

[0051] Step S3: Monitor the occupancy rate of the communication transmission buffer queue in real time, and calculate the dynamic transmission admission probability of the current frame data based on the geological disturbance index and the occupancy rate to assess the carrying capacity of the current communication link.

[0052] Specifically, after calculating the geological value, the system must determine whether the current communication link is congested. This step calculates the dynamic transmission admission probability to determine whether the current frame data should be sent in full or compressed.

[0053] The expression for the dynamic transmission admission probability of the current frame data is:

[0054]

[0055] In the formula, To dynamically transmit admission probabilities, The occupancy rate of the current communication send buffer queue. This is the geological disturbance index. Let be the steepness factor of the decision function. This is the offsetting factor of geological value on cache pressure. In this embodiment, the occupancy rate of the current communication sending cache queue ranges from 0 to 1, and the steepness factor of the decision function is 10.

[0056] For example, assume the current cache occupancy is 0.8 (i.e., 80% full) and the offset factor is 0.1.

[0057] If the geological disturbance index is low (e.g., 1.0, corresponding to mudstone), the dynamic transmission admission probability of the current frame data can be calculated to be 0.0009, which is extremely low, and the system tends to compress the data.

[0058] If the geological disturbance index is high (e.g., 10.0, corresponding to an oil layer), the dynamic transmission admission probability of the current frame data can be calculated to be 0.88, which is a high probability, and the system tends to transmit with high fidelity.

[0059] Thus, when the cache is idle, the probability approaches 1, and the system tends to transmit high-quality data; when the cache is congested, if it is mudstone, the probability approaches 0, and the transmission is automatically downgraded; if it is oil layer, the geological value will offset the impact of cache pressure, forcibly increasing the probability and ensuring that critical data squeezes into the transmission queue, thus realizing intelligent flow control management.

[0060] Step S4: Compare the dynamic transmission admission probability with a preset threshold, dynamically decide the compression encoding method and transmission queue priority of the current frame data based on the comparison result, and execute data transmission.

[0061] Specifically, based on the calculated dynamic transmission admission probability value, the FPGA executes a hierarchical processing strategy and pre-sets a threshold.

[0062] If the dynamic transmission admission probability is greater than the preset threshold, it is determined to enter the high-fidelity mode. This usually occurs when there are key geological strata or sufficient buffer. At this time, lossless compression algorithm (such as LZW or Huffman coding) is used on the energy spectrum data to retain all 256 channels of data. At the same time, it is placed in a high-priority queue and a key frame flag is inserted at the beginning.

[0063] If the dynamic transmission admission probability is less than or equal to a preset threshold, it is determined that the system enters the characteristic mode. This usually occurs when the non-target layer is not in use and the buffer is tight. In this case, a significant lossy compression is performed on the energy spectrum data. Specifically, only the first five principal component analysis (PCA) coefficients of the energy spectrum and the count rates of the five feature windows of carbon, oxygen, silicon, calcium, and iron are extracted and transmitted. The data volume can be compressed to about 5% of the original size, and then placed into a normal queue.

[0064] The remote transmission interface modulates the data onto the cable and transmits it to the ground according to the principle of high-priority queue first.

[0065] Reference Figure 2 This demonstrates how the system of this invention dynamically adjusts the transmission bandwidth according to geological conditions. As shown in the figure, the geological disturbance index peaks near well depths of 2010 meters and 2035 meters, corresponding to oil and gas reservoirs. The actual allocated bandwidth fluctuations are highly synchronized with the geological disturbance index curve, automatically decreasing in mudstone sections and instantly surging in oil and gas reservoirs; whereas the fixed bandwidth mode of existing technologies cannot respond to geological changes.

[0066] Reference Figure 3 A key oil-bearing layer (ROI region) was selected for comparison of the energy dispersive spectral reconstruction quality. As can be seen from the figure, the existing technology uses significant dimensionality reduction due to bandwidth limitations, resulting in waveform distortion and loss of characteristic peaks; while in this invention, due to the dynamic allocation of high bandwidth and lossless compression, the curve almost perfectly overlaps with the original standard spectrum, and the characteristic peaks are clear.

[0067] Thus, this invention breaks the traditional pattern of evenly allocating bandwidth, saving a large amount of bandwidth in the mudstone section that occupies most of the well section, and concentrating it on the oil and gas layer section, which greatly increases the actual data density of the target layer. Combined with the cache negative feedback mechanism, it ensures the integrity of high-value data and achieves millisecond-level edge response.

[0068] This invention also provides a real-time preprocessing system for neutron detection data based on edge computing. The system includes a processor and a memory, the memory storing computer program instructions. When the processor executes the computer program instructions, it implements the real-time preprocessing method for neutron detection data based on edge computing described above according to this invention.

[0069] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and therefore will not be described in detail here.

[0070] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented by computer-readable / executable instructions stored or otherwise maintained on such a computer-readable medium.

[0071] In the description of this specification, "multiple" means at least two, such as two, three or more, etc., unless otherwise expressly and specifically defined.

[0072] While various embodiments of the invention have been shown and described in this specification, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention.

Claims

1. A method for real-time preprocessing of neutron detection data based on edge computing, characterized in that, include: The downhole main control unit receives the pulse signals from the detector in parallel, converts them into discrete energy spectrum data, generates the raw energy spectrum data packet for the current moment, and temporarily stores it. The geological disturbance index of the current frame data is calculated based on the preset feature window weights, the count changes within the feature window, and the full-spectrum information entropy of the current frame, in order to assess the degree of drastic changes in the strata. The expression for the geological disturbance index of the current frame data is: , This is the geological disturbance index. For the first Geological weight coefficients of each feature window. For the current moment In the The gamma count rate within each feature window For the previous moment at the first The gamma count rate within each feature window It is a smoothing constant. The information entropy of the full spectrum of the current frame, This is the theoretical maximum entropy value; The occupancy rate of the communication transmission buffer queue is monitored in real time, and the dynamic transmission admission probability of the current frame data is calculated based on the geological disturbance index and the occupancy rate to assess the carrying capacity of the current communication link. The dynamic transmission admission probability is compared with a preset threshold. Based on the comparison result, the compression encoding method and transmission queue priority of the current frame data are dynamically determined, and data transmission is executed.

2. The real-time preprocessing method for neutron detection data based on edge computing according to claim 1, characterized in that, The information entropy of the current frame's full spectrum is calculated as follows: In the formula, The information entropy of the full spectrum of the current frame, It is the first The proportion of the count of each energy channel to the total count; the theoretical maximum entropy value. For a 256-channel energy spectrum, the constant is given. .

3. The real-time preprocessing method for neutron detection data based on edge computing according to claim 1, characterized in that, The geological weight coefficient is preset according to the importance of each element to oil and gas evaluation, and the smoothing constant is set to 1.0 to simulate the characteristics of Poisson distribution and prevent the denominator from being zero.

4. The real-time preprocessing method for neutron detection data based on edge computing according to claim 1, characterized in that, The expression for the dynamic transmission admission probability of the current frame data is: In the formula, To dynamically transmit admission probabilities, The occupancy rate of the current communication send buffer queue. This is the geological disturbance index. Let be the steepness factor of the decision function. This is the coefficient by which geological value offsets the pressure on the cache.

5. The real-time preprocessing method for neutron detection data based on edge computing according to claim 1, characterized in that, The step of dynamically deciding the compression encoding method and transmission queue priority of the current frame data based on the comparison results includes: If the dynamic transmission admission probability is greater than a preset threshold, it is determined to enter the high-fidelity mode. The energy spectrum data is subjected to a lossless compression algorithm to retain all energy channel data and place it into a high-priority queue. If the dynamic transmission admission probability is less than or equal to a preset threshold, it is determined that the feature mode is entered, lossy compression is performed on the energy spectrum data, the principal component analysis coefficients and feature window count rates of the energy spectrum are extracted, and they are placed into the ordinary queue.

6. The real-time preprocessing method for neutron detection data based on edge computing according to claim 5, characterized in that, The lossy compression in the feature mode includes: performing principal component analysis on the current frame energy spectrum data, extracting and transmitting the first 5 principal component analysis coefficients of the energy spectrum, and the count rates of the 5 feature energy windows of carbon, oxygen, silicon, calcium and iron.

7. The real-time preprocessing method for neutron detection data based on edge computing according to claim 1, characterized in that, The feature windows include address ranges corresponding to the elements carbon, oxygen, silicon, calcium, and iron.

8. The real-time preprocessing method for neutron detection data based on edge computing according to claim 1, characterized in that, The execution data transmission includes: The remote transmission interface reads data according to the principle of dequeuing the highest priority queue first; When in high-fidelity mode, a keyframe flag is inserted into the header of the data frame, and LZW or Huffman encoding is used for encapsulation and transmission.

9. A real-time preprocessing system for neutron detection data based on edge computing, characterized in that, include: processor; A memory storing computer instructions for real-time preprocessing of neutron detection data based on edge computing, wherein when the computer instructions are executed by the processor, the system performs the real-time preprocessing method for neutron detection data based on edge computing according to any one of claims 1-8.

Images