Multi-modal joint-based high-speed transmission system and method for logging while drilling

Abstract

The application provides a logging-while-drilling high-speed transmission system and method based on multi-modal combination, belongs to the technical field of oil exploration and logging-while-drilling, and comprises a downhole acquisition unit, a transmission unit and a ground decoding unit. The downhole acquisition unit continuously acquires multiple types of formation parameters during the drilling process of a drilling tool and generates standardized data frames. The transmission unit selects different transmission modes to transmit the standardized data frames in real time. The ground decoding unit comprises an independent decoding layer, a redundancy check layer and a fusion layer. The independent decoding layer performs wavelet transformation on the standardized data frames to separate out mode signals corresponding to each transmission mode. The redundancy check layer marks conflict data with low consistency by calculating the correlation between different mode signals, and repairs the conflict data. The fusion layer constructs a probability fusion model based on a Bayesian network, and inputs the output data of the redundancy check layer into the probability fusion model to obtain decoded formation parameters. The application can effectively improve the efficient and reliable transmission of downhole data.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of oil exploration and logging while drilling, and particularly relates to a logging while drilling high-speed transmission system and method based on multi-modal combination. BACKGROUND

[0002] The logging while drilling technology is a key technology for real-time acquisition of formation parameters in the process of oil drilling, and the core is efficient transmission of multi-modal logging data, such as resistivity, acoustic wave, gamma, pressure and other parameters. The traditional drilling transmission system mainly adopts single transmission mode of mud pulse or cable transmission, and has problems of low transmission rate and weak anti-interference ability.

[0003] For example, a mud pulse signal denoising method, device, equipment and storage medium disclosed in Chinese patent No. 202411594200.4 constructs a linear mud pulse signal model, and denoises the mud pulse signal based on a standard Kalman filtering algorithm, but because the mud pulse signal is disturbed by nonlinear interference, the fitting degree of the linear model is low, and the standard Kalman filtering algorithm is difficult to adapt to a variable noise environment and has low filtering precision for nonlinear interference, resulting in poor denoising effect.

[0004] For example, a logging while drilling mud pressure wave signal processing method, system, terminal and medium disclosed in Chinese patent No. 202311465807.8 first collects signals as noise samples and obtains frequency domain characteristics when the mud pump is started, collects pressure fluctuation signals and calculates frequency spectrum characteristics when the downhole instrument works, then compares the frequency domain characteristic points of the two to select a bandpass or bandstop filter to filter out noise frequencies, and finally restarts the mud pump, decodes the filtered pressure wave signal to complete the processing work, and repeats the above steps if the filtering effect is poor or the operation parameters change. However, in the logging while drilling pressure wave transmission technology, the mud pressure wave signal is affected by many factors, and the signals collected on the ground are disturbed by mud pump noise and the like.

[0005] In addition, the denoising in the prior art is usually based on preset filter / signal processing method parameters, and engineers adjust the parameters based on experience, resulting in low signal recognition success rate and poor adaptability in different operation blocks. SUMMARY

[0006] The purpose of the present application is to provide a logging while drilling high-speed transmission system and method based on multi-modal combination, which can effectively improve the efficient and reliable transmission of downhole data.

[0007] The present application is implemented by the following technical solutions:

[0008] The multi-modal joint-based logging-while-drilling high-speed transmission system comprises a downhole acquisition unit, a transmission unit and a ground decoding unit. The downhole acquisition unit continuously acquires multiple types of formation parameters during the drilling process of the drilling tool and generates standardized data frames. The transmission unit selects different transmission modes according to the type of formation parameters, the well depth and the mud viscosity to transmit the standardized data frames in real time. The ground decoding unit comprises an independent decoding layer, a redundancy checking layer and a fusion layer. The independent decoding layer performs wavelet transform on the received standardized data frames to separate the modal signals corresponding to each transmission mode. The redundancy checking layer marks the conflict data with low consistency by calculating the correlation between different modal signals, and repairs the conflict data. The fusion layer constructs a probability fusion model based on a Bayesian network, and inputs the output data of the redundancy checking layer into the probability fusion model to obtain the decoded formation parameters. The transmission modes include mud pulse mode, electromagnetic wave mode and optical fiber mode.

[0009] Further, the downhole acquisition unit comprises an ultrasonic imaging device, a resistivity imaging device and conventional logging instruments installed in the drilling tool. The formation parameters acquired by the downhole acquisition unit include ultrasonic imaging data, resistivity imaging data and conventional logging data. The conventional logging instruments include a natural gamma logging instrument, a spontaneous potential logging instrument, an acoustic logging instrument, a density logging instrument, a resistivity logging instrument and a caliper logging instrument. The conventional logging data includes natural gamma related to lithology, spontaneous potential related to lithology and permeability, acoustic travel time related to porosity and permeability, formation density related to porosity, formation resistivity related to oil and water layers, and borehole diameter related to wellbore conditions.

[0010] Further, the transmission unit selects different transmission modes according to the type of formation parameters, the well depth and the mud viscosity to transmit the formation parameters in real time, specifically: when the well depth is less than a set well depth threshold and the mud viscosity is less than a set viscosity threshold, the transmission unit selects the electromagnetic wave mode and the optical fiber mode for joint transmission; otherwise, the transmission unit selects the mud pulse mode and the optical fiber mode for joint transmission. When the electromagnetic wave mode and the optical fiber mode are selected for joint transmission, for ultrasonic imaging data and resistivity imaging data, at least 80% of the total data amount of ultrasonic imaging data and 80% of the total data amount of resistivity imaging data are transmitted by the optical fiber mode, and the remaining data is transmitted by the electromagnetic wave mode; for conventional logging parameters, at least 60% of the conventional logging data is transmitted by the optical fiber mode, and the remaining data is transmitted by the electromagnetic wave mode. When the mud pulse mode and the optical fiber mode are selected for joint transmission, for ultrasonic imaging data and resistivity imaging data, at least 80% of the total data amount of ultrasonic imaging data and 80% of the total data amount of resistivity imaging data are transmitted by the optical fiber mode, and the remaining data is transmitted by the mud pulse mode; for conventional logging parameters, at least 60% of the conventional logging data is transmitted by the optical fiber mode, and the remaining data is transmitted by the mud pulse mode.

[0011] Further, the downhole acquisition unit embeds different identification fields in the header of the standardized data frames corresponding to the ultrasonic imaging data, resistivity imaging data and conventional logging data, and the transmission unit judges the category of the transmitted data according to the identification fields.

[0012] Further, the transmission unit comprises a mud pulse modulator, an electromagnetic wave coupler and an optical fiber slip ring, the mud pulse modulator is used to encode data into pressure waves and form a mud pulse mode with a low frequency transmission channel of 0-10 kHz, the electromagnetic wave coupler is used to encode data into electromagnetic waves and form an electromagnetic wave mode with a medium frequency transmission channel of 10 kHz-1 MHz, and the optical fiber slip ring is used to encode data into optical signals and form an optical fiber mode with a high frequency optical fiber channel.

[0013] Further, the independent decoding layer calculates the modal information according to the formula x . t a is a scale parameter, b is a translation parameter, is a mother wavelet function, when a >100, the modal signal corresponding to the mud pulse mode is obtained , when 10 a ≤100, the modal signal corresponding to the electromagnetic wave mode is obtained , and when a <10, the modal signal corresponding to the optical fiber mode is obtained .

[0014] Further, the redundancy checking layer calculates the mutual information between the modal information and the modal information S i . S j If I ( S i , S j ) < information threshold value, it is determined that the modal information S i and the modal information S j are conflict data, the correlation between the modal information and the modal information S i is calculated according to the formula S j , if C ij < correlation threshold value, the modal information S i and the modal information​S j marking the conflict data points as low consistency, otherwise not marking, wherein, i , j i=1,2,3 and i ≠ j , is the calculation information entropy, is the modal information S i the i-th sample data of the k th sample data, N is the modal information S i the number of sample data of the indicates returning 1 when , otherwise returning 0, 1≤ k ≤ N .

[0015] Further, the redundancy checking layer further comprises an adaptive equalizer, the marked conflict data points are input to the adaptive equalizer, the adaptive equalizer adopts a least mean square error algorithm, and a step factor of the least mean square error algorithm is dynamically adjusted according to an error between an output and an input of the adaptive equalizer, so as to repair the conflict data, and the repaired data and the data not marked are jointly used as new modal information .

[0016] Further, the probability fusion model of the fusion layer is represented as , wherein, is the decoded formation parameter, is a likelihood function, is a prior probability of the known formation parameter, is weighted by a conditional probability of the modal information when the formation parameter is D, and a weighting coefficient is a reliability weight of the known modal information.

[0017] The present application also realizes the following technical solutions:

[0018] The transmission method of the multi-modal joint based drilling while logging high-speed transmission system according to any one of the above, comprising the following steps:

[0019] Step S1: continuously collecting multiple formation parameters during the process of drilling the drilling tool;

[0020] Step S2: selecting different transmission modalities according to the formation parameter category, the well depth and the mud viscosity to transmit the formation parameters to the ground in real time;

[0021] Step S3: the ground first performs wavelet transform on the received formation parameters to separate the modal signals corresponding to each transmission mode, then marks the conflict data with low consistency by calculating the correlation between different modal signals, repairs the conflict data, and finally constructs a probability fusion model based on a Bayesian network, inputs the repaired data and the data not marked into the probability fusion model to obtain the decoded formation parameters, wherein the transmission modes include mud pulse mode, electromagnetic wave mode and optical fiber mode.

[0022] The present application has the following advantages:

[0023] 1、The downhole acquisition unit of the present application continuously acquires multiple types of formation parameters during the drilling process of the drilling tool and generates standardized data frames, the transmission unit selects different transmission modes according to the type of formation parameters, well depth and mud viscosity to transmit the standardized data frames in real time, which can realize efficient collaboration of multi-modal transmission, effectively improve the transmission efficiency, the ground decoding unit includes an independent decoding layer, a redundancy check layer and a fusion layer, the independent decoding layer performs wavelet transform on the received standardized data frames to separate the modal signals corresponding to each transmission mode, the redundancy check layer marks the conflict data with low consistency by calculating the correlation between different modal signals, and repairs the conflict data, the fusion layer constructs a probability fusion model based on a Bayesian network, and the output data of the redundancy check layer is input into the probability fusion model to obtain the decoded formation parameters, which can significantly reduce the influence of the complex downhole environment on the transmission signal, reduce the bit error rate by 2~3 orders of magnitude compared with the traditional system, and support high-definition imaging, real-time logging and other multi-scene applications, and improve the intelligent level of the logging while drilling. BRIEF DESCRIPTION OF DRAWINGS

[0024] The present application will be further described in detail below with reference to the accompanying drawings.

[0025] Figure 1 It is a structural schematic diagram of the system of the present application.

[0026] Figure 2 It is a flowchart of the method of the present application.

[0027] Among them, 1, resistivity imaging device; 2, ultrasonic imaging device; 3, conventional logging instrument; 4, conditioning circuit; 5, modal switching controller; 6, mud pulse modulator; 7, electromagnetic wave coupler; 8, optical fiber slip ring; 9, ground decoding unit; 10, main control computer; 11, drill string; 12, drilling fluid circulation system. DETAILED DESCRIPTION

[0028] As Figure 1As shown, the multi-modal joint-based logging-while-drilling high-speed transmission system includes a downhole acquisition unit, a transmission unit, a ground decoding unit 9 and a display unit. The ground decoding unit 9 is implemented in the host computer 10, and the output data of the ground decoding unit 9 is displayed and stored by the host computer 10. The downhole acquisition unit continuously acquires multiple types of formation parameters during the drilling process of the drilling tool and generates standardized data frames. The transmission unit selects different transmission modes according to the type of formation parameters, the depth of the well and the mud viscosity to transmit the standardized data frames in real time. The ground decoding unit 9 includes an independent decoding layer, a redundancy check layer and a fusion layer. The independent decoding layer performs wavelet transform on the received standardized data frames to separate the modal signals corresponding to each transmission mode. The redundancy check layer marks the conflict data with low consistency by calculating the correlation between different modal signals, and repairs the conflict data. The fusion layer constructs a probability fusion model based on a Bayesian network, and the output data of the redundancy check layer is input into the probability fusion model to obtain decoded formation parameters through joint decoding. The transmission modes include mud pulse mode, electromagnetic wave mode and optical fiber mode.

[0029] The drilling tool includes a vertical drill string 11 and a drilling fluid circulation system 12 connected to the drill string 11. The specific structure of the drilling tool is prior art. The downhole acquisition unit includes an ultrasonic imaging device 2, a resistivity imaging device 1 and a conventional logging instrument 3 installed in the drill string 11, and a conditioning circuit 4 connected to the ultrasonic imaging device 2, the resistivity imaging device 1 and the conventional logging instrument 3, respectively. The formation parameters acquired by the downhole acquisition unit include ultrasonic imaging data, resistivity imaging data and conventional logging data. The conventional logging instrument 3 includes a natural gamma ray instrument, a spontaneous potential logging instrument, an acoustic logging instrument, a density logging instrument, a resistivity logging instrument and a caliper logging instrument. The conventional logging data includes natural gamma ray, spontaneous potential, acoustic time difference, formation density, formation resistivity and borehole diameter. The natural gamma ray is the gamma ray released by the radioactive elements in the formation captured by the natural gamma ray instrument, which can assist in dividing the lithology. The spontaneous potential is the potential difference between the formation and the drilling fluid measured by the spontaneous potential logging instrument, which can assist in determining the lithology and permeability. The acoustic time difference can be related to porosity and lithology. The formation density is measured by the density logging instrument using gamma scattering, which can calculate the porosity. The formation resistivity can identify oil and water layers, and the borehole diameter can reflect the wellbore condition. The downhole acquisition unit also integrates existing temperature and vibration compensation modules to support stable data acquisition in environments of -20℃ to 170℃ and 0 to 20000g. The conditioning circuit 4 uses an NI9237 signal conditioning module to support multi-channel analog signal conditioning and can amplify, filter and level convert the weak signals output by the ultrasonic imaging, resistivity imaging and other devices to meet the logging data preprocessing requirements. The ultrasonic imaging device 2 uses the UIS ultrasonic imaging logging system of CNOOC, and the resistivity imaging device 1 uses the array resistivity logging module of Jerey Company.

[0030] The conditioning circuit 4 performs preprocessing, i.e., converting the ultrasonic imaging data, resistivity imaging data and conventional logging data into standardized data frames. Specifically, the NI 9237 signal conditioning module first filters the ultrasonic imaging data, resistivity imaging data and conventional logging data with its built-in anti-aliasing filter, then amplifies the filtered data with a programmable gain amplifier, and then outputs the amplified data to the digital processing unit in digital form after being converted by a 24-bit ADC. The digital processing unit then identifies the data type according to the channel and maps it to a text-defined identification field, such as converting ultrasonic imaging data 0001 to 0x01, resistivity imaging data 0010 to 0x02, and conventional logging data 0100 to 0x04. The data is then spliced in a fixed format of frame header 0xAA55 + identification field + data length + 9237 output digital data body + CRC16 check bit, to generate a standardized data frame containing synchronization identification, type differentiation, data body and error detection function. Finally, the standardized data frame is output to the transmission unit, completing the process from analog signal conditioning to standardized frame packaging. This process is prior art.

[0031] After preprocessing, the conditioning circuit 4 embeds different identification fields in the header of the standardized data frame. Specifically, the identification field corresponding to the ultrasonic imaging data is "0001", the identification field corresponding to the resistivity imaging data is "0010", and the identification field corresponding to the conventional logging data is "0100".

[0032] The transmission unit includes a mud pulse modulator 6, an electromagnetic wave coupler 7, an optical fiber slip ring 8 and a mode switching controller 5. The mud pulse modulator 6 is used to encode the preprocessed standardized data frame into a pressure wave and form a mud pulse mode with a low frequency transmission channel of 0-10 kHz. The electromagnetic wave coupler 7 is used to encode the preprocessed standardized data frame into an electromagnetic wave and form an electromagnetic wave mode with a medium frequency transmission channel of 10 kHz-1 MHz. The optical fiber slip ring 8 is used to encode the preprocessed standardized data frame into an optical signal and form an optical fiber mode with a high frequency optical fiber channel.

[0033] The mode switching controller 5 of the transmission unit selects different transmission modes to transmit the formation parameters in real time according to the formation parameter category, well depth and mud viscosity. Specifically, when the well depth is less than the set well depth threshold and the mud viscosity is less than the set viscosity threshold, the transmission unit selects the electromagnetic wave mode and the optical fiber mode for joint transmission; otherwise, the transmission unit selects the mud pulse mode and the optical fiber mode for joint transmission. The mode switching controller 5 determines the category of the transmitted data through the identification field, and allocates the transmission ratio according to the data category under the joint transmission mode selected by the transmission unit. Specifically, when the electromagnetic wave mode and the optical fiber mode are selected for joint transmission, for ultrasonic imaging data and resistivity imaging data, the optical fiber mode is used to transmit at least 80% of the total ultrasonic imaging data and 80% of the total resistivity imaging data, and the electromagnetic wave mode is used to transmit the remaining data. For conventional logging parameters, the optical fiber mode is used to transmit at least 60% of the conventional logging data, and the electromagnetic wave mode is used to transmit the remaining data. When the mud pulse mode and the optical fiber mode are selected for joint transmission, for ultrasonic imaging data and resistivity imaging data, the optical fiber mode is used to transmit at least 80% of the total ultrasonic imaging data and 80% of the total resistivity imaging data, and the mud pulse mode is used to transmit the remaining data. For conventional logging parameters, the optical fiber mode is used to transmit at least 60% of the conventional logging data, and the mud pulse mode is used to transmit the remaining data.

[0034] The depth encoder on the drill string 11 can directly measure the well depth, and the mud viscosity is detected by a high-temperature and high-pressure mud viscometer installed on the drill string 11. The well depth threshold is set to 3000m, and the viscosity threshold is set to 50mPa.s.

[0035] The mode switching controller 5 is located in the instrument sub in the middle of the drill string 11. The mud pulse modulator 6 is installed at the lower portion of the drill string 11, near the pumping system. The electromagnetic wave coupler 7 surrounds or is embedded in the drill string 11. The fiber optic slip ring 8 is located at the top of the drill string 11, where it connects to the surface top drive system. A rotary joint connects the dynamic drill string 11 to the static surface optical fiber. Each component is designed as a cylindrical unit that adapts to the structure of the drill string 11, ensuring stable operation in high-temperature, high-pressure, and high-vibration environments.

[0036] The ground decoding unit 9 independent decoding layer according to the formula The received formation parameters x ( t ) to perform wavelet transform, a is the scale parameter, b is the translation parameter, As the mother wavelet function, the modal signals corresponding to the three transmission modes are separated by setting the characteristic scale intervals of mud pulse (low frequency), electromagnetic wave (medium frequency), and optical fiber signal (high frequency). a When >100, the modal signal corresponding to the mud pulse mode is obtained , when setting 10< a When ≤100, the modal signal corresponding to the electromagnetic wave mode is obtained , when setting a When <10, the modal signal corresponding to the optical fiber mode is obtained , which can eliminate inter-modal crosstalk.

[0037] The redundant check layer is based on the formula Calculating modal information S i With modal information S j If the mutual information between I ( S i , S j )<information threshold, then the modal information is determined S i With modal information S j For conflicting data, according to the formula Calculating modal information S i With modal information S j The correlation, if C ij < correlation threshold, the modal information S i With modal information S j Conflicting data points are marked as having low consistency, otherwise they are not marked. i , j =1,2,3 and i ≠ j , To calculate information entropy, For modal information S i No. k Sample data, N For modal information S i The number of sample data, Indicates when Returns 1 if yes, otherwise returns 0, 1≤ k ≤ N Among them, the information threshold is set to 0.2 and the correlation threshold is set to 90%.

[0038] The redundancy check layer also includes an adaptive equalizer. The marked conflicting data points are used as input to the adaptive equalizer. The adaptive equalizer uses the minimum mean square error algorithm. The step size factor of the minimum mean square error algorithm is dynamically adjusted according to the error between the output and input of the adaptive equalizer to repair the conflicting data. The repaired data and the unmarked data are used together as the new modal information. .

[0039] More specifically, let the input signal of the adaptive equalizer be , the expected output signal is , the filter coefficient vector is , the iterative update formula is , , is the step factor. When the mud bubble noise breaks out, the error When the amplitude and power increase significantly, the instantaneous increase , let the filter coefficient vector Rapid update, forcibly catch up with the time-varying channel, and gradually reduce the error after the error is reduced due to parameter adjustment and the system approaches a steady state , fine-tune with small steps to avoid parameter oscillation caused by instantaneous fluctuation of mud bubbles. Excessive power, strong mud pulse interference, temporary constraints The upper limit prevents the equalizer from diverging due to excessive step sizes. In the early stages of iteration, a larger base step size and error feedback acceleration are proactively set to allow the equalizer to quickly capture channel characteristics. In the later stages of iteration, a smaller step size and error smoothing are used to stabilize noise compensation.

[0040] The probability fusion model of the fusion layer is expressed as ,in, is the decoded formation parameter, is the likelihood function, is the prior probability of the known formation parameters, The modal information when the formation parameter is D The conditional probability of The weighted coefficient is the reliability weight of each known modal information. In this embodiment, the reliability weight is set to: 0.6 for the optical fiber mode, 0.4 for the electromagnetic wave mode, and 0.4 for the mud pulse mode.

[0041] Table 1 shows the performance comparison between the present invention and traditional mud pulse transmission:

[0042] Table 1

[0043]

[0044] like Figure 2As shown, the transmission method of the logging-while-drilling high-speed transmission system based on multi-modal joint includes the following steps:

[0045] Step S1: continuously collecting multiple types of formation parameters during the drilling process of the drilling tool;

[0046] Step S2: selecting different transmission modes according to the formation parameter type, well depth, and mud viscosity to transmit the formation parameters to the ground in real time;

[0047] Step S3: the ground first performs wavelet transform on the received formation parameters to separate the modal signals corresponding to different transmission modes, then marks conflict data with low consistency by calculating the correlation between different modal signals, repairs the conflict data, finally constructs a probability fusion model based on a Bayesian network, and inputs the repaired data and the data not marked into the probability fusion model to obtain decoded formation parameters, wherein the transmission modes include mud pulse mode, electromagnetic wave mode, and optical fiber mode.

[0048] The above is only a preferred embodiment of the present application, and therefore cannot limit the scope of the present application. Equivalent changes and modifications made in accordance with the scope of the present application and the content of the specification should still be within the scope of the present application.

Claims

1. A high-speed transmission system for logging while drilling based on multi-modal combination, characterized by: It includes a downhole acquisition unit, a transmission unit and a ground decoding unit. The downhole acquisition unit continuously collects multiple types of formation parameters and generates standardized data frames during the drilling process of the drill bit. The transmission unit selects different transmission modes according to the formation parameter category, well depth and mud viscosity to transmit the standardized data frames in real time. The ground decoding unit includes an independent decoding layer, a redundant check layer and a fusion layer. The independent decoding layer performs wavelet transform on the received standardized data frames to separate the modal signals corresponding to each transmission mode. The redundant check layer calculates the correlation between different modal signals, marks conflicting data with low consistency, and repairs the conflicting data. The fusion layer constructs a probabilistic fusion model based on the Bayesian network. The output data of the redundant check layer is input into the probabilistic fusion model to obtain the decoded formation parameters, wherein the transmission modes include mud pulse mode, electromagnetic wave mode and optical fiber mode. The independent decoding layer is based on the formula The received formation parameters x ( t ) to perform wavelet transform, a is the scale parameter, b is the translation parameter, is the mother wavelet function, when setting a When >100, the modal signal corresponding to the mud pulse mode is obtained , when setting 10< a When ≤100, the modal signal corresponding to the electromagnetic wave mode is obtained , when setting a When <10, the modal signal corresponding to the optical fiber mode is obtained ; The redundancy check layer is based on the formula Calculating modal information S i With modal information S j If the mutual information between I ( S i , S j )<information threshold, then the modal information is determined S i With modal information S j For conflicting data, according to the formula Calculating modal information S i With modal information S j The correlation, if C ij < correlation threshold, the modal information S i With modal information S j Conflicting data points are marked as having low consistency, otherwise they are not marked. i , j =1,2,3 and i ≠ j , To calculate information entropy, For modal information S i No. k Sample data, N Modal information S i The number of sample data, Indicates when Returns 1 if yes, otherwise returns 0, 1≤ k ≤ N ; The redundancy check layer also includes an adaptive equalizer. The marked conflict data points are used as input to the adaptive equalizer. The adaptive equalizer uses a minimum mean square error algorithm. The step size factor of the minimum mean square error algorithm is dynamically adjusted according to the error between the output and input of the adaptive equalizer to repair the conflicting data. The repaired data and the unmarked data are used together as new modal information. ; The probabilistic fusion model of the fusion layer is expressed as ,in, is the decoded formation parameter, is the likelihood function, is the prior probability of the known formation parameters, The modal information when the formation parameter is D The conditional probability of The weighted coefficient is the reliability weight of each known modal information.

2. The high-speed transmission system for logging while drilling based on multi-modal combination according to claim 1, characterized in that: The downhole acquisition unit includes an ultrasonic imaging device, a resistivity imaging device and a conventional logging instrument installed in the drilling tool. The formation parameters acquired by the downhole acquisition unit include ultrasonic imaging data, resistivity imaging data and conventional logging data. The conventional logging instrument includes a natural gamma ray meter, a natural potential logging instrument, an acoustic logging instrument, a density logging instrument, a resistivity logging instrument and a caliper logging instrument. The conventional logging data includes natural gamma related to lithology, natural potential related to lithology and permeability, acoustic wave time difference related to porosity and permeability, formation density related to porosity, formation resistivity related to oil and water layers, and wellbore diameter related to well wall conditions.

3. The high-speed transmission system for logging while drilling based on multi-modal combination according to claim 2, characterized in that: The transmission unit selects different transmission modes to transmit the formation parameters in real time according to the formation parameter type, well depth and mud viscosity, specifically including: when the well depth is less than a set well depth threshold and the mud viscosity is less than a set viscosity threshold, the transmission unit selects the electromagnetic wave mode and the optical fiber mode for joint transmission; otherwise, the transmission unit selects the mud pulse mode and the optical fiber mode for joint transmission; when the electromagnetic wave mode and the optical fiber mode are selected for joint transmission, for the ultrasonic imaging data and the resistivity imaging data, at least 80% of the total data volume of the ultrasonic imaging data and 80% of the total data volume of the resistivity imaging data are transmitted using the optical fiber mode; The remaining data is transmitted using the electromagnetic wave mode. For conventional logging parameters, at least 60% of the conventional logging data is transmitted using the optical fiber mode, and the remaining data is transmitted using the electromagnetic wave mode. When the mud pulse mode and the optical fiber mode are jointly transmitted, for ultrasonic imaging data and resistivity imaging data, at least 80% of the total ultrasonic imaging data and 80% of the total resistivity imaging data are transmitted using the optical fiber mode, and the remaining data is transmitted using the mud pulse mode. For conventional logging parameters, at least 60% of the conventional logging data is transmitted using the optical fiber mode, and the remaining data is transmitted using the mud pulse mode.

4. The high-speed transmission system for logging while drilling based on multi-modal combination according to claim 3 is characterized in that: The downhole acquisition unit embeds different identification fields in the headers of standardized data frames corresponding to ultrasonic imaging data, resistivity imaging data and conventional logging data, and the transmission unit determines the category of the transmitted data according to the identification fields.

5. The high-speed transmission system for logging while drilling based on multi-modal combination according to any one of claims 1 to 4, characterized in that: The transmission unit includes a mud pulse modulator, an electromagnetic wave coupler and an optical fiber slip ring. The mud pulse modulator is used to encode data into pressure waves and form a mud pulse mode with a low-frequency transmission channel of 0 to 10 kHz. The electromagnetic wave coupler is used to encode data into electromagnetic waves and form an electromagnetic wave mode with a medium-frequency transmission channel of 10 kHz to 1 MHz. The optical fiber slip ring is used to encode data into optical signals and form an optical fiber mode with a high-frequency optical fiber channel.

6. The transmission method of the multi-modal combined logging while drilling high-speed transmission system according to any one of claims 1 to 5, characterized in that: The steps include: Step S1: continuously collecting multiple types of formation parameters during the drilling process of the drilling tool; Step S2: Select different transmission modes according to formation parameter type, well depth and mud viscosity to transmit formation parameters to the ground in real time; Step S3: The ground first performs a wavelet transform on the received formation parameters to separate the modal signals corresponding to each transmission mode. Then, by calculating the correlation between different modal signals, conflicting data with low consistency is marked and the conflicting data is repaired. Finally, a probabilistic fusion model is constructed based on the Bayesian network. The repaired data and unmarked data are input into the probabilistic fusion model to obtain the decoded formation parameters, where the transmission modes include mud pulse mode, electromagnetic wave mode and optical fiber mode.

Images