CN120090910B

CN120090910B - Dual-mode FSK demodulation high-speed well logging communication method and system

Info

Publication number: CN120090910B
Application number: CN202510584751.0A
Authority: CN
Inventors: 侯立东; 张岳霞; 白劲松
Original assignee: Deep Blue Tianjin Intelligent Manufacturing Co ltd
Current assignee: Deep Blue Tianjin Intelligent Manufacturing Co ltd
Priority date: 2025-05-08
Filing date: 2025-05-08
Publication date: 2025-07-15
Anticipated expiration: 2045-05-08
Also published as: CN120090910A

Abstract

The application provides a high-speed well logging communication method and system for dual-mode FSK demodulation, which relate to the technical field of communication, and the method comprises the steps of receiving and preprocessing an analog FSK modulation signal, and then executing signal granularity discrimination to dynamically update an internal recognition mode; the method comprises the steps of receiving a preprocessing result, sending the preprocessing result to a first demodulation module and a second demodulation module, obtaining the output result of the module, utilizing an internal recognition mode to fuse the output result, establishing a fused preliminary judgment and an adjacent frequency point set, carrying out soft judgment triggering analysis based on the fused preliminary judgment and the adjacent frequency point set, and carrying out high-speed logging communication after updating the fused preliminary judgment by utilizing the adjacent frequency point set if soft judgment compensation is triggered. The application solves the technical problems of insufficient anti-interference capability in underground strong noise environment, reduced communication reliability and high error rate caused by the dependence of the synchronous mode on the fixed recognition mode in the demodulation process in the prior art, improves the demodulation accuracy of communication signals, and remarkably improves the overall communication quality and stability.

Description

Dual-mode FSK demodulation high-speed well logging communication method and system

Technical Field

The application relates to the technical field of communication, in particular to a dual-mode FSK demodulation high-speed well logging communication method and system.

Background

With the continuous improvement of the requirements of oil and gas exploration and development on real-time performance and data volume, a high-speed well communication technology becomes an important means for information transmission between a well and the ground. The Frequency Shift Keying (FSK) modulation and demodulation method is adopted, so that the method is simple in implementation and good in system stability, and becomes a main technical route in the current underground logging communication. However, the existing FSK demodulation method has obvious disadvantages. Because the synchronization and demodulation processes depend on the fixed templates, the system is easy to be influenced by factors such as strong noise interference, frequency drift and the like in a complex underground electromagnetic environment, has limited anti-interference capability, is difficult to adapt to channel characteristic changes in time, and easily causes communication reliability reduction and error rate increase, so that accurate transmission of logging data is influenced.

Disclosure of Invention

The application provides a dual-mode FSK demodulation high-speed well communication method and a dual-mode FSK demodulation high-speed well communication system, which solve the technical problems of communication reliability reduction and high bit error rate caused by insufficient anti-interference capability in a downhole strong noise environment due to the fact that a synchronous mode depends on a fixed recognition mode in the FSK demodulation process in the prior art, and achieve the technical effects of improving the demodulation accuracy of high-speed well communication signals, and further remarkably improving the overall communication quality and stability.

In view of the above problems, in one aspect, the present application provides a high-speed well communication method for dual-mode FSK demodulation, which includes receiving an analog FSK modulation signal of a downhole communication link, preprocessing the analog FSK modulation signal, performing signal granularity discrimination of a preprocessing result, dynamically updating an internal recognition mode of the dual-mode FSK by using the signal granularity discrimination result, respectively transmitting the preprocessing result to a first demodulation module and a second demodulation module of the dual-mode FSK, wherein the first demodulation module is a main frequency-energy demodulation module, the second demodulation module is a local frequency spectrum dynamic feature extraction module, respectively acquiring output results corresponding to the first demodulation module and the second demodulation module, fusing the output results by using the internal recognition mode, establishing a fusion preliminary decision, searching adjacent frequency points by using the fusion preliminary decision, establishing an adjacent frequency point set, performing soft decision triggering analysis based on the fusion preliminary decision and the adjacent frequency point set, updating the fusion preliminary decision by using the adjacent frequency point set if the soft decision triggering compensation, and performing high-speed preliminary communication after updating.

On the other hand, the application also provides a dual-mode FSK demodulation high-speed well communication system, which comprises a signal granularity judging module, an identification mode updating module, a result sending module, a trigger analysis module and a preliminary judgment and soft-trigger analysis module, wherein the signal granularity judging module is used for receiving an analog FSK modulation signal of a downhole communication link and carrying out signal granularity judgment of a preprocessing result after preprocessing the analog FSK modulation signal, the identification mode updating module is used for dynamically updating an internal identification mode of the dual-mode FSK by utilizing the signal granularity judging result, the result sending module is used for respectively sending the preprocessing result to a first demodulation module and a second demodulation module of the dual-mode FSK, the first demodulation module is a main frequency-energy demodulation module, the second demodulation module is a local frequency spectrum dynamic characteristic extracting module, the result fusing module is used for respectively acquiring output results corresponding to the first demodulation module and the second demodulation module, fusing the output results by utilizing the internal identification mode, establishing a fused preliminary judgment, the adjacent frequency point searching module is used for carrying out adjacent frequency point searching on the fusing judgment, the preliminary frequency point searching module is used for triggering the adjacent frequency point set by utilizing the fusing judgment, the preliminary frequency point searching module is used for carrying out the preliminary frequency searching and the adjacent frequency point searching and the preliminary frequency point searching and the soft-frequency point searching trigger is used for carrying out the updating after the primary frequency searching and the adjacent frequency point is used for carrying out the primary frequency searching communication.

The one or more technical schemes provided by the application have at least the following beneficial effects:

The method has the advantages that the standard processing of the original FSK signal is finished by receiving and preprocessing the analog FSK modulation signal of the underground communication link, then the detail change degree of the signal is judged and analyzed through the signal granularity, a foundation is provided for the follow-up dynamic adaptation demodulation mode, the internal recognition mode of the dual-mode FSK is dynamically updated according to the signal granularity judging result, the internal mode relied on by demodulation can be dynamically adjusted according to the current signal characteristic, the self-adaption capability to channel change is improved, and therefore anti-interference performance is enhanced. And the preprocessing result is respectively sent to a first demodulation module and a second demodulation module of the dual-mode FSK, and coarse demodulation is performed on the basis of the overall frequency and the energy through a dual-path processing mechanism, and weak change is captured through local frequency spectrum feature extraction to form complementation, so that the accuracy and the robustness of demodulation are improved. And respectively acquiring output results corresponding to the first demodulation module and the second demodulation module, fusing the demodulation results of the double paths by using an internal recognition mode, establishing fusion preliminary judgment, and improving demodulation reliability. And searching adjacent frequency points for the fusion preliminary judgment, establishing an adjacent frequency point set, constructing an auxiliary reference set, and preparing subsequent soft judgment compensation. And carrying out soft decision triggering analysis based on the fusion preliminary decision and the adjacent frequency point set, and if soft decision compensation is triggered, utilizing the adjacent frequency point set to carry out fine adjustment or correction on the fusion preliminary decision, thereby obviously reducing the demodulation error probability and further improving the noise immunity. And data reduction and communication are carried out by utilizing the updated fusion preliminary judgment, so that stable and reliable communication in a high-speed well logging environment is ensured, and system-level performance improvement is realized.

In summary, the application realizes the dynamic perception of FSK signal characteristics in a complex underground environment by introducing signal granularity discrimination, updates the internal recognition mode in real time based on the granularity analysis result, obviously improves the adaptability of the demodulation process to channel state change, adopts a dual-mode demodulation structure of main frequency-energy demodulation and local frequency spectrum dynamic characteristic extraction, combines the internal recognition mode to carry out fusion judgment on dual-mode output, improves the accuracy of signal recognition, enhances the integral anti-interference capability, and further corrects potential demodulation deviation by constructing an adjacent frequency point set and triggering soft decision compensation when the fusion preliminary judgment has uncertainty, thereby effectively reducing the error rate. In the whole, the application effectively enhances the accuracy and reliability of signal demodulation, particularly in a complex underground environment, can obviously reduce the error rate, improves the communication quality and meets the actual requirement of a high-speed well on high-reliability data transmission.

The foregoing description is only an overview of the present application, and is intended to be implemented in accordance with the teachings of the present application in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present application more readily apparent.

Drawings

Fig. 1 is a flow chart of a dual-mode FSK demodulation high-speed well communication method according to an embodiment of the present application.

Fig. 2 is a schematic flow chart of performing item-by-item check backtracking based on a check chain in the dual-mode FSK demodulation high-speed well communication method according to the embodiment of the present application.

Fig. 3 is a schematic structural diagram of a dual-mode FSK demodulation high-speed well communication system according to an embodiment of the present application.

Reference numerals illustrate that the signal granularity discriminating module 10, the identifying mode updating module 20, the result transmitting module 30, the result fusing module 40, the adjacent frequency point searching module 50, the triggering analyzing module 60 and the communication executing module 70.

Detailed Description

The embodiment of the application provides a dual-mode FSK demodulation high-speed well communication method and a dual-mode FSK demodulation high-speed well communication system, which solve the technical problems of insufficient anti-interference capability in a downhole strong noise environment and high communication reliability and error rate caused by the fact that a synchronous mode depends on a fixed recognition mode in the FSK demodulation process in the prior art by dynamically updating an internal recognition mode and combining dual-mode feature fusion judgment, and achieve the technical effects of improving the demodulation accuracy of high-speed well communication signals, and further remarkably improving the overall communication quality and stability.

An embodiment of the present application provides a dual-mode FSK demodulation high-speed well communication method, as shown in fig. 1, including:

Step S100, receiving an analog FSK modulation signal of a downhole communication link, preprocessing the analog FSK modulation signal, and then judging the signal granularity of a preprocessing result.

Specifically, the downhole communication link refers to a signal transmission channel between the downhole tool and the surface control center for transmitting logging data and other information. The analog FSK modulated signal is an analog signal generated by a Frequency Shift Keying (FSK) modulation method, and different data bits are represented by changing the frequency of the signal. Granularity of a signal, i.e., the degree of fineness of the signal or the intensity of a characteristic change.

The analog FSK modulated signal is received over a downhole communication link. Taking petroleum drilling logging as an example, geological data (such as resistivity, porosity, etc.) acquired by downhole instruments at different intervals may be converted into analog FSK modulated signals that are sent to a surface system. And then carrying out preprocessing operations such as filtering, noise suppression, amplitude normalization and the like on the analog FSK modulation signal, then carrying out signal granularity discrimination on the preprocessed signal, and discriminating that the signal belongs to a coarse granularity stable change state or a fine granularity rapid change state by analyzing the main frequency change trend, the energy distribution stability and the short-time spectrum expansion condition of the signal, thereby providing a basis for the selection of a subsequent dynamic demodulation mode. For example, when the frequency of the signal changes slowly and the amplitude is large, the coarse-grain signal is discriminated, whereas when the frequency of the signal changes frequently and the amplitude is small, the fine-grain signal is discriminated.

Through the step, the basic cleaning and granularity classification can be completed quickly when the original signal is received, a targeted mode selection basis is provided for the subsequent demodulation process, and the adaptability and the robustness of the whole demodulation process are improved effectively.

Step 200, dynamically updating the internal recognition mode of the dual-mode FSK by using the signal granularity discrimination result.

Specifically, the internal identification mode is a mode or strategy for identifying and processing signals within the dual-mode FSK demodulation system, and different modes correspond to different demodulation parameters and algorithm configurations to adapt to signals with different characteristics. And (3) dynamically updating the internal recognition mode of the dual-mode FSK in real time according to the signal granularity judging result obtained in the step (S100), if the signal granularity judging result is a coarse granularity, starting a demodulation strategy with long time window and low variation sensitivity to improve the demodulation speed and stability, and if the signal granularity judging result is a fine granularity, starting a high confidence coefficient and short time window tracking mode to improve the sensing capability of the fine variation, so that the subsequent demodulation module can extract signal characteristics and judge bits according to the optimal mode.

The step can adapt to the complex underground signal environment by dynamically updating the internal recognition mode, improves the demodulation accuracy, reduces the error recognition probability and ensures that the subsequent demodulation module operates in the optimal working state.

Step S300, the preprocessing result is respectively sent to a first demodulation module and a second demodulation module of the dual-mode FSK, wherein the first demodulation module is a main frequency-energy demodulation module, and the second demodulation module is a local frequency spectrum dynamic characteristic extraction module.

Specifically, the preprocessing result obtained in step S100 is sent to two demodulation modules simultaneously, wherein the first demodulation module is a main frequency-energy demodulation module and is dedicated to extracting the main frequency and energy distribution characteristics of the signal, and preliminary bit judgment is performed through the energy peak value of the main frequency point, and the second demodulation module is a local frequency spectrum dynamic characteristic extraction module and is responsible for extracting dynamic variation trends in the local frequency spectrum of the signal, such as frequency offset, local energy diffusion and other characteristics, and is used for assisting in detecting potential signal drift or weak variation conditions, and the two demodulation modules operate in parallel and respectively output independent preliminary analysis results.

In the step, the signals are demodulated from different angles through two different demodulation modules, so that more comprehensive signal information can be obtained, and the accuracy and reliability of demodulation are improved.

And step 400, respectively obtaining output results corresponding to the first demodulation module and the second demodulation module, and fusing the output results by utilizing the internal recognition mode to establish fusion preliminary judgment.

Specifically, the primary decision is to fuse the output results of the two demodulation modules according to a certain rule to obtain a primary decision result, which is used for representing the bit sequence represented by the signal. And (3) respectively receiving a main frequency-energy judgment result from the first demodulation module and a local frequency spectrum dynamic characteristic extraction result from the second demodulation module, carrying out fusion processing on the two output results by adopting mechanisms such as weighted fusion, characteristic complementation or confidence screening according to the internal recognition mode updated in the step S200, preferentially retaining bit judgment of consistent main frequency and dynamic characteristic support, simultaneously deciding by applying a set fusion rule to information with difference, and finally generating a fusion preliminary decision, thereby providing a basis for further optimization.

The method and the device have the advantages that the two-module information is subjected to deep fusion, so that the stability of main frequency energy is utilized, the sensitivity of local frequency change is also considered, the accuracy and the anti-interference capability of preliminary bit judgment are greatly improved, and a foundation is provided for subsequent further optimization and compensation.

And S500, searching adjacent frequency points for the fusion preliminary judgment, and establishing an adjacent frequency point set.

Specifically, the adjacent frequency point set is a set of frequency points adjacent to the frequency points in the preliminary decision result in the frequency domain, and the frequency points may contain secondary components of useful signals or frequency point information of interference signals, so that the preliminary decision can be verified and optimized in an auxiliary manner. And taking the central frequency point corresponding to the fusion preliminary judgment as a reference, carrying out energy scanning and dynamic characteristic analysis in a certain frequency range around the central frequency point, searching adjacent frequency points with potential signal components, screening effective frequency points according to conditions such as energy threshold value, frequency shift trend consistency and the like, and finally forming a set containing a plurality of adjacent frequency points to assist subsequent soft judgment compensation analysis.

By establishing the adjacent frequency point set, the frequency information which cannot be completely covered by the primary judgment can be effectively captured, redundancy and flexibility in soft judgment are increased, and the recovery capability of weak signals under complex channel conditions is improved.

And step 600, carrying out soft decision triggering analysis based on the fusion preliminary decision and the adjacent frequency point set, and updating the fusion preliminary decision by utilizing the adjacent frequency point set if soft decision compensation is triggered.

Specifically, the soft decision triggering analysis refers to judging whether to perform soft decision compensation on a preliminary decision result according to a certain rule and condition, that is, whether to correct the preliminary decision by using information of an adjacent frequency point set so as to improve the accuracy of the decision. The soft decision compensation means that the frequency point information in the adjacent frequency point set is utilized to adjust and optimize the preliminary decision result so as to reduce misjudgment and improve the reliability of demodulation. Based on the fusion preliminary judgment and the adjacent frequency point set, the soft judgment triggering analysis is executed, whether the preliminary judgment confidence coefficient meets a set threshold value is firstly evaluated, if the preliminary judgment confidence coefficient is lower than the threshold value or judgment conflict exists, a soft judgment compensation flow is triggered, the fusion preliminary judgment with higher confidence coefficient is finally updated by referring to the energy level, dynamic consistency and correlation indexes of each frequency point in the adjacent frequency point set and the main judgment frequency point, and the judgment result is revised after comprehensive weighting.

By setting the soft decision mechanism, decision correction and optimization can be automatically carried out under the condition of large channel quality fluctuation or interference, the adaptability and fault tolerance to underground complex environments are greatly improved, and the accuracy of final logging communication data is improved.

And step S700, performing high-speed logging communication by using the updated fusion preliminary judgment.

Specifically, based on the fusion preliminary judgment result after soft judgment compensation updating, actual high-speed well logging communication is executed, bit streams obtained through judgment are transmitted back to a ground system in real time, and meanwhile, the sending rate and coding parameters are dynamically adjusted according to the state of a communication link, so that stable and low-bit-error-rate data transmission can be realized under the underground high-interference and high-attenuation environment, and the real-time monitoring and control requirements of underground operation are supported. For example, the bit sequence is converted to specific geological parameter values (e.g., resistivity values of the formation, etc.) or other instrument state information. These data information are then transmitted at high speed to the surface control center via the downhole communication link in accordance with a certain communication protocol, such as the DSL protocol. At the surface control center, the received data is further processed and analyzed for guiding the drilling operation, geological analysis and other works.

By adopting the updated high-confidence judgment result to carry out data transmission, the error rate and the retransmission times in the logging communication process can be obviously reduced, the effective load capacity and the instantaneity of a communication link are improved, and the accurate and efficient transmission of underground measurement and control data is ensured.

Further, step S600 includes:

step S610, if the soft decision compensation is not triggered, establishing a preliminary candidate bit stream based on the fusion preliminary decision.

Step S620, checking the consistency of the preliminary candidate bit stream by using the sliding window.

And step 630, triggering strong decision compensation if the consistency check fails, performing backtracking correction of fusing the primary decisions by using the strong decision compensation, and performing high-speed logging communication by using a backtracking correction result.

Specifically, the primary candidate bit stream is a group of primary bit sequences generated according to the fusion primary decision result, and is subject to further checksum processing after being subjected to no depth correction but has a communication structure. If soft decision compensation is not triggered, the fusion preliminary decision is considered to have basic credibility, a preliminary candidate bit stream is directly constructed, and a sliding window consistency check process is further carried out, wherein in the check process, redundancy detection and front-back logic consistency comparison are carried out on local fragments of the bit stream according to the set window length. For example, the window length is set to be 4 bits, the window is placed at the initial position of the initial candidate bit stream, whether the 4 bits meet the preset consistency rule is checked, if yes, whether the 4 bits meet the specific coding rule (such as Manchester coding rule) or whether the requirements of continuity, regularity and the like of data are met, then the window slides forward by one bit position, and the verification is continued, and so on until the verification of the whole initial candidate bit stream is completed. If the consistency check fails, a strong decision compensation mechanism is triggered, decision basis in a plurality of previous sampling periods is traced back, the spectrum structure and decision chains of each node are reevaluated, the original decision with low confidence is removed, the most probable bit sequence is reequipped according to the intensity trend of the frequency point, thus realizing tracing back correction of merging the primary decisions, and finally, the corrected result is used as a new bit stream for communication.

For example, in a certain communication process, a bit stream '1011001' is generated by fusing preliminary decisions, frequent inconsistent support characteristics of '1100' fragments are found when sliding window consistency check (window size is 4) is executed, strong decision compensation is triggered, original energy of a '1' part in spectrum data before backtracking for 5 periods is found to be lower than a threshold value, and then the original energy is corrected to be '1000' through backtracking correction, so that a trusted bit stream '1000001' is finally obtained and used for actual communication transmission.

According to the method, the two-stage mechanism of soft decision compensation and strong decision backtracking is introduced, so that the decision paths can be adaptively adjusted according to uncertainty of different degrees, the overall consistency and accuracy of the data stream are guaranteed while excessive correction is avoided, the robustness and communication reliability of bit stream construction in a downhole high-interference environment are remarkably improved, and solid guarantee is provided for subsequent high-efficiency logging data transmission.

Further, step S630 includes:

In the step S631, in the backtracking verification process, a plurality of data nodes are backtracked, and a verification chain is established, wherein the verification chain comprises a time sequence verification chain and a signal verification chain.

Step S632, starting from the node triggering the strong judgment, performing item-by-item check backtracking based on the check chain to establish the backtracking correction result.

Specifically, after consistency check fails and strong decision compensation is triggered, a backtracking check mechanism is started first, a plurality of historical data nodes are backtracked forward, two types of check chains, namely a sequence check chain and a signal check chain, are respectively established in the backtracking process, wherein the sequence check chain is used for verifying time logic continuity among the nodes, ensuring that bit-to-bit transfer accords with a special sequence jump rule of FSK communication, and the signal check chain is used for detecting continuity of spectral characteristic changes of each node, such as main frequency drift, energy mutation and the like. For the establishment of the time sequence check chain, related historical data nodes are connected according to the time sequence requirement of data transmission, and a time relation rule between each node is determined, for example, a certain node should appear in a specific time after another node. For the establishment of a signal verification chain, the historical data nodes related to the signal characteristics are connected, and the relation rules which are required to be met by the characteristics of the signal frequency, the energy and the like among the nodes are determined.

And then, starting from the current node triggering strong judgment, performing item-by-item check backtracking based on the established check chain, namely respectively executing time consistency check and spectrum characteristic consistency check on each backtracking node, if deviation is found, correcting according to the support degree and the change trend of the adjacent nodes, and finally forming a backtracking correction result through continuous multi-node check, wherein the corrected judgment result is used as a new basis of the bit stream to participate in subsequent high-speed well communication. For example, in a log communication segment, a local anomaly in the bit stream "10011010" is detected, and the first 10 data nodes are traced back after a strong decision is triggered. After the time sequence check chain is established, unreasonable time sequence jump (with extremely small time span) exists between the 4 th node and the 5 th node, and meanwhile, no precursor offset occurs to the frequency spectrum main frequency of the 3 rd to 5 th nodes on the signal check chain. And according to the double checking result of the time sequence and the signal, the bit judgment of the 4 th node is adjusted, the node which is originally judged to be 1 is changed into 0, and the backtracking correction bit stream 10001010 is regenerated, so that the data reliability is greatly improved.

The systematic check chain backtracking is introduced in the strong judgment compensation process, so that the isolated error point can be comprehensively identified and corrected, the continuity correction can be realized according to the dual characteristics of time sequence and frequency spectrum, the overall accuracy of bit judgment in a logging communication link is remarkably improved, the anti-interference capability and the signal integrity are enhanced, the misjudgment rate caused by short-time abnormality is greatly reduced, and the stability and reliability of high-speed logging data transmission are ensured.

Further, as shown in fig. 2, step S632 includes:

And step S632-1, activating a time sequence checking sub-channel, and performing time consistency check based on a time sequence checking chain by using the time sequence checking sub-channel to generate a first attention checking result.

And step S632-2, activating a frequency spectrum verification sub-channel, and utilizing the frequency spectrum verification sub-channel to perform frequency spectrum characteristic consistency verification based on a signal verification chain so as to generate a second attention verification result.

And step S632-3, performing double verification on the first attention verification result and the second attention verification result to finish item-by-item verification backtracking.

Specifically, the time sequence checking sub-channel is a functional module for time consistency checking based on a time sequence checking chain, and the time arrangement and jump rationality of the focusing nodes are realized. The frequency spectrum checking sub-channel is used for a functional module for checking the consistency of frequency spectrum characteristics based on a signal checking chain, and the rationality of characteristic changes such as the frequency and the energy of a focusing node is guaranteed.

In order to complete the successive check backtracking of the fusion preliminary decision, a time sequence check sub-channel is activated first, consistency check is carried out on each data node in a time sequence check chain according to the time evolution rule of the sampling moment and the decision result, and if phenomena such as abnormal time span, abnormal bit jump and the like are detected, the phenomena are marked as a first concern point and the first concern check result is output. For example, during the transmission of logging data, each data node has a corresponding timestamp, and under normal conditions, the time interval between two adjacent data nodes should be Δt, and if a certain interval is actually detected to deviate significantly from Δt, for example, exceeds 1.5 Δt or is less than 0.5 Δt, it is determined that the time consistency check is not passed. At the same time, the time rationality of the change rate of the data is checked, for example, the change rate of a certain geological parameter suddenly exceeds the physical possibility range in a short time, and the change rate is also regarded as a time consistency problem.

And then, activating a spectrum verification sub-channel, analyzing spectrum characteristic evolution conditions such as main frequency, bandwidth, energy and the like node by node based on a signal verification chain, marking the spectrum characteristic evolution conditions as a second attention point if abnormality such as characteristic mutation, continuity fracture and the like is detected, and outputting a second attention verification result. Taking the example of an analog FSK modulated signal in downhole communications, the normal signal spectrum should have a stable dominant frequency location, a reasonable spectral width, and a specific spectral shape. And performing Fast Fourier Transform (FFT) on each traced data node, and extracting the spectrum information of each traced data node. For example, it is checked whether the dominant frequency points fluctuate within an allowable frequency offset range and whether the spectral energy distribution conforms to a preset template. If the main frequency offset of a certain data node is found to be too large or the frequency spectrum has abnormal side lobe growth and other phenomena, judging that the frequency spectrum characteristic verification at the position is not passed.

When the first attention checking result and the second attention checking result are subjected to double checking verification, only when the first attention checking result and the second attention checking result point to the same problem or are mutually evidence, the node is identified to have errors and backtracking correction is performed, otherwise original judgement is kept, and the whole backtracking flow of the item-by-item checking is completed. For example, in a certain log communication data analysis, for 10 retrospective nodes, a timing verification sub-channel detects that a consistency abnormality (a first attention verification result) with a time span shorter than a normal jump period exists between a 6 th node and a 7 th node, and a spectrum verification sub-channel finds that the main frequency drift speed of the 6 th node exceeds a normal change threshold (a second attention verification result), and in double verification, the two abnormal results form complementary verification on node positioning and abnormality types, so that a 6 th node judgment error is confirmed, and correction is performed. In contrast, for node 3, only the spectrum check detects a weak energy anomaly, but the timing check does not find a timing problem, and the anomaly is considered insufficient to trigger correction after double check verification, so original judgement is maintained.

According to the steps, the special time sequence checking sub-channel and the frequency spectrum checking sub-channel are arranged, the consistency detection is independently carried out in a layered and characteristic manner, and finally, the error node is comprehensively judged through double cross verification, so that the accuracy and the robustness of error identification are greatly improved, the error correction risk possibly brought by a single checking method is avoided, and the sensing and repairing capability of tiny abnormality under a complex interference environment is enhanced, so that the high reliability and the stability of the data quality in the communication process of a high-speed well are ensured.

Further, step S300 includes:

And step S310, extracting frequency domain information of the preprocessing result by using the first demodulation module, and performing fast Fourier transform.

And step S320, performing energy detection on the output of the fast Fourier transform to determine a main frequency point with the maximum energy.

And step S330, comparing the main frequency point with a preset frequency threshold value to finish preliminary bit judgment.

And step 340, performing the confidence level identification of the preliminary bit judgment based on the energy value corresponding to the main frequency point.

And step 350, performing preliminary bit decision smoothing based on the confidence level identification based on the energy-time sequence window, and outputting an output result of the first demodulation module.

Specifically, the dominant frequency point refers to the frequency component with the greatest energy, and generally represents the carrier frequency corresponding to the current bit in signal modulation. Confidence identification is an indicator that characterizes the degree of confidence in the preliminary bit decisions, calculated based on the frequency energy intensity. The energy-time sequence window is a sliding analysis area which is jointly limited in energy and time dimensions, and defines an energy change rule in a time range for smoothing decision errors caused by short-time jitter.

The first demodulation module extracts frequency domain information of an input preprocessing result, performs Fast Fourier Transform (FFT) processing on the frequency domain information to obtain frequency spectrum distribution of signals, then performs energy detection on a frequency spectrum curve output by the fast Fourier transform, and searches and determines a frequency point with the strongest energy, namely a main frequency point. And then comparing the extracted main frequency point with a preset frequency threshold value, judging that the main frequency belongs to a high-frequency or low-frequency region, and finishing preliminary bit judgment, for example, the low-frequency corresponds to bit 1 and the high-frequency corresponds to bit 0. Further, according to the energy value corresponding to the main frequency point, a confidence mark is generated, the stronger the energy is, the higher the confidence is, and the confidence is used as an important reference for subsequent processing, and can be realized by establishing a mapping relation in advance, for example, the energy value is divided into different intervals, and each interval corresponds to a confidence value.

And finally, carrying out smoothing processing on the preliminary bit judgment based on the energy-time sequence window and combining with the confidence coefficient identification, wherein the main purpose is to correct isolation judgment jump caused by channel disturbance in a short time and finally output a stable demodulation result of the first demodulation module. An example of the preliminary bit decision smoothing process is that a fixed length sliding window, i.e., an energy-timing window, is first defined, the window length is typically set to 5 to 9 bit decision periods, and can be dynamically adjusted according to system delay and channel stability. And simultaneously recording a preliminary bit judgment value, a main frequency energy value and a confidence coefficient identifier corresponding to each period in the window. In the current sliding window, counting the bit judgment value with the largest occurrence number, and recording the bit judgment value as a main judgment value. If the occurrence proportion of the main decision value exceeds a set threshold value (such as more than 80%), the whole decision of the window is considered to be stable, and the main decision value is output as a smoothing result. If the threshold value is not exceeded, the abnormality determination is entered. If only one or a few (generally one to two) bit decisions in the window are inconsistent with the main decision value, and the confidence levels corresponding to the abnormal points are obviously lower (lower than a preset confidence level threshold value, for example, 60%), the single or few bit decisions are considered to belong to isolated hops, the detected bit decision value of the isolated hopping point is directly replaced by the main decision value, meanwhile, the confidence level is updated to be the average confidence level in the window, and the final decision output of the whole window is recalculated after correction, so that the continuity and the smoothness are ensured. After each time of window smoothing processing is completed, the window slides forward for one bit period, and the same detection and correction are continuously carried out on the data in the new window, so that continuous flow type smoothing judgment is formed. If a plurality of windows continuously appear and cannot form high consistency (for example, continuous 3 windows fail to be judged in consistency), a strong judgment compensation flow is triggered, and backtracking correction is further carried out.

For example, when a section of downhole FSK signal is processed, after fast Fourier transform processing, the frequency of a main frequency point is found to be 17kHz, the energy density is found to be 95%, and a preset frequency threshold is set to be 15kHz, the main frequency point is judged to be bit 0 and marked as high confidence according to 95% of energy intensity, if the judgment results in the subsequent continuous 5 sampling periods are consistent, the judgment results are directly output, and if the main frequency energy is reduced in a certain sampling period and the bit judgment jump occurs, the judgment continuity is maintained by using an energy-time sequence window smoothing algorithm, and bit errors caused by short-time interference are avoided.

According to the method, the frequency domain information is extracted accurately through the fast Fourier transform, bit judgment and confidence assessment are carried out based on the main frequency point with the maximum energy, meanwhile, judgment smoothness is carried out by combining the energy-time sequence window, the accuracy and stability of demodulation are effectively improved, the bit error rate can be remarkably reduced especially in an underground high-noise environment, and the data transmission reliability of the whole high-speed logging communication system is enhanced.

Further, step S300 further includes:

And step S360, carrying out self-adaptive time window segmentation on the preprocessing result, and establishing a time window set.

And step S370, carrying out local spectrum extraction on each time window set.

And step S380, calculating the frequency spectrum variation characteristics of the adjacent time windows, wherein the frequency spectrum variation characteristics comprise a main frequency offset, an energy distribution diffusion characteristic, an energy distribution shrinkage characteristic and a local noise rise rate characteristic.

Step S390, the spectrum variation characteristic combination is used for constructing a local characteristic vector, and bit judgment is carried out to establish an output result of the second demodulation module.

Specifically, the adaptive time window segmentation refers to dynamically determining the time window length according to signal characteristics (such as energy change, time sequence marks and the like), dividing a preprocessing result into a plurality of local time window areas, and facilitating the local extraction and analysis of subsequent spectrum features. First, a preliminary energy analysis is carried out on the preprocessed analog FSK signal, and the signal is divided into a plurality of local time windows by adopting a self-adaptive adjustment time window mechanism according to a set starting threshold. If the current energy level is stable, the time window is kept at a default length (e.g., 100 mus), and if the energy change rate is detected to exceed a preset threshold, the time window length is dynamically shortened (e.g., to 50 mus) to increase the resolution. Finally, a time window set covering the whole signal is obtained, and the length of each time window in the time window set is not identical, so that a foundation is provided for the subsequent local frequency spectrum feature extraction.

Then, within each established time window, a Fast Fourier Transform (FFT) is performed, converting the time domain signal into a frequency domain representation. And extracting the position of a main frequency point and the spectrum energy distribution form of each time window, and recording the change trend index of the local spectrum. To suppress spectral leakage and improve accuracy of spectral features, windowing (e.g., hanning or blackman) is applied to each time window during extraction.

And then, comparing the frequency spectrum data extracted from the two adjacent time windows, and calculating variable characteristic indexes such as a main frequency offset, an energy distribution diffusion characteristic, an energy distribution shrinkage characteristic, a local noise rise rate characteristic and the like. The main frequency offset refers to the difference value between main frequencies of two adjacent time windows and represents the change degree of the main frequency of a signal, the energy distribution diffusion characteristic is represented by a frequency spectrum width increment and reflects the change of the energy dispersion degree on a frequency spectrum, the energy distribution contraction characteristic is represented by a frequency spectrum width contraction and reflects the change of the concentration degree of the energy on the frequency spectrum, the energy distribution contraction characteristic is opposite to the energy distribution diffusion characteristic, the local noise rise rate characteristic is represented by a background energy increase rate of a non-main frequency area and is used for measuring the change speed of noise in the adjacent time windows.

Taking two adjacent time windows as an example, assuming that the main frequency of the former time window is f ₁ and the main frequency of the latter time window is f ₂, the main frequency offset is |f ₂-f₁ |. For the energy distribution spread feature, it can be calculated by comparing the spectral energy distribution of the two time windows. For example, the energy concentration of the preceding time window (the ratio of the energy concentrated in a certain narrow band) and the energy concentration of the following time window are calculated, and if the energy concentration of the following time window is reduced, the energy distribution is considered to have a diffusion tendency, otherwise, the energy distribution has a contraction tendency. The local noise rise characteristic may be obtained by calculating the ratio of the difference in noise intensity over the time interval in adjacent time windows. For example, in a certain frequency band, the noise intensity of the former time window is N ₁, the noise intensity of the latter time window is N ₂, and the time interval is Δt, and the local noise rise rate is characterized by (N ₂-N₁)/Δt. In the characteristic calculation process, normalization processing is adopted, interference of absolute energy magnitude on bit judgment is eliminated, and the characteristic is ensured to reflect relative change trend instead of signal amplitude deviation.

And finally, arranging the main frequency offset, the energy diffusion characteristic, the energy contraction characteristic and the local noise rise rate characteristic obtained by calculation of each time window in a fixed sequence to form a four-dimensional local characteristic vector. And then, classifying each feature vector by using a mode recognition model (such as a K nearest neighbor algorithm, a support vector machine or a lightweight neural network) obtained through training, and judging a bit value corresponding to the current time window. And combining the judgment results of the time windows and outputting the judgment results as a demodulation bit stream of the second demodulation module. The pattern recognition model training process is exemplified by collecting a large number of preprocessed analog FSK signal data samples, and accurately labeling each segment of signal according to a known bit stream. Each sample corresponds to a particular bit tag (e.g., "0" or "1"). And processing each section of acquired signal according to an actual demodulation flow, performing self-adaptive time window segmentation and local frequency spectrum extraction, and calculating four indexes of a main frequency offset, an energy diffusion characteristic, an energy contraction characteristic and a local noise rise rate to form a four-dimensional local characteristic vector. All feature vector data are randomly divided into a training set (such as 80%) and a verification set (such as 20%), so that the training set and the verification set are guaranteed to be representative under different signal-to-noise ratios and different signal conditions. And for the K nearest neighbor algorithm, determining the optimal neighbor number K value, wherein the training process is to store the training set feature vector and the label thereof in a model, and judging based on the nearest neighbor rule when the training set feature vector and the label thereof are classified. And selecting proper kernel functions (such as linear kernel and radial basis kernel RBF) for the support vector machine, performing hyperplane segmentation training, and maximizing class intervals. For a lightweight neural network, designing a small neural network with one or two hidden layers, performing iterative training by adopting an Adam optimizer by using a cross entropy loss function, and adjusting the network weight to minimize training errors. And (5) evaluating indexes such as accuracy, recall rate and the like of the trained model by using the verification set. If the performance is not ideal, the feature selection mode and model parameters (such as k value, kernel function type, network layer number and the like) are adjusted or the training data sample size is increased to perform retraining optimization. And storing the model which is trained and qualified in verification, and performing classification judgment on the real-time feature vector in the subsequent online demodulation process.

Further, step S200 includes:

And S210, when the signal granularity judging result is a first granularity judging result, activating a long-time window smoothing mode, and updating the long-time window smoothing mode into an internal identification mode of the dual-mode FSK, wherein the first granularity judging result is a coarse granularity judging result.

And S220, when the signal granularity judging result is a second granularity judging result, activating a high-confidence tracking mode, and updating the high-confidence tracking mode into an internal identification mode of the dual-mode FSK, wherein the second granularity judging result is a fine granularity judging result.

Specifically, the signal granularity discrimination result refers to the granularity degree of the discrimination signal in the time domain and the frequency domain by analyzing the characteristics of the preprocessed analog FSK signal, and is divided into a coarse granularity discrimination result (i.e., a first granularity discrimination result) and a fine granularity discrimination result (i.e., a second granularity discrimination result).

And (5) performing mode updating judgment according to the signal granularity judgment result obtained in the step S100. When the discrimination result is the first granularity discrimination result (i.e. the coarse granularity discrimination result), the current underground communication link signal is considered to have larger scale change characteristics and poorer stability in short time, so that a long-time window smoothing mode is activated, and the mode is updated to an internal identification mode of the dual-mode FSK. In the long-time window smoothing mode, a larger energy smoothing window is adopted and the time domain accumulated weight is improved, so that interference of random noise on spectrum extraction and bit judgment is restrained, and demodulation stability is improved. For example, for energy detection of a signal, energy values of a plurality of sampling points in a time window are averaged to obtain a smoothed energy value for subsequent main frequency point determination and bit decision. This smoothing process can reduce energy fluctuations due to short-term noise or interference and improve demodulation stability.

When the discrimination result is the second granularity discrimination result (namely the fine granularity discrimination result), the current signal is considered to have good short-time stability and higher local feature legibility, and the high-confidence tracking mode is activated at the moment and is updated to the internal recognition mode of the dual-mode FSK. In the high-confidence tracking mode, the length of an energy smoothing window is reduced, the sensitivity of feature change is increased, and a dynamic confidence computing mechanism is adopted to enhance the response capability to fine main frequency drift and local feature disturbance, so that the real-time judgment precision under high-speed communication is improved.

For example, when the received underground signal is severely fluctuated in a short time and the main frequency point is unstable, a coarse granularity judgment result is output after granularity judgment, the signal is switched to a long-time window smoothing mode, signal characteristic extraction is carried out under an energy smoothing window with the length of 200 mu s, and when the signal shows the characteristics of good continuity, slow change of the main frequency point and lower local noise, a fine granularity judgment result is output, the signal is switched to a high confidence tracking mode, the energy smoothing window is shortened to 50 mu s, and meanwhile dynamic confidence weighting judgment is started to capture the change of each bit finely.

In the demodulation process of the analog FSK modulation signal, the first demodulation module and the second demodulation module work in parallel all the time, and the judgment result of the signal granularity determines a weight distribution mechanism and a confidence fusion mode. And in coarse granularity, the output result of the first demodulation module is mainly used, and the output result of the second demodulation module is used for confidence correction or low weight repetition discrimination. And in fine granularity, the local frequency spectrum dynamic characteristics extracted by the second demodulation module are taken as the main, and the first demodulation module provides candidate support and redundant error correction channels. And giving a fusion preliminary decision after confidence weighting fusion based on output results of the first demodulation module and the second demodulation module.

According to the method, the internal identification mode of the dual-mode FSK is dynamically updated according to the signal granularity discrimination result, so that the demodulation strategy can be adaptively adjusted according to the actual signal state of the communication link, the high stability is maintained under severe channel conditions, the demodulation sensitivity and the demodulation rate are improved under good channel conditions, and the robustness and the overall transmission quality of the high-speed well communication are remarkably improved.

Further, step S100 includes:

step S110, working condition characteristics of the front working condition are obtained, and noise suppression parameters are adaptively matched based on the working condition characteristics.

And step S120, performing noise suppression preprocessing of the analog FSK modulation signal by using the adaptively matched noise suppression parameters.

Specifically, the front-end working condition refers to the current physical environment state in the underground logging communication process, and comprises relative environment parameters such as geological layer type, well depth, lithology, well fluid viscosity, conductivity, vibration intensity and the like which influence signal quality. The working condition features are environmental quantization indexes obtained by the input of a sensor acquisition or regulation end, such as noise background intensity, signal attenuation rate, instantaneous signal to noise ratio and the like, and are used as the basis for matching noise suppression parameters. Firstly, acquiring real-time or near real-time information of a front-end working condition, wherein the information is acquired by a downhole sensing module or transmitted and acquired by a ground system. After the working condition characteristics are extracted from the front-end working condition, the working condition characteristics are compared with a predefined working condition-parameter mapping model, and currently applicable noise suppression parameters including key parameters such as filter bandwidth, cut-off frequency, noise reduction weight, suppression gain threshold and the like are determined according to the matching result. These parameters will then be used in the noise suppression process.

And performing multi-stage noise suppression on the analog FSK signal according to the selected parameter configuration. Common processes include bandpass filtering to preserve the FSK modulation frequency range (e.g., 2kHz-5 kHz) to remove out-of-band interference, spectral subtraction to reduce noise after estimating the noise spectrum to increase the net signal duty cycle, adaptive gain adjustment to automatically adjust the signal strength to avoid feature loss due to too low signal amplitude, and anti-abrupt smoothing to suppress abnormal energy values due to sudden impact or transient spikes.

The steps realize robust adaptation to different underground communication environments by acquiring the working condition characteristics and executing the self-adaptive noise suppression, effectively suppress background noise and environmental interference, improve the signal preprocessing quality, enhance the accuracy of subsequent signal granularity discrimination and demodulation, and provide a solid foundation for dual-mode FSK communication.

Further, step S700 includes:

Step S710, an early warning monitoring index set is established, wherein the early warning monitoring index set comprises a signal strength change index, a main frequency drift rate index, a bit error rate index and a confidence coefficient distribution abnormal index.

And step 720, performing early warning trigger analysis of the analog FSK modulation signal by using the early warning monitoring index set, and reporting an early warning signal.

Specifically, according to the communication stability requirement of the analog FSK modulation signal, a multi-dimensional early warning monitoring index set comprising a signal strength change index, a main frequency drift rate index, a bit error rate index and a confidence coefficient distribution abnormal index is established. The signal strength change index is used for measuring fluctuation conditions of the signal strength and reflects the stability of the signal in the transmission process. The primary frequency drift rate index focuses on the rate of change of the primary frequency of the signal in time, reflecting the stability of the signal frequency. If the dominant frequency drift rate is too high, this means that the signal source is unstable or subject to external interference. The bit error rate index refers to the probability of occurrence of error bits in the digital signal transmission process, directly reflects the reliability of signal transmission, and is one of key indexes for evaluating the communication quality. The confidence coefficient distribution abnormal index is used for monitoring the distribution condition of the confidence coefficient of each bit decision in the signal demodulation process. Normal confidence profiles often have a certain regularity and abnormal profiles may be indicative of potential problems in the demodulation process, such as increased noise interference or signal characteristic variations.

The calculation examples of the early warning monitoring indexes are as follows, for the signal strength change indexes, energy detection needs to be carried out on the input analog FSK signals, in each sliding window, the maximum energy value Emax, the minimum energy value Emin and the root mean square energy Erms are calculated, the energy fluctuation degree delta E=Emax-Emin is defined, and when the delta E exceeds a set fluctuation threshold value or abnormal decline (such as decline exceeds a set percentage) occurs in Erms, the signal strength is judged to be abnormal, and the early warning of the signal strength change is triggered.

For the dominant frequency drift rate index, the dominant frequency point f _main (t) of each time slice is continuously extracted based on a Fast Fourier Transform (FFT). The variation of the main frequency of the adjacent time slices is calculated, the drift amount= -f _main(t+Δt)-f_main (T) | is calculated, then the main frequency drift rate in unit time is calculated, when the drift rate exceeds a set threshold value T _f (for example, 1 kHz/ms), the main frequency is judged to be unstable, and the main frequency drift early warning is triggered.

For the bit error rate index, the statistical bit decision result is compared with the reference bit stream (or by self-consistency backtracking) in the sliding window to obtain the bit error number N _err. The total decision bit number N _total is recorded simultaneously, and the bit error rate is calculated as ber=n _err/N_total. When the BER value continuously exceeds a set threshold (e.g., 1% or a dynamically adaptive threshold), an abnormal bit error rate warning is triggered.

For the confidence distribution anomaly index, a corresponding confidence value C (t) (such as energy difference degree, discrimination probability and the like) is extracted when each bit is judged, the confidence distribution is counted in a sliding window, wherein the confidence distribution comprises a mean value mu _C, a standard deviation sigma _C and a low confidence sample ratio (such as a sample ratio lower than a certain threshold), and then anomaly situations are checked, namely the confidence mean value is reduced by more than a threshold (such as more than 20 percent), the confidence standard difference is increased frequently (such as more than 15 percent), the low confidence sample ratio is increased sharply (such as more than 30 percent), and confidence anomaly early warning is triggered when any anomaly occurs.

And continuously monitoring and analyzing the indexes in real time, performing early warning trigger analysis according to the monitoring analysis result, wherein each early warning index can be independently judged, and a comprehensive early warning rule can be set, for example, an arbitrary single index is continuously abnormal for 3 times or two indexes are simultaneously abnormal for one time, namely, the whole early warning is triggered, and an early warning signal is reported. The early warning signal is used for prompting the current communication quality to be reduced and the potential failure risk to be increased, and supporting the subsequent communication link switching, parameter self-adaptive adjustment or manual intervention operation. In addition, information such as the occurrence time of early warning, specific numerical values of related indexes and the like can be recorded, so that fault detection and analysis can be performed later.

Through early warning trigger analysis and early warning signal report, can in time remind relevant personnel to pay attention to the state of communication link to take corresponding solution measure, help improving reliability and stability of high-speed logging communication, reduce the risk of logging data loss or mistake because of signal quality problem leads to, ensure going on smoothly of logging operation.

In summary, the dual-mode FSK demodulation high-speed well communication method provided by the embodiment of the present application has the following beneficial effects:

according to the embodiment of the application, through a multi-stage and multi-mode parallel processing architecture, the accurate demodulation and real-time early warning of the analog FSK modulation signal under the conditions of strong noise interference and complex working conditions are realized. Firstly, front-end working condition characteristics are acquired, noise self-adaptive parameter matching is carried out, dynamic adaptability of signal preprocessing is achieved, then signal granularity discrimination is introduced to dynamically sense FSK signal characteristics in a complex underground environment, and a long-time window smooth mode under coarse granularity or a high-confidence tracking mode under fine granularity is respectively activated according to signal granularity discrimination results. And running the dual-mode demodulation structure with main frequency-energy demodulation and local frequency spectrum dynamic characteristic extraction in parallel in both modes. The first demodulation module (main frequency-energy demodulation module) rapidly extracts main frequency points and energy through fast Fourier analysis, performs confidence weighting and bit decision smoothing, and outputs a first demodulation result, and the second demodulation module (local frequency spectrum dynamic feature extraction module) generates dynamic feature vectors such as main frequency offset, energy diffusion/contraction, noise rise rate and the like through self-adaptive time window division, local frequency spectrum feature extraction and variable quantity calculation, and outputs a second demodulation result after performing bit decision. And the dual-mode output is subjected to fusion judgment by combining an internal recognition mode, and the bit selection mechanism is dynamically adjusted by combining the noise level, the granularity category and the confidence coefficient, so that the stability under the anti-interference and variable environments is improved, the accuracy of signal recognition is improved, and the overall anti-interference capability is enhanced. On the basis of fusion judgment, in order to cope with short-term identification instability caused by spectrum offset, an adjacent frequency point set is further constructed and used for assisting in judging whether a soft judgment compensation mechanism is triggered, if soft judgment compensation is triggered, the adjacent frequency point set is used for updating fusion preliminary judgment, if soft judgment compensation is not triggered, a consistency check and backtracking correction stage is entered, and a time sequence sub-channel and a spectrum sub-channel are introduced for focus point comparison and correction verification. On the basis, an early warning and monitoring index set which takes signal intensity change, main frequency drift rate, bit error rate and confidence coefficient distribution as cores is established, and rapid early warning and response of demodulation abnormality are realized through sliding window analysis and dynamic threshold judgment.

In the whole, the embodiment of the application effectively enhances the accuracy and reliability of signal demodulation, particularly in a complex underground environment, can obviously reduce the error rate, improves the communication quality and meets the actual requirement of a high-speed well on high-reliability data transmission.

In a second embodiment, as shown in fig. 3, based on the same inventive concept as the previous embodiment, an embodiment of the present application provides a dual-mode FSK demodulation high-speed well communication system, which includes:

The signal granularity discriminating module 10 is configured to receive an analog FSK modulated signal of a downhole communication link, and perform signal granularity discrimination of a preprocessing result after preprocessing the analog FSK modulated signal.

The identification pattern updating module 20 is configured to dynamically update the internal identification pattern of the dual-mode FSK according to the signal granularity discrimination result.

And the result sending module 30 is configured to send the preprocessing result to a first demodulation module and a second demodulation module of the dual-mode FSK, where the first demodulation module is a main frequency-energy demodulation module, and the second demodulation module is a local spectrum dynamic feature extraction module.

And the result fusion module 40 is configured to obtain output results corresponding to the first demodulation module and the second demodulation module, fuse the output results by using the internal recognition mode, and establish a fusion preliminary decision.

The adjacent frequency point searching module 50 is configured to search the adjacent frequency points for the fused preliminary decision, and establish an adjacent frequency point set.

And the trigger analysis module 60 is configured to perform soft-decision trigger analysis based on the fused preliminary decision and the adjacent frequency point set, and if soft-decision compensation is triggered, update the fused preliminary decision with the adjacent frequency point set.

The communication execution module 70 is configured to perform high-speed logging communication by using the updated fusion preliminary decision.

Further, the trigger analysis module 60 according to the embodiment of the present application is further configured to perform the following steps:

If the consistency check fails, the strong decision compensation is triggered, the strong decision compensation is used for carrying out backtracking correction of the fusion preliminary decision, and the backtracking correction result is used for carrying out high-speed logging communication.

And starting from the node triggering strong judgment, carrying out item-by-item check backtracking based on the check chain so as to establish a backtracking correction result.

the method comprises the steps of starting a time sequence checking sub-channel, starting a frequency spectrum checking sub-channel, starting a second attention checking result, and performing double checking verification on the first attention checking result and the second attention checking result to finish item-by-item checking backtracking, wherein the time sequence checking sub-channel is used for performing time consistency checking based on a time sequence checking chain to generate a first attention checking result, the frequency spectrum checking sub-channel is used for performing frequency spectrum characteristic consistency checking based on a signal checking chain to generate a second attention checking result, and the first attention checking result and the second attention checking result are used for performing double checking verification.

Further, the result sending module 30 of the embodiment of the present application is further configured to perform the following steps:

The method comprises the steps of extracting frequency domain information of a preprocessing result by using a first demodulation module, performing fast Fourier transform, performing energy detection on output of the fast Fourier transform, determining a main frequency point with maximum energy, comparing the main frequency point with a preset frequency threshold value to finish preliminary bit judgment, performing confidence level identification of the preliminary bit judgment on the basis of an energy value corresponding to the main frequency point, performing preliminary bit judgment smoothing on the basis of the confidence level identification on the basis of an energy-time sequence window, and outputting an output result of the first demodulation module.

The preprocessing result is subjected to self-adaptive time window segmentation to establish time window sets, local frequency spectrum extraction is carried out on each time window set, frequency spectrum variation characteristics of adjacent time windows are calculated, the frequency spectrum variation characteristics comprise main frequency offset, energy distribution diffusion characteristics, energy distribution shrinkage characteristics and local noise rising rate characteristics, local characteristic vectors are constructed by combining the frequency spectrum variation characteristics, and bit judgment is carried out to establish an output result of the second demodulation module.

Further, the identifying pattern updating module 20 of the embodiment of the present application is further configured to perform the following steps:

and when the signal granularity distinguishing result is a second granularity distinguishing result, activating a high-confidence tracking mode, and updating the high-confidence tracking mode into an internal recognition mode of the dual-mode FSK, wherein the second granularity distinguishing result is a fine granularity distinguishing result.

Further, the signal granularity discriminating module 10 of the embodiment of the present application is further configured to execute the following steps:

And performing noise suppression pretreatment of the analog FSK modulation signal by using the self-adaptively matched noise suppression parameters.

Further, the communication execution module 70 of the embodiment of the present application is further configured to execute the following steps:

And establishing an early warning monitoring index set, wherein the early warning monitoring index set comprises a signal strength change index, a main frequency drift rate index, a bit error rate index and a confidence coefficient distribution abnormal index, and performing early warning trigger analysis of the analog FSK modulation signal by using the early warning monitoring index set to report an early warning signal.

The foregoing detailed description of a dual-mode FSK demodulation high-speed well communication method will be clear to those skilled in the art, and the system disclosed in the second embodiment has corresponding functional modules and beneficial effects as far as the system disclosed in the first embodiment corresponds to the method disclosed in the first embodiment, and relevant points refer to the description of the method section.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A dual-mode FSK demodulated high-speed well communication method, the method comprising:

Receiving an analog FSK modulation signal of a downhole communication link, preprocessing the analog FSK modulation signal, and then judging the signal granularity of a preprocessing result;

Dynamically updating an internal recognition mode of the dual-mode FSK by using a signal granularity discrimination result, comprising:

when the signal granularity judging result is a first granularity judging result, activating a long-time window smoothing mode, and updating the long-time window smoothing mode into an internal identification mode of the dual-mode FSK, wherein the first granularity judging result is a coarse granularity judging result;

When the signal granularity judging result is a second granularity judging result, activating a high-confidence tracking mode, and updating the high-confidence tracking mode into an internal identification mode of the dual-mode FSK, wherein the second granularity judging result is a fine granularity judging result;

The preprocessing result is respectively sent to a first demodulation module and a second demodulation module of the dual-mode FSK, wherein the first demodulation module is a main frequency-energy demodulation module, and the second demodulation module is a local frequency spectrum dynamic characteristic extraction module;

respectively obtaining output results corresponding to a first demodulation module and a second demodulation module, and fusing the output results by utilizing the internal recognition mode to establish fusion preliminary judgment;

searching adjacent frequency points for the fusion preliminary judgment, and establishing an adjacent frequency point set;

performing soft decision triggering analysis based on the fusion preliminary decision and the adjacent frequency point set, and if soft decision compensation is triggered, updating the fusion preliminary decision by using the adjacent frequency point set;

and carrying out high-speed logging communication by utilizing the updated fusion preliminary judgment.

2. The dual-mode FSK demodulated high speed well communication method of claim 1 wherein said soft decision trigger analysis based on said fused preliminary decision and said set of adjacent frequency points comprises:

if the soft decision compensation is not triggered, establishing a primary candidate bit stream based on the fusion primary decision;

checking the consistency of the preliminary candidate bit stream by utilizing a sliding window;

And if the consistency check fails, triggering strong decision compensation, performing backtracking correction of fusing the preliminary decisions by using the strong decision compensation, and performing high-speed logging communication by using a backtracking correction result.

3. The dual-mode FSK demodulated high speed well communication method of claim 2 wherein said performing a backtracking correction of a fused preliminary decision using said strong decision compensation comprises:

In the backtracking verification process, backtracking a plurality of data nodes, and establishing a verification chain, wherein the verification chain comprises a time sequence verification chain and a signal verification chain;

And starting from the node triggering the strong judgment, performing item-by-item check backtracking based on the check chain so as to establish the backtracking correction result.

4. A dual-mode FSK demodulated high speed borehole communication method according to claim 3, wherein said performing a itemized check backtrack based on said check chain comprises:

activating a time sequence checking sub-channel, and performing time consistency check based on a time sequence checking chain by using the time sequence checking sub-channel to generate a first attention checking result;

Activating a frequency spectrum verification sub-channel, and performing frequency spectrum feature consistency verification based on a signal verification chain by using the frequency spectrum verification sub-channel to generate a second attention verification result;

And performing double verification on the first attention verification result and the second attention verification result to finish item-by-item verification backtracking.

5. The method for high-speed well logging communication by dual-mode FSK demodulation according to claim 1, wherein said transmitting said preprocessing results to said first demodulation module and said second demodulation module of dual-mode FSK, respectively, comprises:

extracting frequency domain information of the preprocessing result by using the first demodulation module, and performing fast Fourier transform;

performing energy detection on the output of the fast Fourier transform, and determining a main frequency point with the maximum energy;

comparing the main frequency point with a preset frequency threshold value to finish preliminary bit judgment;

performing confidence level identification of preliminary bit judgment based on the energy value corresponding to the main frequency point;

And performing preliminary bit decision smoothing based on confidence level identification based on the energy-time sequence window, and outputting an output result of the first demodulation module.

6. The method for high-speed well communication by dual-mode FSK demodulation according to claim 5, wherein said sending the preprocessing result to the first demodulation module and the second demodulation module of the dual-mode FSK, respectively, further comprises:

Performing self-adaptive time window segmentation on the preprocessing result, and establishing a time window set;

Extracting local frequency spectrum of each time window set;

Calculating the frequency spectrum variation characteristics of adjacent time windows, wherein the frequency spectrum variation characteristics comprise a main frequency offset, an energy distribution diffusion characteristic, an energy distribution shrinkage characteristic and a local noise rise rate characteristic;

And combining the spectrum variation characteristics to construct a local characteristic vector, and executing bit judgment to establish an output result of the second demodulation module.

7. A dual-mode FSK demodulated high speed well communication method according to claim 1 wherein said preprocessing said analog FSK modulated signal comprises:

Acquiring working condition characteristics of a front-end working condition, and adaptively matching noise suppression parameters based on the working condition characteristics;

and performing noise suppression pretreatment on the analog FSK modulation signal by using the self-adaptive matched noise suppression parameters.

8. The dual-mode FSK demodulated high-speed logging communication method of claim 1, wherein said high-speed logging communication using updated fused preliminary decisions comprises:

establishing an early warning monitoring index set, wherein the early warning monitoring index set comprises a signal strength change index, a main frequency drift rate index, a bit error rate index and a confidence coefficient distribution abnormal index;

And carrying out early warning triggering analysis of the analog FSK modulation signal by using the early warning monitoring index set, and reporting an early warning signal.

9. A dual-mode FSK demodulated high speed well communication system for performing a dual-mode FSK demodulated high speed well communication method according to any of claims 1-8, comprising:

the signal granularity judging module is used for receiving the analog FSK modulation signal of the underground communication link and executing the signal granularity judgment of the pretreatment result after the analog FSK modulation signal is pretreated;

The recognition mode updating module is used for dynamically updating the internal recognition mode of the dual-mode FSK by utilizing the signal granularity discrimination result;

The result sending module is used for respectively sending the preprocessing result to a first demodulation module and a second demodulation module of the dual-mode FSK, wherein the first demodulation module is a main frequency-energy demodulation module, and the second demodulation module is a local frequency spectrum dynamic characteristic extraction module;

the result fusion module is used for respectively acquiring output results corresponding to the first demodulation module and the second demodulation module, and utilizing the internal recognition mode to fuse the output results so as to establish fusion preliminary judgment;

The adjacent frequency point searching module is used for searching the adjacent frequency points for the fusion preliminary judgment and establishing an adjacent frequency point set;

the trigger analysis module is used for carrying out soft decision trigger analysis based on the fusion preliminary decision and the adjacent frequency point set, and if soft decision compensation is triggered, the fusion preliminary decision is updated by utilizing the adjacent frequency point set;

And the communication execution module is used for carrying out high-speed logging communication by utilizing the updated fusion preliminary judgment.