Ultrasonic echo denoising method based on drilling fluid performance parameters and complex solid phase influence

Abstract

The invention discloses an ultrasonic echo denoising method based on drilling fluid performance parameters and complex solid phase influence, and belongs to the field of ultrasonic measurement while drilling, and the method comprises the following steps: collecting underground ultrasonic echo signals and drilling fluid performance parameters in real time, and constructing a drilling fluid interference feature vector; predicting an ultrasonic echo signal-to-noise ratio through a sound wave attenuation model based on the drilling fluid interference feature vector; taking the predicted signal-to-noise ratio as a constraint, and performing frequency band division on the ultrasonic echo signal by adopting improved self-adaptive variational mode decomposition to obtain a plurality of intrinsic mode functions; calculating a solid-phase particle size weighted correlation coefficient for each intrinsic mode function, and dividing the intrinsic mode function into a signal dominant component and a noise dominant component based on the solid-phase particle size weighted correlation coefficient; denoising the noise dominant component, and superposing the noise dominant component with the signal dominant component to obtain a denoised ultrasonic echo signal; and outputting the de-noised ultrasonic echo signal to the measurement-while-drilling system. According to the invention, both noise suppression and signal fidelity can be considered.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of while-drilling ultrasonic measurement, and particularly relates to an ultrasonic echo denoising method based on drilling fluid performance parameters and complex solid phase influence. BACKGROUND

[0002] In the oil and gas drilling engineering, the measurement while drilling (MWD) technology is a core means for realizing accurate control of a wellbore trajectory and real-time monitoring of formation parameters. The ultrasonic echo signal detection is widely used in key scenes such as hole diameter measurement, wellbore imaging and drill bit wear monitoring due to the advantages of non-contact measurement, high spatial resolution and fast response speed. In the prior art, the processing of the ultrasonic echo signal mainly depends on traditional signal denoising methods, which mainly include three types of technical paths: one is wavelet threshold denoising based on time-frequency analysis, which decomposes the signal into different frequency bands by using a preset basis function, and applies a threshold to suppress the noise dominant frequency band; two is empirical mode decomposition (EMD) and improved algorithm (such as EEMD), which adaptively decomposes the signal into intrinsic mode functions (IMF) by using the time scale characteristics of the signal itself, and realizes denoising by screening effective components; three is variational mode decomposition (VMD), which decomposes the signal into a plurality of modal components with sparse characteristics by constructing a constrained variational model, and separates the signal and noise based on energy or correlation criteria. These methods have achieved certain application effect under the conventional drilling fluid environment (low density, low solid phase content), and provide a basic technical support for while-drilling ultrasonic measurement.

[0003] However, the existing technology has significant limitations in complex drilling fluid environments, making it difficult to meet actual engineering needs. On the one hand, the traditional method does not consider the dynamic influence of drilling fluid performance parameters (such as density, viscosity, solid phase volume fraction, and particle size distribution) on ultrasonic signals. When the drilling fluid is in a high-density (>1.8 g / cm³) and high-solid phase (>30%) state, the ultrasonic signal will be severely attenuated and distorted due to strong viscous absorption and particle scattering, resulting in loss of signal characteristics or noise residue during the denoising process. On the other hand, the key parameters of existing decomposition algorithms (such as wavelet basis function, EMD screening times, VMD modal number and penalty factor) are mostly dependent on artificial experience setting and cannot adapt to real-time changes in drilling fluid performance. In complex interference scenarios, modal aliasing or over-decomposition may occur, resulting in limited signal-to-noise ratio improvement (usually <10 dB) after denoising, making it difficult to support high-precision measurement-while-drilling needs. In addition, the existing technology lacks modeling of ultrasonic propagation physical characteristics, and relying solely on data-driven denoising strategies can easily lead to signal distortion, affecting the accuracy of subsequent well trajectory calculation and formation evaluation. SUMMARY

[0004] To solve the above technical problems, the present application provides an ultrasonic echo denoising method based on drilling fluid performance parameters and complex solid phase influence to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above-mentioned purpose, in a first aspect, the present application provides an ultrasonic echo denoising method based on drilling fluid performance parameters and complex solid phase influence, comprising: S1. Real-time acquisition of downhole ultrasonic echo signals and drilling fluid performance parameters, and construction of drilling fluid interference feature vectors; S2. Based on the drilling fluid interference feature vectors, predict the signal-to-noise ratio of the ultrasonic echo through the sound wave attenuation model; S3. With the predicted signal-to-noise ratio as a constraint, use an improved adaptive variational modal decomposition to divide the frequency band of the ultrasonic echo signal and obtain multiple intrinsic modal functions; S4. Calculate the solid phase particle size weighted correlation coefficient for each intrinsic modal function, and divide the intrinsic modal function into signal dominant components and noise dominant components based on the solid phase particle size weighted correlation coefficient; S5. Denoising processing of the noise dominant component and superposition with the signal dominant component to obtain the denoised ultrasonic echo signal; S6. Output the denoised ultrasonic echo signal to the measurement-while-drilling system.

[0006] Preferably, in step S1, the drilling fluid performance parameters include density, funnel viscosity, solid phase volume fraction, median particle size, and 90% particle size; and the drilling fluid interference feature vectors are constructed according to the drilling fluid performance parameters.

[0007] Preferably, in step S2, the acoustic attenuation model is an acoustic scattering-viscous coupling attenuation model; based on the drilling fluid interference feature vector, the ultrasonic echo signal-to-noise ratio is predicted using the acoustic scattering-viscous coupling attenuation model.

[0008] Preferably, the formula for the sound attenuation model is: ; in, This is the viscous absorption term caused by the drilling fluid viscosity. This is the scattering attenuation term caused by solid particles. The frequency is the ultrasonic frequency.

[0009] Preferably, in step S3, the improved adaptive variational mode decomposition employs a multi-objective optimization function, using the predicted signal-to-noise ratio as a constraint, to jointly optimize the number of modes and the penalty factor; wherein, the multi-objective optimization function includes reconstruction error, spectral separation degree, and signal-to-noise ratio penalty term.

[0010] Preferably, the multi-objective optimization function is: ; in, , , These are the weighting coefficients. For based on The penalty term is used to guide the parameters to converge toward a higher signal-to-noise ratio.

[0011] Preferably, in step S4, the solid-phase particle size weighted correlation coefficient is calculated by multiplying the solid-phase interference weight function by the correlation coefficient; the signal dominant component and the noise dominant component are divided according to the solid-phase particle size weighted correlation coefficient; wherein, the solid-phase interference weight function is calculated based on the solid-phase volume fraction and the median particle size.

[0012] Preferably, the solid particle size weighted correlation coefficient is: ; in, For solid-phase interference weighting function, For intrinsic mode functions, The original signal, is the absolute value of the correlation coefficient between the i-th intrinsic mode function and the original signal.

[0013] Preferably, in step S5, the denoising process employs a dual Gaussian impulse model combined with a wavelet packet thresholding strategy. The dominant noise component is fitted using a double Gaussian pulse model; the fitted signal is obtained based on the double Gaussian pulse model. According to the residual of the double Gaussian pulse model, wavelet packet threshold denoising is performed, and the purified component is obtained by superimposing the fitted signal.

[0014] Preferably, the formula of the double Gaussian pulse model is: ; Wherein, 、 is an amplitude coefficient, 、 is a time center, 、 is an attenuation width.

[0015] Compared with the prior art, the present application has the following advantages and technical effects: The present application introduces the physical characteristics of drilling fluid into the denoising process by synchronously collecting drilling fluid performance parameters and constructing an interference feature vector, avoids denoising distortion caused by "pure data driving", and realizes accurate signal processing under physical constraints; and solves the problem that the traditional denoising method is not associated with drilling fluid performance parameters.

[0016] The present application uses the constraint conditions generated by the physical model to guide the adaptive optimization of the decomposition algorithm parameters, without manual intervention, and can dynamically adapt to the complex scene of drilling fluid performance changing with well depth and formation, avoiding modal aliasing or over-decomposition; and breaks through the limitation that the parameters of the traditional decomposition algorithm depend on manual setting.

[0017] The present application performs weighted correlation analysis on the real-time collected downhole parameters and the solid phase parameters in the interference feature vector, rather than relying only on the characteristics of the signal itself, can more accurately identify the effective signal component in the environment of high solid phase and high density drilling fluid, and reduce the loss of effective signal; and improves the accuracy of distinguishing signal and noise components.

[0018] The present application uses the pulse attenuation model fitting combined with wavelet packet threshold processing for the noise dominant component, first fits the signal characteristics remaining in the noise, and then specifically suppresses the residual noise, avoiding the problems of "noise not completely removed" or "signal characteristics lost" caused by a single denoising method, and ensuring the integrity and clarity of the denoised signal. BRIEF DESCRIPTION OF DRAWINGS

[0019] The accompanying drawings, which form a part of this application, are included to provide a further understanding of the application and are incorporated in and constitute a part of this application. The embodiments of this application and their description are used to explain the application and are not intended to limit the application. In the drawings: Figure 1 The ultrasonic echo denoising method flowchart of the embodiment of the present application; Figure 2 The parameter optimization flowchart of the embodiment of the present application; Figure 3 ​A signal contrast chart before and after denoising of an embodiment of the present application; Figure 4 A system overall architecture chart of an embodiment of the present application. DETAILED DESCRIPTION

[0020] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0021] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a group of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in a different order.

[0022] Embodiment one As shown in the drawings, Figure 1 In the present embodiment, a method for removing ultrasonic echo based on drilling fluid performance parameters and complex solid phase under the influence of ultrasonic echo is provided, comprising: S1. Real-time acquisition of downhole ultrasonic echo signal and drilling fluid performance parameters, and construction of drilling fluid interference characteristic vector; Further, in step S1, the drilling fluid performance parameters include density, funnel viscosity, solid phase volume fraction, median particle size and 90% particle size; and the drilling fluid interference characteristic vector is constructed according to the drilling fluid performance parameters.

[0023] Specifically, the drilling fluid performance parameters include density , funnel viscosity , solid phase volume fraction , particle size distribution and , and the drilling fluid interference characteristic vector is .

[0024] S2. Based on the drilling fluid interference characteristic vector, the ultrasonic echo signal-to-noise ratio is predicted by the sound wave attenuation model; Further, in step S2, the sound wave attenuation model is a sound wave scattering-viscosity coupling attenuation model; and the ultrasonic echo signal-to-noise ratio is predicted by the sound wave scattering-viscosity coupling attenuation model according to the drilling fluid interference characteristic vector.

[0025] Further, the formula of the sound attenuation model is: ; Wherein, is the viscous absorption term caused by the viscosity of the drilling fluid, is the scattering attenuation term caused by the solid phase particles, is the ultrasonic frequency.

[0026] Specifically, the embodiment establishes a mapping relationship between an ultrasonic echo signal-to-noise ratio (SNR) and a feature vector , and obtains a SNR prediction function . .

[0027] S3. With the predicted SNR as a constraint, an improved adaptive variational mode decomposition is used to perform frequency band division on the ultrasonic echo signal, and a plurality of intrinsic mode functions are obtained. Further, in step S3, the improved adaptive variational mode decomposition uses a multi-objective optimization function to jointly optimize the mode number and the penalty factor with the predicted SNR as a constraint. The multi-objective optimization function includes a reconstruction error, a spectral separation degree, and a SNR penalty term.

[0028] Further, the multi-objective optimization function is: ; wherein, , , is a weight coefficient, is a penalty term based on , which is used to guide the parameters to converge to a high SNR direction.

[0029] Specifically, as shown in Figure 2 , the embodiment uses an improved adaptive variational mode decomposition method , constructs a multi-objective optimization function with as a constraint condition, and jointly optimizes the mode number and the penalty factor of the intrinsic mode function to realize adaptive frequency band division of the signal.

[0030] S4. A solid phase particle size weighted correlation coefficient is calculated for each intrinsic mode function, and the intrinsic mode function is divided into a signal dominant component and a noise dominant component based on the solid phase particle size weighted correlation coefficient. Further, in step S4, the solid phase particle size weighted correlation coefficient is calculated by the product of a solid phase interference weight function and a correlation coefficient. The signal dominant component and the noise dominant component are divided according to the solid phase particle size weighted correlation coefficient. The solid phase interference weight function is calculated based on a solid phase volume fraction and a median particle size.

[0031] Further, the solid phase particle size weighted correlation coefficient is: ; wherein, is a solid phase interference weight function, is an intrinsic mode function, is an original signal, is the absolute value of the correlation coefficient between the i-th intrinsic mode function and the original signal.

[0032] S5. The noise-dominant component is denoised and superimposed on the signal-dominant component to obtain the denoised ultrasonic echo signal. Furthermore, in step S5, the denoising process employs a dual Gaussian impulse model combined with a wavelet packet thresholding strategy. The dominant noise component is fitted using a double Gaussian pulse model; the fitted signal is obtained based on the double Gaussian pulse model. Based on the residual of the double Gaussian pulse model, wavelet packet thresholding is performed for denoising, and then superimposed with the fitted signal to obtain the purified component.

[0033] Furthermore, the formula for the double Gaussian pulse model is: ; in, , For amplitude coefficient, , Centered on time , The decay width is determined by the particle swarm optimization algorithm (PSO). Estimate from noisy signals.

[0034] In this embodiment, the double Gaussian attenuation model is used to model the energy attenuation characteristics of ultrasonic pulses under solid-state interference.

[0035] S6. Output the denoised ultrasonic echo signal to the drilling measurement system.

[0036] Specifically, in this embodiment, the denoised signal is input into the subsequent ultrasonic ranging or imaging module to achieve high-precision downhole measurement.

[0037] In addition, a detailed implementation scheme is provided for the above embodiments, which specifically includes the following steps: Step S1: Data acquisition and feature vector construction; 1. Ultrasonic echo signal With 5 Sampling rate acquisition: 1024 points are collected per pulse cycle (approximately 0.2). ); 2. Synchronously read drilling fluid parameters: density ( ), Funnel viscosity ( Solid volume fraction (%), median particle size ( ), 90% particle size ( 3. Construct the drilling fluid interference feature vector , data type is 32-bit floating-point array.

[0038] Step S2: Signal-to-noise ratio prediction model calculation 1. Sound attenuation model calculation: ; Where, =500 (Ultrasonic center frequency); =( -25)×0.1 (estimated dynamic viscosity, unit · ); =1450+3.5×( -20) (sound speed empirical formula, unit ); Obtained by scattering theory pre-computation lookup table.

[0039] 2. Signal-to-noise ratio prediction: Using ridge regression model ; coefficient ~ Fitted by historical well data (R2≥0.92), stored in ; Real-time calculation , output is a scalar value (unit ).

[0040] Step S3: Adaptive frequency band decomposition; 1. Multi-objective optimization function: Where, is the reconstruction error (mean square error of the original signal and and); is the spectral overlap degree (sum of the reciprocal of the center frequency interval of each ).

[0041] 2. Particle swarm optimization (PSO) parameters: Particle number 30, iteration 50 times; search range: ∈[4,12] (integer), ∈[1000,8000]; optimal solution Used for VMD decomposition, obtaining component.

[0042] ​​​​Step S4: Signal and noise component discrimination; 1. Calculate solid phase interference weight: 2. Calculate weighted correlation coefficient: , Correlation coefficient calculation window length takes the main lobe width of the pulse (about 40 sampling points).

[0043] 3. Discrimination rule: > 0.6: Signal dominant component (retained); ≤ 0.6: Noise dominant component (enter step S5 processing).

[0044] Step S5: Noise component denoising and reconstruction; 1. Double Gaussian pulse fitting: Model: ; Fitting target: Minimize fitting error (<5%), parameter range: , ∈[0,0.5 max(IMF_noise)]; , ∈[0,100] (sampling point index); , ∈[1,20].

[0045] 2. Wavelet packet threshold denoising: The fitting residual is 6 wavelet packet 4-layer decomposition; Adaptive threshold calculation, soft threshold function processing; Reconstruct the denoised residual signal.

[0046] 3. Superimpose double Gaussian model and denoised residual to get purified component.

[0047] 4. Superimpose all signal dominant components and purified components to generate .

[0048] Step S6: Output and integration.

[0049] 1. Output through -485 / Output, data format is 12-bit fixed point number (0-3.3 corresponding to 0-4095); 2. Transmission protocol: Address 0x01, function code 0x10 (write multiple registers), data length 512 words; 3. After receiving the data, the system uses it for wellbore trajectory calculation (distance measurement error <1cm) or wellbore imaging (resolution <5mm).

[0050] Laboratory simulation: Configuration density 1.2-2.2 Drilling fluid with a solid content of 10-40% is injected with standard particles ( =50-100 ), compare the ranging error before and after denoising; signal comparison graphs before and after denoising, such as Figure 3 As shown.

[0051] This implementation scheme ensures the reliability of the method in high-temperature, high-pressure, and high-vibration downhole environments, and can be directly embedded into existing systems. The system meets the real-time requirements of measurement while drilling.

[0052] Example 2 Based on the same inventive concept, this embodiment also provides a method based on downhole measurement while drilling (MDR) The system hardware platform mainly includes an ultrasonic signal acquisition module, a drilling fluid performance parameter monitoring module, a data processing unit, and output interfaces. The overall system architecture is as follows: Figure 4 As shown, the modules are as follows: Ultrasonic signal acquisition module: using a center frequency of 500 piezoelectric transducers with an outer diameter ≤25 Withstand pressure ≥140 Bandwidth ≥ 40%, sampling rate ≥ 5 Used to transmit ultrasonic pulses and receive echo signals .

[0053] Drilling fluid performance parameter monitoring module: integrated Coriolis density meter (measurement) ), rotational viscometer (for measuring) Laser diffraction particle size analyzer (for measuring particle size distribution) Data update cycle ≤ 10 The accuracy meets the requirements. ±0.005 , ±2 .

[0054] Data processing unit: based on 32-bit floating point 200MHz , ≥512 , ≥1 , built-in temperature compensation module (correcting 0-150℃ sound speed drift, precision ±0.5%), integrated Algorithm, double Gaussian fitter, wavelet packet denoiser.

[0055] Output interface: support bus or -485 physical layer, protocol is , baud rate 115200, data bits 8, stop bit 1, no check bit, isolation voltage 2500 , transmission delay <50 .

[0056] The above, only for the preferred specific embodiments of the present application, but the scope of the present application is not limited to this, any skilled in the art of the technical personnel in the technical range disclosed by the present application, can easily think of changes or replacement, should be covered in the scope of the present application. Therefore, the scope of protection of the present application should be subject to the scope of protection of claims.

Claims

1. An ultrasonic echo noise reduction method based on drilling fluid performance parameters and the influence of complex solid phases, characterized in that, Includes the following steps: S1. Real-time acquisition of downhole ultrasonic echo signals and drilling fluid performance parameters to construct drilling fluid interference feature vectors; S2. Based on the drilling fluid interference feature vector, predict the ultrasonic echo signal-to-noise ratio using the acoustic attenuation model; S3. Using the predicted signal-to-noise ratio as a constraint, the ultrasonic echo signal is divided into frequency bands by an improved adaptive variational mode decomposition to obtain multiple intrinsic mode functions; S4. Calculate the solid particle size weighted correlation coefficient for each intrinsic mode function, and divide the intrinsic mode function into signal-dominant components and noise-dominant components based on the solid particle size weighted correlation coefficient; S5. The noise-dominant component is denoised and superimposed on the signal-dominant component to obtain the denoised ultrasonic echo signal. S6. Output the denoised ultrasonic echo signal to the drilling measurement system.

2. The method according to claim 1, characterized in that, In step S1, the drilling fluid performance parameters include density, funnel viscosity, solid volume fraction, median particle size, and 90% particle size; based on the drilling fluid performance parameters, a drilling fluid interference feature vector is constructed.

3. The method according to claim 1, characterized in that, In step S2, the acoustic attenuation model is an acoustic scattering-viscous coupling attenuation model; based on the drilling fluid interference feature vector, the ultrasonic echo signal-to-noise ratio is predicted using the acoustic scattering-viscous coupling attenuation model.

4. The method according to claim 3, characterized in that, The formula for the sound attenuation model is: ； in, This is the viscous absorption term caused by the drilling fluid viscosity. This is the scattering attenuation term caused by solid particles. The frequency is the ultrasonic frequency.

5. The method according to claim 1, characterized in that, In step S3, the improved adaptive variational mode decomposition employs a multi-objective optimization function, using the predicted signal-to-noise ratio as a constraint, to jointly optimize the number of modes and the penalty factor; wherein, the multi-objective optimization function includes reconstruction error, spectral separation degree, and signal-to-noise ratio penalty term.

6. The method according to claim 5, characterized in that, The multi-objective optimization function is: ； in, , , These are the weighting coefficients. Based on The penalty term is used to guide the parameters to converge toward a higher signal-to-noise ratio.

7. The method according to claim 1, characterized in that, In step S4, the solid-phase particle size weighted correlation coefficient is calculated by multiplying the solid-phase interference weight function by the correlation coefficient; based on the solid-phase particle size weighted correlation coefficient, the signal dominant component and the noise dominant component are divided; wherein, the solid-phase interference weight function is calculated based on the solid-phase volume fraction and the median particle size.

8. The method according to claim 7, characterized in that, The solid-phase particle size weighted correlation coefficient is: ； in, For solid-phase interference weighting function, For intrinsic mode functions, The original signal, is the absolute value of the correlation coefficient between the i-th intrinsic mode function and the original signal.

9. The method according to claim 1, characterized in that, In step S5, the denoising process employs a dual Gaussian impulse model combined with a wavelet packet thresholding strategy. The dominant noise component is fitted using a double Gaussian pulse model; the fitted signal is obtained based on the double Gaussian pulse model. Based on the residual of the double Gaussian pulse model, wavelet packet thresholding is performed for denoising, and then superimposed with the fitted signal to obtain the purified component.

10. The method according to claim 9, characterized in that, The formula for the double Gaussian pulse model is: ； in, , For amplitude coefficient, , Centered on time , This represents the attenuation width.

Images