A method and device for estimating the channel response of mud pulse telemetry
By constructing a local observation model for the LFM probe segment and a local channel estimation model for the pilot segment, and combining sparse regularization and local Markov support transfer, the problem of unstable channel response under short pilot and strong interference conditions was solved, and stable and accurate estimation of the mud pulse telemetry channel was achieved.
Patent Information
- Application Number
- CN202611424321.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-15
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2046-09-15
AI Technical Summary
Under conditions of short pilot signals and strong interference, the channel response estimation of mud pulse telemetry is unstable, and existing technologies are unable to achieve accurate channel estimation over the entire time delay range.
By constructing a local observation model of the LFM probe segment to estimate the sparse regularized channel response, the LFM coarse delay support set is determined, and the prior support score of each delay tap in the pilot band is calculated using the local Markov support transition matrix. A local channel estimation model for the pilot band is constructed, and the channel response is estimated using the regularized least squares method.
It improves the stability and accuracy of channel estimation under short pilot and strong interference conditions, solves the problem of unstable channel response over the entire time delay range, and realizes stable and accurate estimation of mud pulse telemetry channel.
Smart Images

Figure CN122951906B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of oil drilling logging mud pulse telemetry signal processing technology, and in particular to a mud pulse telemetry channel response estimation method and device. Background Technology
[0002] Mud pulse telemetry utilizes pressure waves in drilling fluid to transmit downhole measurement data to the surface, and is widely used in measurement-while-drilling (MSW) and logging-while-drilling (LOD) applications. A wellhead pressure sensor acquires the wellhead pressure signal and outputs a digital pressure signal. The surface receiving equipment performs frame synchronization, channel estimation, equalization, and data demodulation on this digital pressure signal. As the pressure wave propagates along the drill string and drilling fluid column towards the surface, it is reflected and coupled by propagation boundaries such as drill pipe joints, pipe diameter changes, downhole tool structures, and surface manifolds, forming multiple propagation components with different time delays and attenuation coefficients. Therefore, the mud pulse telemetry channel exhibits significant multipath and long-memory characteristics; the current sampled value in the received digital pressure signal is formed by the superposition of multiple propagation components with different time delays.
[0003] The mud pulse telemetry channel response consists of a set of time delay taps and their corresponding channel coefficients. The time delay taps represent the discrete time delay positions corresponding to the propagating components, the channel coefficients represent the amplitude and phase of the corresponding propagating components, and the channel support represents the set of time delay positions where the effective non-zero taps in the channel response are located. Only by accurately determining the channel support and the channel coefficients corresponding to each effective tap can the ground receiving equipment equalize the inter-symbol interference caused by multipath propagation and complete subsequent data demodulation.
[0004] Existing methods for estimating channel response in mud pulse telemetry typically utilize known training sequences in pilot bands to construct pilot training convolution matrices and solve for the channel response based on the correspondence between the transmitted pilot signal and the received pilot observation signal. Since the pilot sequence is a known, deterministic sequence for the ground receiving equipment, pilot bands can accurately estimate the channel coefficients corresponding to candidate delay taps. Compared to relying solely on synchronization signals for delay localization, pilot bands can obtain more accurate channel tap positions and channel coefficients.
[0005] However, the number of observation samples in the pilot band is limited by the communication frame length and effective data transmission efficiency. When the number of pilot samples is less than or close to the number of delay taps to be estimated, directly estimating the channel response within the entire delay tap range from zero delay to the maximum preset delay will result in an excessively high dimensionality of the pilot training convolution matrix. Under this condition, periodic harmonics from the mud pump, random noise, and unmodeled disturbances will enter the channel response solution process and be misjudged as valid channel taps, causing false tap activation, main propagation component position shift, and channel coefficient estimation errors. Therefore, although the pilot band has high channel response estimation accuracy, its estimation stability is directly affected by the number of pilot samples and received interference.
[0006] The LFM detection segment exhibits stable matched correlation characteristics. Mud pump noise primarily manifests as a narrowband periodic component composed of the fundamental frequency and its integer harmonics, whose frequency structure differs from the instantaneous frequency structure of the LFM signal over time. Therefore, LFM template matched correlation can utilize this structural difference to stably detect the communication frame position under strong interference conditions from mud pump harmonics and random noise, and stably obtain the coarse delay support corresponding to the channel propagation components.
[0007] However, the main lobe of the LFM matched correlation output has a finite width, and the correlation responses of adjacent multipath components can overlap. The discrete sampling interval, the width of the matched correlation main lobe, and the superposition of multiple propagation components collectively limit the delay localization accuracy of LFM. Therefore, the LFM probe segment obtains a coarse-grained location of the delay region where the channel propagation components are located, which is suitable for determining coarse delay supports, but cannot directly replace the pilot segment to complete accurate channel response estimation.
[0008] Therefore, the LFM probe segment and the pilot segment have different characteristics in the channel estimation process: the LFM probe segment relies on a wide time-frequency structure and matched correlation to obtain stable coarse delay support, but its delay positioning accuracy is limited; the pilot segment relies on known training sequences to obtain accurate channel tap coefficients, but its estimation results are easily affected by mud pump harmonics, random noise, and short pilot sample conditions. Using the LFM probe segment alone cannot obtain an accurate channel response, while using the pilot segment alone for estimation across the entire delay tap range is prone to generating spurious channel taps.
[0009] Furthermore, the LFM probe segment, guard interval, and pilot segment are arranged sequentially along the time direction of the communication frame, with the LFM probe segment and pilot segment corresponding to different time positions within the same communication frame. During the time interval between the LFM probe segment and the pilot segment, changes in propagation conditions, equivalent processing delay, and sampling position can cause finite delay tap position changes in the channel support. Therefore, the coarse delay support obtained from the LFM probe segment cannot be directly copied to the pilot segment according to the original tap positions. Direct copying would miss effective channel taps with finite drift; completely abandoning the LFM coarse delay support would still require estimation over the entire delay tap range of the pilot segment, and the stable delay prior provided by the LFM probe segment cannot be utilized.
[0010] Therefore, a method for estimating the channel response of mud pulse telemetry is needed to solve the problem of unstable channel estimation over the entire time delay range under short pilot and strong interference conditions. Summary of the Invention
[0011] The purpose of this application is to provide a mud pulse telemetry channel response estimation method, which can improve the problem of unstable channel estimation over the entire time delay range under short pilot and strong interference conditions.
[0012] To achieve the above objectives, this application provides the following solution.
[0013] In a first aspect, this application provides a method for estimating the channel response of mud pulse telemetry, including: Based on the pressure digital signal, a preset standard LFM linear frequency modulation template, and preset communication frame structure parameters, the LFM probe segment received observation vector and the pilot segment received observation vector are determined. Based on the LFM probe segment received observation vector and the LFM convolutional regression matrix, a local observation model for the LFM probe segment is constructed. Sparse regularized channel response estimation is performed on the local observation model to determine the LFM coarse delay support set and the support quality of each coarse delay support. A local Markov support transition matrix is constructed, and based on the local Markov support transition matrix, the LFM coarse delay support set, and the support quality of each LFM coarse delay support, the prior support score of each delay tap in the pilot segment is calculated. Using the prior support scores of each delay tap in the pilot segment, a local candidate delay set for the pilot segment is determined. Based on the pilot segment received observation signal and the pilot training convolutional matrix, a local channel estimation model for the pilot segment is constructed. Based on the local candidate delay set and the pilot sequence, a regularized least squares method is used to solve the local channel estimation model for the pilot segment to determine the mud pulse telemetry channel response estimate.
[0014] In a second aspect, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the mud pulse telemetry channel response estimation method described in the first aspect.
[0015] Based on the specific embodiments provided in this application, the following technical effects are disclosed.
[0016] This application provides a method for estimating the channel response of mud pulse telemetry. First, based on the pressure digital signal, a preset standard LFM linear frequency modulation template, and preset communication frame structure parameters, the LFM probe segment received observation vector and the pilot segment received observation vector are determined. Then, based on the LFM probe segment received observation vector and the LFM convolutional regression matrix, a local observation model of the LFM probe segment is constructed. Sparse regularized channel response estimation is performed on the local observation model of the LFM probe segment to determine the LFM coarse delay support set and the support quality of each coarse delay support. This allows the coarse delay support obtained from the LFM probe segment to be transferred to the pilot segment based on the finite tap drift relationship between the LFM coarse delay support and the pilot segment channel support. A local Markov support transition matrix is further constructed. Based on this local Markov support transition matrix, the LFM coarse delay support set, and the support quality of each LFM coarse delay support, the prior support score of each delay tap in the pilot band is calculated. Using the prior support scores of each delay tap in the pilot band, a local candidate delay set for the pilot band is determined. Thus, after determining the local candidate delay set based on the prior support scores, the ground receiving equipment only constructs a pilot training convolution matrix within this local candidate delay set and estimates the channel response, thereby reducing the number of delay taps to be estimated and lowering the risk of mud pump harmonics and random noise being misclassified as valid channel taps. Finally, based on the received observation signal and the pilot training convolution matrix in the pilot band, a local channel estimation model for the pilot band is constructed. Based on the local candidate delay set and the pilot sequence, a regularized least squares method is used to solve the local channel estimation model to determine the mud pulse telemetry channel response estimate. Stable and accurate estimation of the mud pulse telemetry channel response is achieved. This application utilizes the LFM coarse delay support obtained stably from the LFM probe segment and transfers it to the pilot segment to form a local candidate delay set for the pilot segment. Then, within the local candidate delay set of the pilot segment, the pilot sequence is used to accurately estimate the mud pulse telemetry channel response. Thus, by combining the advantages of the stability of the LFM probe segment and the accuracy of the pilot segment, the problem of unstable channel estimation across the entire time delay range under short pilot and strong interference conditions is solved. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 A schematic diagram illustrating the specific process of a mud pulse telemetry channel response estimation method provided in this application.
[0019] Figure 2This is a flowchart illustrating the mud pulse telemetry channel response estimation method provided in this application.
[0020] Figure 3 This is a schematic diagram of the local candidate delay set transfer in the guide band for LFM coarse delay support.
[0021] Figure 4 This is a schematic diagram comparing the complete set of time delay taps with the local candidate delay set of the pilot band. Detailed Implementation
[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0023] To make the objectives, features and advantages of this application more apparent and understandable, the application will now be described in further detail with reference to the accompanying drawings and specific embodiments.
[0024] like Figure 1 As shown, this application provides a method for estimating the channel response of mud pulse telemetry, including steps 101-107.
[0025] Step 101: Determine the LFM probe segment receiving observation vector and the pilot band receiving observation vector based on the pressure digital signal, the preset standard LFM linear frequency modulation template, and the preset communication frame structure parameters.
[0026] In some embodiments, the preset communication frame structure includes an LFM probe segment, a guard interval, and a pilot segment arranged sequentially along the time direction of the communication frame; the preset communication frame structure parameters include at least the number of LFM probe segment samples, the number of guard interval samples, the number of pilot segment samples, and the maximum preset delay tap number.
[0027] In some embodiments, step 101 specifically includes: normalizing and matching a preset standard LFM linear frequency modulation template and a pressure digital signal to determine an LFM matching correlation sequence; determining the position corresponding to the maximum value of the LFM matching correlation sequence as the starting sample position of the LFM detection segment; using the starting sample position of the LFM detection segment as the reference time of the communication frame; determining the actual sample index set of the LFM detection segment, the actual sample index set of the protection interval, and the actual sample index set of the pilot segment based on the reference time of the communication frame and preset communication frame structure parameters; determining the received observation vector of the LFM detection segment based on the actual sample index set of the LFM detection segment; and determining the received observation vector of the pilot segment based on the actual sample index set of the pilot segment.
[0028] Step 102: Construct a local observation model for the LFM probe segment based on the received observation vector and LFM convolutional regression matrix.
[0029] In some embodiments, step 102 specifically includes: constructing an LFM convolutional regression matrix based on the preset standard LFM linear frequency modulation template; and establishing a local observation model for the LFM probe segment based on the received observation vector of the LFM probe segment and the LFM convolutional regression matrix, wherein the local observation model for the LFM probe segment is: .
[0030] in, The channel response corresponding to the LFM probe segment. The LFM probe segment observation error is caused by a combination of random noise, mud pump harmonics, and unmodeled disturbances. Receive observation vectors for the LFM probe segment.
[0031] Step 103: Perform sparse regularized channel response estimation on the local observation model of the LFM probe segment to determine the LFM coarse delay support set and the support quality of each coarse delay support.
[0032] In some embodiments, step 103 specifically includes: Sparse regularized channel response estimation is performed on the local observation model of the LFM probe segment to determine the channel response estimation result of the LFM probe segment.
[0033] The LFM coarse delay support set is determined based on the maximum tap amplitude and the preset minimum effective tap amplitude of the LFM probe segment channel response estimation result.
[0034] The support quality of each coarse delay support is calculated based on the channel tap energy corresponding to each coarse delay support in the LFM coarse delay support set.
[0035] Step 104: Construct a local Markov support transition matrix, and calculate the prior support score of each delay tap in the pilot band based on the local Markov support transition matrix, the LFM coarse delay support set, and the support quality of each LFM coarse delay support.
[0036] In some embodiments, step 104 specifically includes: determining the nominal delay tap drift and drift neighborhood radius based on the historical tap position difference between the LFM probe segment channel support and the pilot segment channel support; constructing a local Markov support transition matrix based on the nominal delay tap drift and the drift neighborhood radius; and calculating the prior support score of each delay tap in the pilot segment based on the local Markov support transition matrix, the LFM coarse delay support set, and the support quality of each LFM coarse delay support.
[0037] Step 105: Determine the local candidate delay set of the pilot band by using the prior support score of each delay tap of the pilot band.
[0038] In some embodiments, step 105 specifically includes: Based on the prior support score of each delay tap in the pilot band, the pilot band delay taps with a prior support score greater than zero are determined as the pilot band local candidate delay set.
[0039] Step 106: Construct a pilot band local channel estimation model based on the received observation signal and the pilot training convolution matrix.
[0040] In some embodiments, step 106 specifically includes: constructing a pilot band local training convolution matrix based on the pilot sequence; and constructing a pilot band local channel estimation model based on the pilot band received observation vector and the pilot band local training convolution matrix.
[0041] In some embodiments, the pilot band local channel estimation model is: ;in, This is the channel response coefficient vector within the local candidate delay set of the pilot band; The pilot band observation error is caused by a combination of random noise, mud pump harmonics, and unmodeled disturbances. This is the local training convolution matrix for the pilot band; This is the pilot band receive observation vector.
[0042] Step 107: Based on the local candidate delay set and pilot sequence of the pilot band, the local channel estimation model of the pilot band is solved by regularized least squares method to determine the mud pulse telemetry channel response estimate.
[0043] This application divides the pressure digital signal of the wellhead pressure sensor based on the communication frame timing to determine the processed pressure digital signal. The processed pressure digital signal includes an LFM detection segment, a protection interval, and a pilot segment. The LFM detection segment, protection interval, and pilot segment are arranged sequentially along the time direction of the communication frame, with the LFM detection segment at the beginning, the protection interval between the LFM detection segment and the pilot segment, and the pilot segment after the protection interval. The three frame segments do not overlap in time.
[0044] The preset standard LFM linear frequency modulation template waveform is matched and correlated with the pressure digital signal. The starting sample position of the LFM detection segment and the reference time of the communication frame are determined according to the matching correlation peak. The received observation signals corresponding to the LFM detection segment and the pilot segment are determined according to the reference time of the communication frame.
[0045] Based on the received observation signal corresponding to the LFM probe segment and the preset standard LFM linear frequency modulation template waveform, an LFM local observation model is constructed. Sparse channel response estimation is performed on the LFM local observation model to obtain the LFM probe segment channel response estimation result. Based on the channel response estimation result, the LFM coarse delay support and the support quality of each coarse delay support are determined.
[0046] A local Markov support transition matrix is constructed based on the nominal time delay tap drift and drift neighborhood radius between the LFM probe segment and the pilot segment. Based on the local Markov support transition matrix and the support quality of each coarse delay support, the prior support score of each time delay tap of the pilot is calculated.
[0047] All pilot delay taps with prior support scores greater than zero are identified as the pilot band local candidate delay set, and a pilot training convolution matrix is constructed based on the pilot band local candidate delay set.
[0048] A pilot band local channel estimation model is established based on the pilot band received observation signal and the pilot training convolution matrix. The pilot band local channel estimation model is solved using the regularized least squares method to obtain the mud pulse telemetry channel response. This application can accurately estimate the mud pulse telemetry channel response and solve the problem of unstable channel estimation over the entire time delay range under short pilot and strong interference conditions.
[0049] In practical applications, the mud pulse telemetry channel response estimation method provided in this application is described in detail. This embodiment executes steps S1 to S6 sequentially for the pressure digital signal within the same communication frame. The output of the previous step serves as the input for the next step, ultimately outputting the mud pulse telemetry channel response. Details are as follows.
[0050] Step S1: Acquire the pressure digital signal and preset communication frame structure parameters.
[0051] A wellhead pressure sensor acquires wellhead pressure signals and converts them into discrete digital pressure signals. A surface receiving device then acquires these digital pressure signals, denoted as […]. y [ n ],in, This is the index for discrete sampling points.
[0052] The ground receiving equipment simultaneously acquires preset communication frame structure parameters. The preset communication frame structure, along the time direction, sequentially includes an LFM probe segment, a guard interval, and a pilot segment, with the three segments not overlapping in time. Specifically, the LFM probe segment is used to determine the reference time of the communication frame and provide coarse channel delay support; the guard interval is used to isolate the tail transition response of the LFM probe segment from the pilot segment; and the pilot segment includes a pilot sequence known to the ground receiving equipment, used to estimate the mud pulse telemetry channel response.
[0053] The preset communication frame structure parameters include at least the number of LFM probe segment samples. Number of samples for protection interval Number of pilot band samples and the maximum preset delay number of taps .
[0054] The complete set of delay taps, ranging from zero delay to the maximum preset delay. Represented as: .
[0055] The ground receiving equipment also acquires a preset standard LFM linear frequency modulation template. and known pilot sequences .
[0056] The input to step S1 is the wellhead pressure signal acquired by the wellhead pressure sensor and the preset communication frame structure parameters. The output of step S1 is the pressure digital signal. y [ n Preset standard LFM linear frequency modulation template Given pilot sequence And preset communication frame structure parameters.
[0057] The purpose of step S1 is to provide unified input data and a preset communication frame structure (frame structure) basis for subsequent communication frame synchronization, LFM coarse delay support determination, coarse delay support transfer, and pilot band channel response estimation.
[0058] Step S2: Determine the reference time of the communication frame and the actual sample range of each frame segment; wherein, each frame segment corresponds to the LFM detection segment, the guard interval and the pilot segment.
[0059] The ground receiving equipment will have a pre-set standard LFM linear frequency modulation template. Normalized matched correlation was performed with the pressure digital signal y[n] to obtain the LFM matched correlation sequence: .
[0060] in, The candidate starting sample location, The conjugate of the preset standard LFM linear frequency modulation template, To prevent positive numbers with a denominator of zero; Starting with candidate sample positions Starting from the first One pressure digital signal sample; The normalized LFM matching correlation value at the candidate starting sample location.
[0061] The position corresponding to the maximum value of the LFM-matched correlation sequence is determined as the starting sample position of the LFM probe segment. : .
[0062] In this embodiment, the starting sample position of the LFM detection segment is... It is also determined as the reference time for the communication frame.
[0063] To determine the reliability of LFM matching correlation peaks, the second largest correlation peak is identified outside the neighborhood of the largest correlation peak, and the confidence level of the synchronization peak is calculated. : .
[0064] in, The main lobe neighborhood of the maximum correlation peak; The starting sample position of the detected LFM probe segment The corresponding LFM normalized matching correlation value. When the synchronization peak confidence... Not less than the preset synchronization confidence threshold At that time, the communication frame synchronization is confirmed to be successful; when the synchronization peak confidence level is high... Less than the preset synchronization confidence threshold When this happens, a communication frame synchronization failure flag is output, and subsequent channel estimation for the current communication frame is not performed.
[0065] After successful communication frame synchronization, the actual sample index set of the LFM probe segment is determined based on the communication frame reference time and preset communication frame structure parameters. : .
[0066] Determine the actual sample index set for the protection interval : .
[0067] Determine the actual sample index set of the pilot band. : .
[0068] Based on the actual sample index set of the LFM probe segment, extract the received observation vector of the LFM probe segment. : .
[0069] Based on the actual sample index set of the pilot band, extract the pilot band received observation vector. : .
[0070] The pressure digital signal corresponding to the protection interval is not involved in the determination of LFM coarse delay support and pilot band channel response estimation.
[0071] In this application, to achieve communication frame synchronization, an LFM probe segment is set in the preset communication frame structure. The instantaneous frequency of the preset standard LFM linear frequency modulation template changes continuously with time, and has a wide time-frequency coverage range. After matching and correlating the preset standard LFM linear frequency modulation template with the received pressure digital signal, a clear correlation peak can be formed. The ground receiving equipment determines the starting sample position of the LFM probe segment and the reference time of the communication frame based on the position of the correlation peak, and determines the coarse delay region where the channel propagation component is located based on the main delay peak in the correlation response. The coarse delay range represents the range of delay tap positions that may correspond to the main propagation component of the channel.
[0072] Step S2 uses the pressure digital signal output in step S1, the preset standard LFM linear frequency modulation template, and the preset communication frame structure parameters to obtain the communication frame reference time. LFM detection segment receives observation vector and pilot band received observation vector .
[0073] Step S2 uses the stable matched correlation characteristics of the LFM probe segment to determine the position of the communication frame and matches the preset communication frame structure with the actual sample range in the pressure digital signal, providing a unified time reference for the channel processing of the subsequent two frame segments.
[0074] like Figure 1 As shown, step S2 corresponds to "LFM matching correlation and communication frame reference time determination". Figure 2 As shown, the LFM detection segment, protection interval, and pilot segment are arranged sequentially along the time direction, with the protection interval located between the LFM detection segment and the pilot segment.
[0075] Step S3: Determine the LFM coarse delay support set and the support mass of each coarse delay support.
[0076] The ground receiving equipment constructs an LFM convolutional regression matrix based on different delay copies of a preset standard LFM linear frequency modulation template. The LFM convolutional regression matrix's first... The column represents the preset standard LFM linear frequency modulation template delay. The waveform after sampling points is represented by the following matrix elements: .
[0077] in, For the sample index in the LFM probe segment's received observation vector, For the delay tap index of the LFM probe segment; The first digit of the LFM convolutional regression matrix line, number Column elements; For the preset standard LFM linear frequency modulation template discrete sequence in the index The sampled values at that location.
[0078] A local observation model for the LFM probe segment is established based on the received observation vectors of the LFM probe segment and the LFM convolutional regression matrix: .
[0079] in, The channel response corresponding to the LFM probe segment. The LFM probe segment observation error is caused by a combination of random noise, mud pump harmonics, and unmodeled disturbances.
[0080] Sparse regularized channel response estimation for the local observation model of the LFM probe segment: .
[0081] in, The results are the channel response estimation results for the LFM probe segment; For LFM detector segment receiving observation vectors, This is the LFM convolutional regression matrix. The channel response corresponding to the LFM probe segment. The sparse regularization coefficient; Represents the L2 norm, It represents the first norm.
[0082] When the maximum tap amplitude of the LFM probe channel response estimation result is less than the preset minimum effective tap amplitude When the maximum tap amplitude of the LFM coarse delay support determination is not less than the preset minimum effective tap amplitude, the output flag is displayed. At that time, the LFM coarse delay support set is determined based on the relative amplitude of each delay tap. : in, The first element in the LFM probe segment channel response estimation vector is... The amplitude of each delay tap; The delay tap index to be determined; To traverse the complete set of delay taps To find the index of the maximum tap amplitude; This is the largest value among all delay taps; The threshold for determining the amplitude of the coarse delay support is given, and it satisfies the following conditions: .
[0083] LFM coarse delay support set Each element in the array represents the position of the delay tap in the LFM probe channel response that satisfies the amplitude determination condition.
[0084] Calculate the first... The support quality of the LFM coarse delay support: .
[0085] in, For the first The support quality of each LFM coarse delay support reflects its reliability relative to other coarse delay supports and satisfies the following: .
[0086] Step S3 uses the LFM probe segment received observation vector output from step S2 and the preset standard LFM linear frequency modulation template output from step S1 to construct the LFM convolutional regression matrix and the LFM probe segment local observation model, obtaining the LFM probe segment channel response estimation result, and further obtaining the LFM coarse delay support set. and the support mass of each coarse delay support .
[0087] Step S3 is used to determine the coarse delay region where the channel propagation component is located by using the stable wide time-frequency structure of the LFM probe segment, and to convert the credibility of each coarse delay support into support quality, so as to provide position and quality priors for subsequent coarse delay support pilot band transfer.
[0088] like Figure 2 As shown, step S3 corresponds to "Determination of LFM coarse delay support set and support mass". Figure 3 As shown, the LFM coarse delay support set is generated by the LFM probe segment and serves as the input for the local Markov support transfer.
[0089] Step S4: Construct the local Markov support transition matrix and calculate the pilot band prior support score. The ground receiving equipment determines the nominal delay tap drift based on the historical tap position difference between the LFM probe band channel support and the pilot band channel support. and drift neighborhood radius .
[0090] In the initialization phase where there is no historical successful estimation of communication frames, let: .
[0091] And order: .
[0092] That is, each LFM coarse delay support guides the frequency band with the same index delay tap and its left and right adjacent delay taps for transfer.
[0093] In the most recent When estimating the communication frame for the first historical successful frame, let the first frame be... The delay tap index corresponding to the maximum support mass of the LFM probe segment in a given historical communication frame is: The delay tap index corresponding to the maximum channel tap in the pilot band is Then the first Tap drift amount corresponding to each historical communication frame for: .
[0094] The nominal delay tap drift is determined to be the most recent The median of the tap drift of a historical communication frame, rounded down: .
[0095] in, For median operations; For integer operations; For the first The tap drift amount corresponding to the communication frame was successfully estimated in the past.
[0096] The drift neighborhood radius is determined to be the nearest The maximum deviation of a historical communication frame from the nominal delay tap drift: .
[0097] Ground receiving equipment based on nominal time delay tap drift and drift neighborhood radius Construct the local Markov support transition matrix: .
[0098] in, For LFM detection segment The first delay tap guide frequency band The transfer weight of a time-delay tap This is the delay tap index for the LFM probe segment. This is the delay tap index for the pilot band.
[0099] This embodiment uses a uniformly distributed transfer weight within the drift neighborhood: .in, For a complete set of delay taps; For a complete set of delay taps China satisfies The number of pilot band delay taps. Specifically, this can be expressed as: .
[0100] For each LFM coarse delay support The corresponding transition weights satisfy: .
[0101] The local Markov support transfer matrix is used to characterize the relationship between the time delay taps of the LFM probe band and the support transfer of each time delay tap in the guide band within the nominal time delay tap drift and the drift neighborhood.
[0102] Based on the local Markov support transfer matrix and the support mass of each LFM coarse delay support, the pilot band's first... Prior support score for each delay tap: .
[0103] in, For pilot band number The prior support score for each delay tap. The higher the prior support score, the more reliable the delay tap in that pilot band is derived from the coarse delay support of the high-quality LFM.
[0104] Step S4 utilizes the LFM coarse delay support set and the support quality of each coarse delay support output in step S3, and constructs a local Markov support transfer matrix using the nominal time delay tap drift and drift neighborhood radius, to obtain the prior support score for each time delay tap in the pilot band. .
[0105] Step S4 converts the coarse delay support position and support quality in the LFM probe segment to the time delay tap coordinates of the pilot segment, and allows limited tap position changes between the LFM probe segment channel support and the pilot segment channel support through the drift neighborhood.
[0106] like Figure 2 and Figure 3 As shown, step S4 corresponds to "local Markov support transfer". Each LFM coarse delay support is... To nominally transfer positions, and to its left and right. Within each delay tap range, a corresponding pilot band candidate delay region is formed.
[0107] Step S5: Determine the local candidate delay set for the pilot band.
[0108] Based on the prior support scores of each delay tap in the pilot band obtained in step S4, the ground receiving equipment forms a pilot band local candidate delay set by combining all pilot band delay taps with prior support scores greater than zero. : .
[0109] The local candidate delay set of the pilot band can also be equivalently represented as: .
[0110] When the drift neighborhoods corresponding to different LFM coarse delay supports overlap, the overlapping delay taps are retained only once using set union operation.
[0111] In this embodiment, all delay taps formed by LFM coarse delay support through local Markov support transfer are included in the pilot band local candidate delay set. Different delay tap combinations are no longer enumerated from the pilot band local candidate delay set, nor are the highest quality delay taps selected. The effectiveness of each candidate delay tap is jointly characterized by the channel response coefficients estimated in step S6 and the corresponding prior support scores.
[0112] The term "local" refers to the set of complete delay taps ranging from zero delay to the maximum preset delay. The local candidate delay set of the pilot band satisfies the following: .
[0113] When the local candidate delay set of the pilot band is empty, increase the drift neighborhood radius by one delay tap: .
[0114] The number 1 indicates that the drift neighborhood radius will be increased by one time delay tap.
[0115] Then steps S4 and S5 are repeated. When the drift neighborhood radius reaches the preset maximum drift neighborhood radius... Afterwards, when the local candidate delay set of the pilot band is still empty, output a coarse delay support transfer failure flag.
[0116] Step S5 uses the prior support scores of each delay tap of the pilot band output in step S4 to obtain the local candidate delay set of the pilot band. .
[0117] Step S5 reduces the range of channel response to be estimated in the pilot band from the complete set of time delay taps to a partial set of time delay taps related to LFM coarse delay support, thereby reducing the number of channel coefficients to be estimated under short pilot conditions.
[0118] like Figure 2 As shown, step S5 corresponds to "determining the local candidate delay set for the pilot band". Figure 3 As shown, the union of candidate delay regions formed by multiple LFM coarse delay-supported transitions in the complete delay tap set constitutes the pilot band local candidate delay set. Among them, Figure 3 The upper part sequentially shows the LFM detection segment, guard interval, and pilot segment along the time direction; Figure 3 The lower part sequentially shows the LFM coarse delay support set, nominal delay tap drift and drift neighborhood radius, local Markov support transition matrix, pilot band prior support score, and pilot band local candidate delay set.
[0119] Step S6: Estimate the mud pulse telemetry channel response within the local candidate delay set of the pilot band. Ground receiving equipment based on known pilot sequences Constructing complete pilot training convolutional matrices using different delayed copies The first complete pilot training convolution matrix The column represents the known pilot sequence delay. The waveform after sampling points is represented by the following matrix elements: in, The complete pilot training convolution matrix is equivalent to the complete set of time delay taps. The corresponding complete pilot training convolution matrix has a size of Each column and The complete delay taps in the data correspond one-to-one; The row index of the complete pilot training convolution matrix is... Column index is Element. The total number of samples in the entire pilot band is defined by the preset communication frame structure parameters. The set of local candidate delays for the pilot band is retained from the complete pilot training convolution matrix. The corresponding columns yield the pilot band local training convolution matrix. : .
[0120] in, This selects all rows of the matrix.
[0121] A pilot band local channel estimation model is established based on the pilot band received observation vector and the pilot band local training convolution matrix: .
[0122] in, This is the channel response coefficient vector within the local candidate delay set of the pilot band; The pilot band observation error is caused by a combination of random noise, mud pump harmonics, and unmodeled disturbances. This is the pilot band receive observation vector.
[0123] Construct a prior weighted regularization matrix based on the pilot band prior support score obtained in step S4. : .
[0124] in, The value is set to a positive number to prevent the denominator from being zero. The candidate tap sequence number is the local candidate delay set within the pilot band. For the set of A complete latency index for each candidate latency tap; For complete delay index The corresponding pilot band prior support score; This represents the number of delay taps in the local candidate delay set of the pilot band. This is to form a diagonal matrix with the elements within the parentheses as the main diagonal elements.
[0125] For delay taps with larger pilot band prior support scores, the corresponding regularization penalty is smaller; for delay taps with smaller pilot band prior support scores, the corresponding regularization penalty is larger.
[0126] The channel response coefficients within the local candidate delay set of the pilot band are estimated using a priori weighted regularized least squares method (i.e., the channel response coefficients are solved using the following weighted regularized least squares objective function): .
[0127] in, The pilot band regularization coefficient; This is the local training convolution matrix for the pilot band; This is the channel response coefficient vector within the local candidate delay set of the pilot band; This is the channel response coefficient estimation vector within the local candidate delay set of the pilot band.
[0128] The closed-form solution to the above prior weighted regularized least squares problem is: .
[0129] in, This is the conjugate transpose of the prior weighted regularized matrix; This is the conjugate transpose of the pilot band local training convolution matrix.
[0130] Setting the channel taps outside the pilot band local candidate delay set to zero yields the complete mud pulse telemetry channel response. Let the pilot band local candidate delay set, sorted in ascending order of delay tap index, be [equation missing]. ,in, This represents the number of delay taps in the local candidate delay set of the pilot band. For the first Candidate latency tap indexes. For a complete latency tap index. Its channel response estimation coefficients are expressed as: .
[0131] in, The first complete mud pulse telemetry channel response Channel response estimation coefficients for each delay tap; This is the channel response coefficient estimation vector within the local candidate delay set of the pilot band; The vector represents the first Each element.
[0132] The mud pulse telemetry channel response estimation vector, composed of all time-delay taps, is expressed as: .
[0133] in, The first complete mud pulse telemetry channel response Channel response estimation coefficients for each delay tap; For mud pulse telemetry channel response estimation, from all The mud pulse telemetry channel response estimation vector is composed of the channel response estimation coefficients of each time delay tap. This is the maximum preset number of delay taps; This is a transpose.
[0134] To determine the validity of the mud pulse telemetry channel response estimation results, the normalized pilot fitting residual is calculated. : .
[0135] in, To preset the residual normalization stability constant, satisfying This is used to keep the denominator of the normalized pilot fitting residual non-zero when the energy of the received observation vector in the pilot band is zero or close to zero. Dimensions and Consistent.
[0136] Normalized pilot fitting residuals Not greater than the preset residual threshold At that time, confirm that the mud pulse telemetry channel response estimation is successful, and output the mud pulse telemetry channel response estimate. .
[0137] Normalized pilot fitting residuals Greater than the preset residual threshold When the drift neighborhood radius reaches the preset maximum drift neighborhood radius, add a time delay tap and re-execute steps S4 to S6. If the normalized pilot fitting residual is still greater than the preset residual threshold, then the channel estimation failure flag for the current communication frame is output.
[0138] Step S6 uses the pilot band received observation vector output in step S2, the known pilot sequence output in step S1, the pilot band local candidate delay set output in step S5, and the pilot band prior support score output in step S4 to obtain the mud pulse telemetry channel response estimate. .
[0139] Step S6 utilizes the accurate estimation capability of the pilot sequence to determine the specific coefficients of each channel tap under the local time delay prior constraint provided by the LFM coarse delay support. This combines the stable positioning of the LFM probe segment and the accurate estimation of the pilot segment coefficients to reduce the degree to which the full-time-delay tap range estimation is affected by short pilots, mud pump harmonics, and random noise.
[0140] like Figure 2 As shown, step S6 corresponds to "pilot band local channel response estimation". Figure 4 As shown, the ground receiving equipment retains only the columns in the complete pilot training convolution matrix that correspond to the local candidate delay set of the pilot band, and estimates the mud pulse telemetry channel response within the local candidate delay set of the pilot band. Figure 4 The complete lateral delay range in the table represents the complete delay tap set. Candidate set 1 and candidate set 2 represent candidate delay regions formed by local Markov support transfer from different LFM coarse delay supports, respectively. The union of candidate set 1 and candidate set 2 constitutes the pilot band local candidate delay set. .
[0141] Technical Effect Verification: To verify the technical effect of the method described in this application, the number of taps in the pilot band local candidate delay set can be compared with the number of taps in the complete delay tap set. The proportion of the local candidate delay set is defined. for: .
[0142] in, This represents the number of delay taps in the local candidate delay set of the pilot band. A local candidate delay set ratio of less than one indicates that this application reduces the range to be estimated for the pilot band channel response from the complete delay tap set to a partial set of candidate delay taps.
[0143] Alternatively, the channel estimation method with complete time delay tap range and the channel estimation method with pilot band local candidate delay set provided in this application can be used respectively to compare the normalized pilot fitting residual, effective channel tap positioning error and mud pulse telemetry channel response estimation error of the two methods.
[0144] This application utilizes the LFM probe segment to stably obtain the coarse delay support of the channel. Based on the finite tap drift relationship, the coarse delay support is transferred to the pilot segment to form a local candidate delay set of the pilot segment. Within the local candidate delay set, the pilot sequence is used to accurately estimate the mud pulse telemetry channel response. This simultaneously leverages the advantages of stable positioning of the LFM probe segment coarse support and accurate estimation of the pilot segment channel coefficients, solving the problem of unstable channel estimation across the entire time delay range under short pilot and strong interference conditions.
[0145] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.
[0146] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0147] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).
[0148] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, etc., and are not limited to these.
[0149] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0150] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A method for estimating the channel response of mud pulse telemetry, characterized in that, include: Based on the pressure digital signal, the preset standard LFM linear frequency modulation template, and the preset communication frame structure parameters, determine the LFM detection segment receiving observation vector and the pilot segment receiving observation vector; Based on the LFM probe segment received observation vectors and the LFM convolutional regression matrix, a local observation model for the LFM probe segment is constructed. Sparse regularized channel response estimation is performed on the local observation model of the LFM probe segment to determine the LFM coarse delay support set and the support quality of each coarse delay support; Construct a local Markov support transition matrix, and based on the local Markov support transition matrix, the LFM coarse delay support set, and the support quality of each LFM coarse delay support, calculate the prior support score of each delay tap in the pilot band. Using the prior support scores of each delay tap in the pilot band, a local candidate delay set for the pilot band is determined; Based on the received observation signal of the pilot band and the pilot training convolution matrix, a pilot band local channel estimation model is constructed. Based on the local candidate delay set and pilot sequence of the pilot band, the local channel estimation model of the pilot band is solved by the regularized least squares method to determine the mud pulse telemetry channel response estimate.
2. The mud pulse telemetry channel response estimation method according to claim 1, characterized in that, Based on the pressure digital signal, the preset standard LFM linear frequency modulation template, and the preset communication frame structure parameters, the LFM probe band received observation vector and the pilot band received observation vector are determined, specifically including: Normalize and match the preset standard LFM linear frequency modulation template and the pressure digital signal to determine the LFM matching correlation sequence; The position corresponding to the maximum value of the LFM matching correlation sequence is determined as the starting sample position of the LFM probe segment; The starting sample position of the LFM probe segment is used as the reference time of the communication frame. Based on the communication frame reference time and preset communication frame structure parameters, determine the actual sample index set of the LFM detection segment, the actual sample index set of the protection interval, and the actual sample index set of the pilot segment; The LFM detection segment's received observation vector is determined based on the actual sample index set of the LFM detection segment; The pilot band receive observation vector is determined based on the actual sample index set of the pilot band.
3. The mud pulse telemetry channel response estimation method according to claim 1, characterized in that, Based on the received observation vectors and LFM convolutional regression matrix of the LFM probe segment, a local observation model for the LFM probe segment is constructed, specifically including: Based on the preset standard LFM linear frequency modulation template, construct an LFM convolutional regression matrix; Based on the received observation vectors of the LFM probe segment and the LFM convolutional regression matrix, a local observation model for the LFM probe segment is established. The local observation model for the LFM probe segment is as follows: ; in, This is the LFM convolutional regression matrix; The channel response corresponding to the LFM probe segment; The LFM probe segment observation error is caused by a combination of random noise, mud pump harmonics, and unmodeled disturbances. Receive observation vectors for the LFM probe segment.
4. The mud pulse telemetry channel response estimation method according to claim 1, characterized in that, Sparse regularized channel response estimation is performed on the local observation model of the LFM probe segment to determine the LFM coarse delay support set and the support quality of each coarse delay support, specifically including: Sparse regularized channel response estimation is performed on the local observation model of the LFM probe segment to determine the channel response estimation result of the LFM probe segment; The LFM coarse delay support set is determined based on the maximum tap amplitude and the preset minimum effective tap amplitude of the LFM probe channel response estimation result. The support quality of each coarse delay support is calculated based on the channel tap energy corresponding to each coarse delay support in the LFM coarse delay support set.
5. The mud pulse telemetry channel response estimation method according to claim 1, characterized in that, A local Markov support transition matrix is constructed, and based on the local Markov support transition matrix, the LFM coarse delay support set, and the support quality of each LFM coarse delay support, the prior support score of each delay tap in the pilot band is calculated, specifically including: Based on the historical tap position difference between the LFM probe segment channel support and the pilot segment channel support, determine the nominal delay tap drift and drift neighborhood radius; Based on the nominal time delay tap drift amount and the drift neighborhood radius, construct a local Markov support transition matrix; Based on the local Markov support transition matrix, the LFM coarse delay support set, and the support quality of each LFM coarse delay support, the prior support score of each delay tap in the pilot band is calculated.
6. The mud pulse telemetry channel response estimation method according to claim 1, characterized in that, Using the prior support scores of each delay tap in the pilot band, a local candidate delay set for the pilot band is determined, specifically including: Based on the prior support score of each delay tap in the pilot band, the pilot band delay taps with a prior support score greater than zero are determined as the pilot band local candidate delay set.
7. The mud pulse telemetry channel response estimation method according to claim 1, characterized in that, Based on the received observation signal and pilot training convolution matrix of the pilot band, a pilot band local channel estimation model is constructed, specifically including: Based on the pilot sequence, construct the pilot segment local training convolution matrix; A pilot band local channel estimation model is constructed based on the pilot band received observation vector and the pilot band local training convolution matrix.
8. The mud pulse telemetry channel response estimation method according to claim 1, characterized in that, The pilot band local channel estimation model is as follows: ; in, This is the channel response coefficient vector within the local candidate delay set of the pilot band; The pilot band observation error is caused by a combination of random noise, mud pump harmonics, and unmodeled disturbances. This is the local training convolution matrix for the pilot band; This is the pilot band receive observation vector.
9. The mud pulse telemetry channel response estimation method according to claim 1, characterized in that, The preset communication frame structure includes an LFM probe segment, a guard interval, and a pilot segment arranged sequentially along the time direction of the communication frame; the preset communication frame structure parameters include at least the number of LFM probe segment samples, the number of guard interval samples, the number of pilot segment samples, and the maximum preset delay tap number.
10. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the mud pulse telemetry channel response estimation method according to any one of claims 1-9.
Citation Information
Patent Citations
Time sequence multiple sparse Bayesian channel and impulse noise joint estimation method
CN121644279A
Channel estimation method of steel penetration system
US20210266196A1
