Drilling fluid outlet flow anti-disturbance monitoring method and device, medium and program product

Abstract

The invention provides a drilling fluid outlet flow anti-disturbance monitoring method and device, a medium and a program product, and the method comprises the steps: obtaining a liquid level echo signal through a millimeter-wave radar liquid level meter installed above a drilling fluid outlet buffer tank, and inhibiting an environment interference factor through time domain and frequency domain fusion adaptive filtering of dynamically adjusting a weight according to a signal-to-noise ratio; inputting the filtering signal into a BP neural network to obtain phase information influenced by the liquid level disturbance interference factor, and combining a physical information neural network to obtain a radar wave attenuation coefficient corresponding to the phase information; on the basis of a liquid level fluctuation and radar echo attenuation quantitative model, liquid level height variation is obtained by integrating phase information and radar wave attenuation coefficient inversion calculation, the current liquid level height is obtained by combining the liquid level height with radar ranging liquid level height, and the relative flow change of the drilling fluid in the buffer tank is judged. High-precision flow monitoring is achieved under the conditions of liquid level disturbance and environment disturbance, and well control safety and the intelligent level of drilling operation are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of oil well exploitation, and particularly relates to a drilling fluid outlet flow anti-disturbance monitoring method, device, medium and program product. BACKGROUND

[0002] During drilling operation, the change of the drilling fluid liquid level in the outlet buffer tank can directly reflect the change of the wellbore flowback flow, and is an important basis for well control monitoring and preventing safety accidents such as well kick and well leakage. In the prior art, the liquid level monitoring methods include manual observation, target type flow sensor measurement and ultrasonic liquid level sensor measurement.

[0003] However, the manual observation method is susceptible to factors such as negligence and fatigue of the operator, and has problems of untimely response and easy missed detection; the target type flow sensor is a contact type measurement device, and the target body is easy to be attached with mud to increase the mass and cause measurement error, and the installation position is usually welded on the elevated tank pipe, which is inconvenient to maintain; the ultrasonic liquid level sensor has the advantage of non-contact measurement, but is susceptible to signal attenuation in high-temperature steam, electromagnetic interference and metal reflection environments, and the measurement stability significantly decreases when the liquid surface fluctuates violently, so it usually needs to be installed at a position with relatively stable liquid surface, which limits its application range.

[0004] Therefore, the existing monitoring methods generally lack comprehensive processing and intelligent discrimination mechanisms for environmental interference and liquid surface disturbance, resulting in insufficient accuracy and stability of the liquid level measurement results, and it is difficult to meet the demand of intelligent oilfield for high-precision and intelligent liquid level monitoring. SUMMARY

[0005] In view of the deficiencies in the prior art, the present application provides a drilling fluid outlet flow anti-disturbance monitoring method, device, medium and program product, at least to solve the problem that the measurement precision is reduced due to the influence of liquid surface fluctuation and environmental factors in the drilling fluid outlet flow monitoring process in the prior art.

[0006] To achieve the above-mentioned purpose and other advantages, some embodiments of the present application provide the following aspects:

[0007] In a first aspect, some embodiments of the present application provide a drilling fluid outlet flow anti-disturbance monitoring method, comprising:

[0008] using a millimeter wave radar liquid level meter installed above the drilling fluid outlet buffer tank, emitting millimeter wave radar waves to the liquid surface in the buffer tank and receiving reflected echo signals;

[0009] performing time domain and frequency domain filtering processing on the echo signals, dynamically adjusting the fusion weight of the time domain filtering output and the frequency domain filtering output according to the signal-to-noise ratio, to obtain a filtered signal filtering out environmental interference echoes;

[0010] inputting the filtered signal into a BP neural network model pre-trained to output phase information affected by a liquid surface disturbance interference factor;

[0011] inputting the phase information and feature data representing environmental parameters, radar parameters and medium parameters into a physical information neural network model pre-trained to output a radar wave attenuation coefficient corresponding to the phase information;

[0012] inputting the radar wave attenuation coefficient and the phase information into a constructed liquid surface fluctuation and radar echo attenuation quantification model to obtain a liquid surface height variation;

[0013] obtaining a current liquid surface height by combining the liquid surface height variation with a radar ranging formula, and determining a relative flow rate change of the drilling fluid in the outlet buffer tank according to the current liquid surface height.

[0014] In a second aspect, some embodiments of the present application further provide an electronic device, which comprises:

[0015] one or more processors; and a memory storing computer program instructions which, when executed, cause the processor to perform the drilling fluid outlet flow rate anti-disturbance monitoring method according to any one of the above.

[0016] In a third aspect, some embodiments of the present application further provide a computer-readable storage medium having stored thereon computer programs and / or instructions which, when executed by a processor, implement the drilling fluid outlet flow rate anti-disturbance monitoring method according to any one of the above.

[0017] In a fourth aspect, some embodiments of the present application further provide a computer program product comprising computer programs and / or instructions which, when executed by a processor, implement the drilling fluid outlet flow rate anti-disturbance monitoring method according to any one of the above.

[0018] Compared with the related art, in the scheme provided by the embodiments of the present application, by introducing an adaptive fusion filtering mechanism in the time domain and the frequency domain in the signal processing process of the millimeter wave radar liquid level meter, the environmental electromagnetic interference and the multipath reflection noise can be effectively suppressed, and the signal-to-noise ratio of the liquid surface echo signal is improved; the phase information of the liquid surface reflection echo is accurately extracted from the filtered echo signal by using a pre-trained BP neural network model, and the physical constraints of the liquid surface fluctuation and the medium propagation loss are introduced by combining a physical information neural network model, so that the radar wave attenuation coefficient is accurately inverted; further, by using a liquid surface fluctuation and radar echo attenuation quantitative model, the phase information and the radar wave attenuation coefficient are comprehensively solved to obtain the liquid surface height change, and the current liquid surface height is obtained by combining the radar ranging formula, so that the relative flow change of the drilling fluid in the buffer tank can still be stably and accurately determined in the case of liquid surface disturbance and environmental interference, and the anti-disturbance capability and the measurement precision of the drilling fluid outlet flow monitoring are significantly improved. Through the above scheme, the outlet drilling fluid return change can be mastered in real time during drilling, potential accidents such as well kick or lost circulation can be predicted in advance, the demand for high-end intelligent equipment in the construction of smart oilfield is met, and the occurrence of well control accidents is prevented and the drilling safety is improved. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other embodiments can be obtained by those skilled in the art without creative labor.

[0020] Figure 1 is a flowchart of a drilling fluid outlet flow anti-disturbance monitoring method provided by the embodiments of the present application;

[0021] Figure 2 is a structural schematic diagram of a drilling fluid outlet buffer tank provided by the embodiments of the present application, which is installed with a millimeter wave radar liquid level meter;

[0022] Figure 3 is a comparison diagram of outlet flow change characteristics of a traditional ultrasonic measurement method and a millimeter wave radar measurement method provided by the embodiments of the present application;

[0023] Figure 4 is a well outlet flow monitoring comparison diagram of a traditional ultrasonic measurement method and a millimeter wave radar measurement method provided by the embodiments of the present application;

[0024] Figure 5 is a structural schematic diagram of an electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0025] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0026] The following terms are used herein.

[0027] Millimeter wave radar liquid level meter: refers to a device for non-contact liquid level measurement using electromagnetic waves in the millimeter wave band (generally 30-300 GHz). It transmits millimeter wave electromagnetic signals to the liquid surface through an antenna and receives the echo signals reflected back by the liquid surface, and calculates the liquid level height according to the time difference or frequency difference between the transmitted and received signals. Such sensors have the characteristics of small beam angle, strong resistance to steam, dust and temperature changes, and are suitable for precise liquid level measurement under complex working conditions.

[0028] Ultrasonic liquid level meter: refers to a device for non-contact liquid level measurement using the propagation characteristics of ultrasonic waves in air or liquid medium. It transmits ultrasonic wave pulses of a certain frequency to the liquid surface through a transducer, and the ultrasonic wave encounters the liquid surface and forms an echo signal. The device calculates the liquid level height according to the time difference between the transmission and echo reception combined with the sound velocity. Ultrasonic liquid level meters are easy to install and do not pollute the measured medium, and are suitable for liquid level monitoring under normal temperature and pressure conditions.

[0029] BP neural network model (Back Propagation Neural Network): refers to a multi-layer feedforward artificial neural network model based on the error backpropagation algorithm. The model is composed of an input layer, one or more hidden layers and an output layer. It calculates the output result through forward propagation and adjusts the output error to each connection weight and threshold value through backward propagation according to the gradient descent method, to realize modeling and prediction of the nonlinear mapping relationship between input and output. The BP neural network model has the characteristics of self-learning, self-adaptation and strong fault tolerance, and is widely used in pattern recognition, prediction modeling and nonlinear regression analysis fields.

[0030] Physics-Informed Neural Network (PINN) is a deep learning model that directly embeds physical laws or governing equations into the neural network training process. This model introduces physical constraints into the loss function, ensuring that the network, while learning the mapping between input and output data, also satisfies the conservation laws, boundary conditions, or differential equation constraints of the established physical model. This improves the model's generalization ability in scenarios with insufficient samples or extrapolation. PINN models are suitable for prediction and inversion problems that require the integration of domain-specific prior knowledge and data-driven methods.

[0031] First Embodiment

[0032] The first embodiment of this application relates to a method for monitoring drilling fluid outlet flow rate without disturbance, referring to... Figure 1 As shown, the method may include the following steps:

[0033] Step S1: Using a millimeter-wave radar level gauge installed above the drilling fluid outlet buffer tank, a millimeter-wave radar wave is emitted towards the liquid surface in the buffer tank and the reflected echo signal is received.

[0034] Specifically, regarding step S1, refer to Figure 2 As shown, a millimeter-wave radar level gauge 100, installed above the drilling fluid outlet buffer tank 112, is fixed to the center of the top of the buffer tank 112 by a support rod 120, with the antenna beam perpendicularly pointing towards the liquid surface 113 inside the buffer tank. The level gauge typically operates in the 78–80 GHz frequency band, with a beam angle not exceeding 6° and a maximum power consumption of 0.54 W. The millimeter-wave electromagnetic signal emitted by the millimeter-wave radar level gauge is reflected at the liquid surface to form an echo signal. The time difference between transmission and reception can be preliminarily calculated by the system to obtain the raw measurement data of the liquid level height.

[0035] In practical applications, although millimeter-wave radar level gauges have a certain resistance to steam and dust interference, they are still affected by various factors in the special scenario of drilling fluid outlet buffer tanks. Among these factors, the undulation of the liquid surface 113 changes the radar wave incident angle and causes changes in the distribution of scattered energy. When drilling fluid in the outlet overhead pipe 142 impacts the buffer tank 112 at a high flow rate, it creates violent fluctuations, turbulence, and splashing in the drilling fluid 125, accompanied by turbulence and bubbles. These factors exacerbate the phase noise and amplitude fluctuations of the echo signal, constituting liquid surface interference. Furthermore, multipath reflections from the inner wall of the buffer tank, reflections blocked by metal components, electromagnetic noise introduced by strong electromagnetic radiation sources on site (such as motors, frequency converters, and radio transmitting equipment), and the superposition effect of direct and reflected waves when the liquid surface is close to the antenna all constitute environmental interference. These interferences will be superimposed on the effective liquid surface echo to varying degrees, causing distortion of the original measurement data. Therefore, it is necessary to suppress these interference factors in subsequent steps.

[0036] Step S2: Time domain and frequency domain filtering processing is performed on the echo signal, and a fusion weight of time domain filtering output and frequency domain filtering output is dynamically adjusted according to a signal-to-noise ratio, so as to obtain a filtered signal filtering out environmental interference echoes.

[0037] For step S2, specifically, time domain and frequency domain fusion adaptive filtering processing is performed on the echo signal output by the millimeter wave radar liquid level meter. The time domain filtering is a smoothing processing of the echo signal in the time sequence dimension to suppress the burst random noise; the frequency domain filtering is to identify and attenuate the interference components in a specific frequency range through spectrum analysis. The fusion weight of the two types of filtering outputs is dynamically adjusted according to the signal-to-noise ratio (SNR, Signal-to-Noise Ratio), which means that according to the ratio of the effective signal power and the noise power of the current echo signal, the relative weight of the time domain filtering and the frequency domain filtering is changed in real time to adapt to the change of the intensity of different environmental interference. The environmental interference includes but is not limited to tank wall multi-path reflection echo, shielding reflection of metal objects to radar waves, interference echo generated by strong electromagnetic field in the field, and multi-wave superposition effect under short distance measurement conditions. After the above fusion filtering processing, the signal-to-noise ratio of the echo signal originally affected by the multi-path reflection and electromagnetic interference is significantly improved.

[0038] Step S3: The filtered signal is input into a pre-trained BP neural network model to output phase information affected by the liquid surface disturbance interference factor.

[0039] For step S3, specifically, the filtered signal is the signal after the time domain and frequency domain fusion filtering processing of the original millimeter wave radar echo, but still retains the interference components introduced by the liquid surface disturbance interference factor (such as bubbles, turbulence, and spray) in the echo formation process. Different liquid surface disturbance interference factors will cause different modes of interference signals in the radar wave propagation and reflection path, and these interference signals exist in the original echo at the same time as the phase information and are jointly manifested as phase shift characteristics. Therefore, when the BP neural network extracts the phase information affected by the liquid surface disturbance interference factor based on the filtered signal, the main components in the propagation phase difference and the complex signal phase are identified as the phase delay components caused by the liquid surface geometric change, and the disturbance components in the interference phase and the intermediate frequency signal phase are identified as the phase shift specific to the liquid surface disturbance interference factor. The phase shift is used to represent the non-geometric path phase change introduced by the liquid surface disturbance interference factor, and is deducted as a disturbance compensation in the liquid surface height inversion calculation process, so as to eliminate the influence of the liquid surface disturbance on the measurement accuracy.

[0040] Step S4: The phase information and the feature data representing the environmental parameters, radar parameters and medium characteristic parameters are input into a pre-trained physical information neural network model to output the radar wave attenuation coefficient corresponding to the phase information.

[0041] For step S4, specifically, the environmental parameters are external physical conditions that affect the propagation of millimeter wave signals and the reflection of liquid surfaces, including but not limited to temperature, humidity, air pressure, air flow disturbance, suspended particle concentration, and electromagnetic interference strength. The radar parameters are the working characteristics of the millimeter wave radar level meter, including but not limited to working frequency, transmission power, antenna gain, beam width, and modulation mode. The medium characteristic parameters are the electromagnetic characteristics of the liquid surface and the liquid medium, including but not limited to dielectric constant, conductivity, density, medium surface roughness, and wave characteristics.

[0042] Due to the disturbance of the liquid surface (such as ripples, sloshing, and fluctuation), the millimeter wave echo signal will be shifted in phase, and amplitude attenuation will also be introduced. If only the original phase information is used to do height inversion without considering this attenuation effect, the echo power term in the radar ranging formula will be distorted, resulting in a deviation in the calculation of the liquid surface height. The phase information itself is only a signal feature, while the attenuation coefficient is a physical quantity directly corresponding to the radar propagation model (directly related to free space propagation loss and medium absorption loss). Therefore, in this step, the phase information and the feature data representing the environmental parameters, radar parameters, and medium characteristic parameters are input into the pre-trained physical information neural network model, and through the model, the complex phase disturbance pattern is mapped to the physical quantity corresponding to the radar propagation model, i.e., the radar wave attenuation coefficient. This attenuation coefficient not only reflects the comprehensive effect of free space propagation loss and medium absorption loss, but also can be dynamically updated according to the real-time measured environmental and medium characteristics, so as to continuously correct the liquid surface height inversion result under different temperature, humidity, dielectric constant, and other working conditions, ensuring the stability and accuracy of the liquid level measurement.

[0043] Step S5: input the radar wave attenuation coefficient and the phase information into the constructed liquid surface fluctuation and radar echo attenuation quantification model to obtain the liquid surface height change.

[0044] For step S5, specifically, the radar wave attenuation coefficient output by step S4 is comprehensively utilized with the phase information obtained in step S3 to quantify the trend of liquid level height change over time. The pre-constructed liquid level fluctuation and radar echo attenuation quantification model is a statistical-physical hybrid model established on the basis of analyzing a large amount of simulation and measured data. The model jointly models the phase change pattern caused by liquid level disturbance and the amplitude attenuation characteristics of the echo signal to form a mapping relationship between phase-attenuation-height change. In actual operation, the phase information can reflect the relative position change caused by the slight fluctuation of the liquid level, and the radar wave attenuation coefficient is used to represent the influence of liquid level disturbance and medium characteristics on the millimeter wave propagation path. By inputting both into the quantification model, false height fluctuations caused by medium absorption, scattering and wave surface distortion can be eliminated, and a more approximate true height change under the static liquid level is obtained. Therefore, the liquid level height change quantity obtained by inversion with the help of the quantification model is more smooth and stable, which can significantly improve the robustness and accuracy of subsequent liquid level-based flow change determination.

[0045] Step S6: Combine the liquid level height change quantity with the radar ranging formula to obtain the current liquid level height, and determine the relative flow change of the drilling fluid in the outlet buffer tank according to the current liquid level height.

[0046] For step S6, specifically, the radar ranging formula is used to calculate the one-way distance from the liquid level to the antenna according to the round-trip propagation time t of the radar wave transmitted by the millimeter wave level gauge and the liquid surface reflected echo, and the formula is:

[0047]

[0048] Wherein, c is the propagation speed of electromagnetic wave in air (approximately take the speed of light 3x10 8 m / s), t is the round-trip propagation time of the echo signal, and 2 is divided to obtain the one-way propagation distance d.

[0049] The ranging time t can be obtained by performing frequency spectrum analysis on the echo signal received by the millimeter wave radar. Specifically, the main peak frequency component corresponding to the target liquid surface echo is extracted by frequency spectrum analysis, and the round-trip propagation time t of the signal is calculated according to the difference between the frequency component and the transmitted signal. Substitute the t value into the radar ranging formula to obtain the liquid level height reference value, and combine the liquid level height change quantity Δh obtained by inversion in step S5 to calculate the current liquid level height h:

[0050]

[0051] The calculation method can compensate the height deviation caused by factors such as liquid surface disturbance and signal attenuation on the basis of the conventional ranging result of the millimeter wave liquid level meter, so as to obtain a height value closer to the real liquid surface. Finally, the system determines the relative flow change of the drilling fluid in the outlet surge tank according to the current liquid surface height and in combination with the geometric size or calibration curve of the surge tank. Since the outlet of the surge tank 150 is connected to the vibrating screen, the outlet speed is basically stable, so the monitoring result can accurately reflect the flow change of the drilling fluid in the outlet surge tank, thereby realizing disturbance-resistant monitoring of the outlet flow. For example, when the drilling fluid flow in the outlet overhead tank pipe 142 increases and the flow rate accelerates, the liquid surface disturbance will be enhanced and a rising trend will occur. The system can accurately monitor the rising of the liquid surface and timely predict the blowout accident that may occur. Conversely, when the monitoring result shows that the liquid surface of the outlet surge tank has a continuous downward trend under the condition that the outlet speed of the outlet remains basically stable, and the downward amplitude exceeds the preset threshold, it can be determined that there is a risk of drilling fluid loss in the well, thereby predicting the occurrence of a lost circulation accident.

[0052] Compared with the related art, in the scheme provided by the embodiments of the present application, by introducing an adaptive fusion filtering mechanism in the time domain and the frequency domain in the signal processing process of the millimeter wave radar liquid level meter, environmental electromagnetic interference and multipath reflection noise can be effectively suppressed, and the signal-to-noise ratio of the liquid surface echo signal can be improved. The phase information of the liquid surface reflection echo is accurately extracted from the filtered echo signal by using a pre-trained BP neural network model, and the physical constraints of liquid surface fluctuation and medium propagation loss are introduced by combining a physical information neural network model, so as to realize accurate inversion of the radar wave attenuation coefficient. Further, by using a liquid surface fluctuation and radar echo attenuation quantization model, the phase information and the radar wave attenuation coefficient are comprehensively solved to obtain the liquid surface height change, and the current liquid surface height is obtained by combining the radar ranging formula, so that the relative flow change of the drilling fluid in the surge tank can still be stably and accurately determined in the presence of liquid surface disturbance and environmental interference, and the disturbance-resistant capability and measurement accuracy of the drilling fluid outlet flow monitoring are significantly improved. Through the above scheme, real-time monitoring of the outlet drilling fluid return change can be realized during drilling, potential accidents such as blowout or lost circulation can be predicted in advance, the demand for high-end intelligent equipment in the construction of smart oilfields is met, and the occurrence of well control accidents can be prevented and the safety of drilling can be improved.

[0053] Second embodiment

[0054] The second embodiment of the present application relates to a drilling fluid outlet flow disturbance-resistant monitoring method, which is an improvement based on the first embodiment. The specific improvement is that in the second embodiment of the present application, a specific implementation of time-frequency domain joint adaptive filtering is provided, that is, step S2 can further include the following steps:

[0055] Step S201: The echo signal is segmented and windowed, and then converted into a frequency domain signal by fast Fourier transform.

[0056] The echo signal x is segmented and windowed by an analysis window with a length of N, and each segment is denoted as x(n) (0≤n≤N-1). Fast Fourier transform (FFT) is performed on each segment to obtain a frequency domain signal X(k):

[0057]

[0058] where N is the window length, and k is the frequency point index. The frequency domain signal X(k) contains superimposed components of the real echo of the liquid surface and environmental interference.

[0059] Step S202: The frequency domain signal is updated in adaptive weight by using a frequency domain least mean square algorithm to obtain a frequency domain filtered signal.

[0060] In this step, the frequency domain least mean square (FLMS) algorithm is used to update the adaptive weight of the frequency domain signal X(k), that is, the frequency domain weight W1(k) is adjusted by an iterative formula, that is, the weight coefficient of the filter in the frequency domain is represented as:

[0061] W1(k+1)=W1(k)+μ·E(k)·X(k)

[0062] where μ is a convergence factor, and its value range is limited to 0<μ<1 / ρ max , ρ max represents the maximum eigenvalue of the input signal correlation matrix; and E(k) is a frequency domain error signal.

[0063] Step S203: The frequency domain error signal is calculated according to the difference between the expected response signal and the frequency domain filtered signal.

[0064] E(k)=D(k)-Y(k)

[0065] where E(k) is the frequency domain error signal; D(k) is the expected response signal; Y(k)=W1(k)·X(k), “·” represents a point-by-point multiplication operation, and Y(k) is the output of the frequency domain filter, that is, the frequency domain filtered signal.

[0066] After obtaining the output signal processed by the frequency domain filtering, it is compared with the preset expected response signal. The expected response signal can be generated from a prior model of the liquid level real change, reference sensor data or historical statistical mean value, and can represent the ideal performance of the liquid level signal in the frequency domain without interference. By comparing the amplitude and phase difference of the filtered signal and the expected response point by point, a residual signal in the frequency domain is obtained.

[0067] The objective of the frequency domain least mean square is to minimize the mean square error between the expected response D(k) and the frequency domain output Y(k) = W1(k) X(k). The weight is adjusted along the negative gradient direction of the error to the weight in each iteration, i.e., W1(k+1) = W1(k) + μE(k) X(k), E(k) = D(k) - Y(k), and μ is subject to 0 < μ < 1 / ρ max The constraint guarantees stable convergence. The real liquid surface echo has the highest correlation with the expected response D(k) in the frequency spectrum,

[0068] Therefore, in the iteration, the E(k) and X(k) terms corresponding to the frequency points are long-term of the same sign, the weight W1(k) is continuously positively reinforced at these frequency points, which is equivalent to increasing the gain of the real liquid surface component. The correlation of multipath reflection, narrowband electromagnetic interference and other frequency points with D(k) is low or even inconsistent in phase, so that the E(k) and X(k) terms at these frequency points are alternately positive and negative or the average is close to zero, and the update terms long-term cancel each other out, and the weight is difficult to grow, which is equivalent to suppressing / not amplifying the interference frequency points. Through this process, the weight coefficient of the frequency domain filter is dynamically corrected, thereby enhancing the real signal component of the liquid surface and suppressing multipath reflection and electromagnetic interference.

[0069] Step S204: converting the frequency domain error signal and fusing it with the output of the time domain adaptive filter to obtain an error-compensated time domain filtered signal.

[0070] In the complex scene of the drilling fluid outlet buffer tank, due to the scattering caused by the liquid surface fluctuation, the superposition of multipath echoes and the electromagnetic interference on site, it is difficult to completely eliminate the noise by relying on a single domain filter. By feeding back the frequency domain error signal to the time domain filter, the residual interference component can be further weakened in the time sequence, thereby improving the fidelity and robustness of the filtered output to the real echo of the liquid surface.

[0071] In this embodiment, step S204 specifically includes:

[0072] Step S2041: converting the frequency domain error signal into a time domain error signal through inverse fast Fourier transform.

[0073] The frequency domain error signal E(k) obtained in step S203 is converted back to the time domain through inverse fast Fourier transform (IFFT) to obtain a time domain error signal e(n), and the calculation formula is as follows:

[0074]

[0075] Wherein, N is the number of frequency points, and Re(·) represents the real part.

[0076] Step S2042: updating the weight parameters of the time domain adaptive filter based on the time domain error signal using the normalized least mean square algorithm.

[0077] Based on the time domain error signal e(n), the weight parameter W2(n) of the time domain adaptive filter at time n is iteratively updated by using a normalized least mean square (NLMS) algorithm, and the update formula is:

[0078]

[0079] Wherein, a is a time domain step factor, ε is a small constant to prevent the denominator from being zero, and φ(n) represents an input signal vector composed of consecutive sampling values of the echo signal x within a preset sampling window. The update process can quickly track time-varying interference and improve filtering accuracy.

[0080] Step S2043: The input signal vector composed of consecutive sampling values of the echo signal within a preset sampling window is filtered by using the time domain adaptive filter with updated weight parameters to obtain a time domain filtered signal.

[0081] At this time, the time domain filtered signal y time (n) can be obtained by the following formula:

[0082]

[0083] Wherein, represents the transpose of the weight parameter at time n, and φ(n) represents the input signal vector. The output y time (n) of the time domain filter depends on the real-time update of the weight parameter.

[0084] Step S205: The fusion weight is adaptively calculated according to the signal-to-noise ratio estimation result, the time domain filtered signal and the frequency domain filtered output are weighted and fused according to the fusion weight, and a filtered signal in which environmental interference echoes are filtered out is obtained.

[0085] Specifically, assuming that the fusion weight is λ, the determination method of the fusion weight λ can be: the time domain features and the frequency domain features of the frequency-modulated continuous millimeter wave radar echo signal are extracted respectively, and the similarity indexes of the two are calculated; when the similarity is high, λ takes an intermediate weight to keep the balance of the time domain and frequency domain filtering results; when the similarity is low, the value of λ is dynamically adjusted in combination with the signal-to-noise ratios of the time domain and frequency domain outputs, so that the side with a higher signal-to-noise ratio has a greater weight in the fusion result, so as to balance between maintaining signal details and suppressing interference. Therefore, λ is dynamically adjusted according to the signal-to-noise ratio of the echo signal to balance between maintaining signal details by time domain filtering and suppressing interference by frequency domain filtering, and finally the fusion output can be represented as:

[0086] y out (n) = λ·y time(n) + (1 - l) - Re(IFFT(Y(k)))

[0087] wherein y out (n) represents the time-frequency fusion filter output; y time (n) is the time domain adaptive filter output; Re(IFFT(Y(k))) is the real part signal after inverse fast Fourier transform of the frequency domain filter output Y(k).

[0088] It can be found that in the scheme provided by the embodiments of the present application, the time domain filter signal and the frequency domain filter output are weighted and synthesized according to the adaptive fusion weight by adopting the time-frequency weighting fusion strategy based on the signal-to-noise ratio. The present application can fully exert the advantages of the time domain filter in suppressing random noise and the advantages of the frequency domain filter in weakening the periodic interference components, so as to effectively reduce the influence of environmental noise, mechanical vibration, electromagnetic interference and other factors on the echo signal while ensuring the integrity of the signal. Thus, the fluctuation of the liquid level real echo signal in amplitude and phase is smoothed and corrected, the stability and reliability of the signal detection are improved, and the accuracy and robustness of the liquid level measurement result are enhanced.

[0089] Third embodiment

[0090] The third embodiment of the present application relates to a drilling fluid outlet flow disturbance resistance monitoring method, which is an improvement based on the first embodiment. The specific improvement is that in the third embodiment of the present application, a specific implementation of training the BP neural network model is provided, which specifically includes the following steps:

[0091] Collect experimental simulation data and radar measured data under different working conditions as a first data training set, and use the phase information corresponding to the calibration data of the laboratory simulated multiphase flow fluctuation as the first label sample corresponding to the first data training set;

[0092] Filter and preprocess the first data training set to extract first multi-dimensional signal feature parameters for characterizing amplitude-frequency characteristics and phase characteristics, and form a first input feature vector;

[0093] Perform forward propagation calculation based on the first input feature vector to obtain a first predicted output;

[0094] Calculate the first error value between the first predicted output and the first label sample, and the first error value is the difference between the predicted phase information and the calibrated phase information;

[0095] When the first error value exceeds a preset first error threshold, update the weight parameters of the BP neural network model through back propagation until the convergence condition is met, so as to complete the training of the BP neural network model.

[0096] Specifically, experimental simulation data and radar measured data under different working conditions are collected to form a first data training set. The experimental simulation data is derived from a multiphase flow simulation test in a laboratory, which can reflect the typical fluctuation characteristics of the drilling fluid flow under different disturbance conditions; the radar measured data is derived from continuous observation results of the millimeter wave radar at the outlet of the drilling fluid, which contains signal strength, amplitude and phase change under actual working conditions. The phase information corresponding to the calibration data of the laboratory simulation multiphase flow fluctuation is used as the first label sample for supervised neural network training.

[0097] The first data training set is filtered and preprocessed to filter out noise interference and abnormal points to ensure the stability of the training data. Then, multi-dimensional signal feature parameters that can represent amplitude-frequency characteristics and phase characteristics are extracted from the processed data, including signal strength, amplitude, peak frequency, initial phase, interference phase, intermediate frequency signal phase, complex signal phase, and propagation phase difference. The above feature parameters are combined to form a first input feature vector.

[0098] In the model calculation stage, the first input feature vector is input into the input layer of the BP neural network in sequence, and is subjected to weighted mapping and nonlinear transformation of the hidden layer, performs forward propagation calculation, and obtains a first prediction output. The prediction output corresponds to the predicted estimation value of the phase information of the input signal, which is used to simulate the echo phase characteristics of the drilling fluid surface under actual working conditions. Then, a first error value between the prediction output and the first label sample is calculated. The error value essentially reflects the fitting degree of the BP neural network under the current weight parameter, that is, the difference between the predicted phase information and the laboratory calibration phase information. When the error value exceeds the first error threshold, the system performs a back propagation mechanism: through layer-by-layer back propagation of error gradient and chain derivation, the connection weights and bias parameters in the network are corrected, so that the network can gradually approach the target label distribution. With the increase of the number of iterations, the first error value gradually decreases to the convergence condition range, and the BP neural network model completes the training process.

[0099] It can be found that, in the scheme provided by the embodiments of the present application, through the training process, since the training samples cover both the liquid surface stable state and various echo phase signals containing interference factors such as turbulence, bubbles and spray, the BP neural network model can not only gradually fit the phase characteristics corresponding to the real height of the liquid surface, but also model the phase disturbances caused by different interference factors in a differentiated manner. With the nonlinear mapping and error back propagation mechanism, the network model automatically extracts and separates the characteristic performances of the liquid surface intrinsic echo and external interference signals in the phase layer, reflects the interference effects of different interference factors in a differentiated manner, and thus realizes the simultaneous identification and characterization of the real component and the interference component of the liquid surface. In the reasoning stage, when the filtered echo signal is input into the BP neural network model, the network model can output the real liquid surface echo phase component and the phase disturbance component caused by different interference factors. In the subsequent calculation of the liquid level height, according to the identification result of the phase disturbance component by the model, the disturbance components are corrected or weighted and weakened in real time, so that the calculation process mainly depends on the phase characteristics of the real liquid surface echo phase component. This realizes the adaptive compensation of the liquid surface disturbance, thereby ensuring the stability and accuracy of the liquid level height estimation.

[0100] It should be noted that the third embodiment of the present application can also be an improvement based on any one or more of the first embodiment to the second embodiment.

[0101] Fourth embodiment

[0102] The fourth embodiment of the present application relates to a drilling fluid outlet flow anti-disturbance monitoring method, and the fourth embodiment is an improvement based on the first embodiment, and the specific improvement lies in that in the fourth embodiment of the present application, a specific implementation of training a physical information neural network model is provided, and specifically includes the following steps:

[0103] The radar wave attenuation data measured in the laboratory microwave darkroom and the radar measured data under different working conditions are collected as a second data training set, and the radar wave attenuation calibrated by the radar wave attenuation data is taken as a second label sample corresponding to the second data training set;

[0104] The second data training set is combined with the corresponding environmental parameters, radar working parameters and medium boundary conditions, second multi-dimensional signal feature parameters for characterizing the liquid surface echo propagation characteristics and phase response characteristics are extracted, and a second input feature vector is constructed;

[0105] The wave equation derived from the Maxwell equations and the medium boundary conditions is embedded in the physical information neural network model as a physical constraint, and a forward propagation calculation is performed based on the second input feature vector to obtain a second prediction output;

[0106] a loss value between the second predicted output and the second label sample, and a residual value of a physical constraint wave equation, and summing the two with a weight to obtain a second error value;

[0107] When the second error value exceeds a preset second error threshold, performing a back propagation to update a weight parameter of the physical information neural network model until a convergence condition is met, to complete training of the physical information neural network model.

[0108] Specifically, first, radar wave attenuation data measured in a laboratory microwave darkroom and radar measured data under different working conditions are collected to form a second data training set, and radar wave attenuation obtained by laboratory darkroom calibration is taken as a second label sample. The laboratory darkroom data can provide a standard attenuation benchmark of liquid surface reflection under the condition of no external interference, and the field measured data reflects the real attenuation characteristics of radar echoes under the action of multiple interference factors under complex working conditions. Therefore, in the model training process, a functional relationship between the radar wave attenuation coefficient and the interfered signal and the data set can be established; with the continuous accumulation of effective data, different liquid surface disturbance interference factors (such as bubbles, splashes, and turbulence) cause interference signals corresponding to different phase information to gradually input into the physical information neural network model, and when the error between the predicted output and the label exceeds the threshold limit of the hidden layer, the network continuously updates the parameters through back propagation, thereby realizing dynamic suppression of interference components and accurate approximation of real liquid surface echo attenuation characteristics.

[0109] Subsequently, the above-mentioned second data training set is combined with corresponding environmental parameters, radar working parameters, and medium characteristic parameters, from which multi-dimensional signal feature parameters representing liquid surface echo propagation characteristics and phase response characteristics are extracted to form a second input feature vector. The input vector not only contains data-driven signal features, but also introduces description information reflecting physical boundary conditions and propagation characteristics.

[0110] In terms of model construction, the physical information neural network model is composed of a feature extraction layer, a time sequence processing layer, a physical constraint layer, and an output layer. Its core feature is that a wave equation derived from Maxwell's equations and medium boundary conditions is embedded in the loss function as a physical constraint. In other words, when performing forward propagation calculation, the model not only learns the mapping relationship between the input and the output based on the input data, but also needs to satisfy the basic physical laws of electromagnetic wave propagation in a multi-medium environment. Through this mechanism, the predicted output is the radar wave attenuation coefficient corresponding to the liquid surface echo signal.

[0111] Further, the medium boundary condition is derived by taking the limit of the integral form of Maxwell's equations at the interface, which is used to establish the continuity constraints of the tangential components of the electric field and the magnetic field, and the continuity constraints of the normal components of the electric displacement vector and the magnetic induction vector at the interface between the air and the drilling fluid, thereby establishing a corresponding relationship between the macroscopic electromagnetic field distribution and the microscopic properties of the dielectric constant and the magnetic permeability of the drilling fluid medium.

[0112] Specifically, the medium boundary condition is derived by taking the limit of the integral form of Maxwell's equations at the interface. On the interface between the air and the drilling fluid, the following continuity constraints are satisfied:

[0113] The tangential components of the electric field and the magnetic field remain continuous at the interface, i.e., the tangential electric field intensity and the tangential magnetic field intensity on the air side and the drilling fluid side are equal.

[0114] The normal components of the electric displacement vector and the magnetic induction vector remain continuous at the interface, i.e., the normal component of the electric displacement vector satisfies the proportional relationship of the dielectric constant, and the normal component of the magnetic induction satisfies the proportional relationship of the magnetic permeability.

[0115] Through the above constraints, a clear corresponding relationship between the macroscopically observable electromagnetic field distribution and the microscopic properties of the drilling fluid medium (dielectric constant and magnetic permeability) can be established. In actual modeling, when electromagnetic waves are incident on the liquid surface, their reflection and transmission characteristics will be modulated by the drilling fluid medium properties. By introducing the above boundary conditions, the model can not only accurately represent the phase shift and amplitude attenuation of the liquid surface reflection signal, but also identify the dielectric constant disturbance caused by changes in bubble content, fluctuations in mud density, etc., thereby providing more physically constrained prior conditions in subsequent signal processing and disturbance compensation.

[0116] In the training phase, the loss value between the second prediction output of the model and the label sample is calculated to measure the accuracy of data fitting. At the same time, the residual value of the physically constrained wave equation at the prediction output is calculated to reflect the physical consistency. The weighted sum of the two gives the second error value. If this error value exceeds the preset threshold, the weight parameters of the neural network are updated through backpropagation to gradually correct the nonlinear mapping ability of the network. After multiple iterations, the model converges, ensuring both the fitting degree of the prediction result and the training data and the satisfaction of the basic physical laws of electromagnetic wave propagation.

[0117] It can be found that in the scheme provided by the embodiment of the application, the wave equation derived from the Maxwell equation set and the air and drilling fluid interface boundary condition is embedded in the loss function in the form of residual, and the attenuation amount calibrated in the microwave darkroom is used as a supervised label, and the radar measured data under different working conditions and the environmental parameters, the radar working parameters and the medium characteristic parameters are jointly trained, so that the radar wave attenuation coefficient output by the model meets the data fitting and the physical consistency at the same time. Therefore, on the one hand, the inaccuracy of the pure data driven model when migrating under the working conditions such as the bubble content, the temperature and the mud formula change is inhibited, and the generalization ability and the robustness under the complex well site environment are significantly improved; on the other hand, the pseudo solution that violates the electromagnetic boundary condition is avoided, the attenuation coefficient estimation is more stable and interpretable, and the dependence on large-scale labeled data is reduced. The physically consistent radar wave attenuation parameter provides more accurate input for the subsequent liquid level fluctuation and echo attenuation quantization model, so that the liquid level height inversion still maintains high precision in the strong interference scene.

[0118] It should be noted that the fourth embodiment of the application can also be an improvement on the basis of any one or more of the first embodiment to the third embodiment.

[0119] Fifth embodiment

[0120] The fifth embodiment of the application relates to a drilling fluid outlet flow disturbance monitoring method, and the fifth embodiment is an improvement on the basis of the first embodiment, and the specific improvement is that in the fifth embodiment of the application, a specific implementation manner of nonlinear inversion calculation of the liquid level height change amount is provided, that is, step S5 can further include the following steps:

[0121] Step S501: based on the ratio of the real part and the imaginary part of the phase information, and the inverse tangent value is calculated to obtain the liquid level phase change amount.

[0122] The filtered millimeter wave echo signal can be expressed as a complex signal z(n)=A(n)e jφtot(n) wherein A(n) represents the amplitude envelope of the echo signal at the sampling time n, j is an imaginary unit, used to construct a complex signal expression, so that the amplitude and phase of the echo signal can be represented in a unified complex plane, φ tot (n) represents the total phase of the echo signal, which includes the geometric phase φ geo (n) caused by the geometric height change of the liquid level and the instantaneous disturbance phase φ int (n) of the liquid level disturbance interference factor, and satisfies:

[0123] φ tot (n)=φ geo (n)+φ int (n)

[0124] The BP neural network model decomposes the filtered millimeter wave echo signal, identifies the disturbance phase φ caused by the liquid level disturbance interference factor, and reflects it to the complex form of the signal interference factor SIF: int (n), and reflects it to the complex form of the signal interference factor SIF:

[0125]

[0126] Wherein, <·> represents the average in the time window W.

[0127] Therefore, the equivalent average phase shift of the disturbance phase obtained from SIF is the overall influence of the liquid level disturbance factor on the phase in the time window scale, that is, the liquid level phase change amount Can be expressed as:

[0128]

[0129] Wherein, The real part of SIF is equal to cos(φ int ), The imaginary part of SIF is equal to sin(φ int ).

[0130] Step S502: Multiply the liquid level phase change amount by the light speed correlation coefficient to obtain the initial change value of the liquid level height.

[0131] Multiply the liquid level phase change amount obtained in step S501 by the light speed correlation coefficient To obtain the initial change value of the liquid level height:

[0132]

[0133] Wherein, Δh0 represents the uncorrected liquid level height change value, c is the light speed, which is used for phase to space distance calibration. The initial change value is the direct inversion result of the liquid level fluctuation to the electromagnetic wave phase modulation, but the influence of the tank damping and the radar wave attenuation has not been considered.

[0134] Step S503: Based on the tank damping coefficient, the radar wave attenuation coefficient and the liquid level fluctuation speed, the initial change value is nonlinearly corrected to weaken the measurement error caused by the interference factor, and the corrected liquid level height change value is obtained, wherein the liquid level fluctuation speed is obtained by spectrum analysis on the echo signal.

[0135] In order to eliminate the error caused by various interference factors in actual working condition, it is necessary to nonlinearly correct the initial change value Δh0. Therefore, the tank damping coefficient β, the radar wave attenuation coefficient α and the liquid level fluctuation speed v are introduced, the liquid level fluctuation and radar echo attenuation quantitative model is established, and the corrected liquid level height change value is obtained:

[0136]

[0137] The modified model can effectively represent the dynamic coupling relationship of the liquid level fluctuation velocity on the echo energy attenuation and phase shift, so that the inverted liquid level change quantity is more stable, and the anti-disturbance performance and measurement accuracy are improved.

[0138] It can be found that in the scheme provided by the embodiments of the application, the signal interference factor is introduced in the inversion process of the liquid level change quantity, and a nonlinear correction model is established in combination with the tank damping coefficient, the radar wave attenuation coefficient and the liquid level fluctuation velocity, so that the additional interference caused by the liquid level disturbance interference factor on the phase information can be quantified and compensated, thereby significantly improving the accuracy and robustness of the liquid level inversion. Therefore, the influence of the liquid level disturbance on the radar measurement accuracy is effectively suppressed, and a more stable and anti-disturbance monitoring effect on the drilling fluid outlet flow is realized.

[0139] It should be noted that the fifth embodiment of the application can also be improved on the basis of any one or more of the first to fourth embodiments.

[0140] In order to verify the effectiveness of the drilling fluid outlet flow anti-disturbance monitoring method described in the embodiments of the application, the monitoring results thereof are compared with those of the existing ultrasonic liquid level sensor. The comparison results are shown in FIG. 6, wherein the upper graph is the outlet flow change characteristic curve of the logging ultrasonic sensor, and the lower graph is the outlet liquid level change characteristic curve of the radar liquid level meter. As shown in the figure, when the buffer tank liquid level rises and falls suddenly, both sensors can reflect the change trend of the liquid level at the corresponding time point, indicating that the radar liquid level meter of the application has real-time and responsiveness in liquid level sudden change detection. Figure 3

[0141] However, from the amplitude change trend of the lateral mutation point, the ultrasonic liquid level sensor has inconsistent fluctuation characteristics with the radar liquid level meter at multiple time points, for example, at 630 min, 1266 min, 2761 min, 2991 min and 3221 min, the mutation amplitude or trend of the ultrasonic liquid level sensor deviates irregularly from the time before and after, while the radar liquid level meter maintains a more consistent mutation trend in the same interval. In addition, the mutation in the red box area in the comparison graph, the monitoring curve of the ultrasonic liquid level sensor shows more frequent and irregular fluctuations, while the curve of the radar liquid level meter remains stable as a whole, and can accurately reflect the real change of the liquid level mutation.

[0142] ​Therefore, compared with the existing ultrasonic sensor, the anti-disturbance monitoring method based on the radar liquid level meter of the present application not only ensures timely response at the mutation point, but also maintains high anti-interference and accuracy under the condition of violent fluctuation of the liquid surface, effectively avoids measurement distortion caused by interference factors, and thus improves the reliability of drilling fluid outlet flow monitoring.

[0143] In order to further verify the reliability of the method in the scene of violent fluctuation of the liquid surface, the radar liquid level meter is installed closer to the surge area where the mud is poured, and compared with the ultrasonic liquid level sensor, the monitoring results are as shown in Figure 4 The red curve is the measurement result of the ultrasonic liquid level sensor, and the blue curve is the measurement result of the radar liquid level meter.

[0144] As can be seen from the figure, during the change of the liquid surface, the overall change trend of the two sensors remains consistent, but there are obvious sharp burrs and unstable fluctuations in the curve of the ultrasonic liquid level sensor, which shows high sensitivity to the disturbance of the liquid surface. The monitoring curve of the radar liquid level meter under the same working condition remains smooth, and there is no obvious burr or abnormal mutation, which can stably reflect the real change of the liquid surface.

[0145] Therefore, the anti-disturbance monitoring method of the present application enables the radar liquid level meter to still maintain stable measurement under the condition of violent fluctuation of the liquid, and has stronger anti-disturbance ability. At the same time, the method has lower dependence on the installation position, and even if the monitoring is carried out in the area where the liquid disturbance is most obvious, it can still ensure high measurement accuracy and reliability, thereby significantly improving the applicability under complex working conditions.

[0146] The step division of the above various methods is only for the purpose of clear description, and when implemented, one step can be combined or some steps can be split and decomposed into multiple steps, as long as the same logical relationship is included, all within the protection scope of the present application; adding irrelevant modifications or introducing irrelevant designs in the algorithm or flow, but not changing the core design of the algorithm and flow, are within the protection scope of the present application.

[0147] In addition, some embodiments of the present application also provide an electronic device. The electronic device can be various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, etc. The electronic device can also be various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices.

[0148] The electronic device includes one or more processors, and a memory storing computer program instructions that, when executed, cause the processors to perform the drilling fluid outlet flow disturbance monitoring method provided by the above embodiments. Figure 5 An exemplary structural diagram of the electronic device is disclosed. The electronic device includes one or more processors 1101, a memory 1102, and interfaces for connecting components, including high-speed interfaces and low-speed interfaces. The components are connected to each other by different buses, and can be mounted on a common motherboard or otherwise mounted as needed. The processor can process instructions executed within the electronic device, including instructions stored in the memory or on the memory to display graphical information of a GUI on an external input / output device, such as a display device coupled to the interface. In some other embodiments, multiple processors and / or buses can be used with multiple memories and multiple memory, if necessary. Also, multiple electronic devices can be connected, each device providing part of the necessary operations. Among them, the components shown herein, their connections and relationships, and their functions are only examples, and are not intended to limit the implementation of the present application described and / or claimed herein.

[0149] The electronic device can also include an input device 1103 and an output device 1104. The processor 1101, the memory 1102, the input device 1103 and the output device 1104 can be connected by bus or other means, Figure 5 The input device 1103 and the output device 1104 are connected by bus in the middle.

[0150] The input device 1103 can receive input digital or character information, and generate key signal input related to user settings and function control of the electronic device, such as touch screen, keypad, mouse, trackpad, touchpad, pointing stick, one or more mouse buttons, trackball, joystick, etc. The output device 1104 can include a display device, an auxiliary lighting device (e.g., LED), and a tactile feedback device (e.g., a vibration motor), etc. The display device can include but is not limited to liquid crystal display, light emitting diode display and plasma display. In some embodiments, the display device can be a touch screen.

[0151] To provide for interaction with a user, the electronic device can be a computer. The computer has a display device (e.g., a cathode ray tube or LCD monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, etc.); and input from the user can be received in any form (e.g., acoustic input, speech input, tactile input, etc.).

[0152] In the embodiments of the present application, the computer program / instruction is stored on the computer readable medium, and the computer program / instruction is executed by the processor to implement the drilling fluid outlet flow anti-disturbance monitoring method provided by the above embodiments. The computer readable medium can be included in the electronic device described in the above embodiments, or can exist separately and not be assembled into the device. The computer readable medium carries one or more computer readable instructions.

[0153] The memory 1102 can be used to store non-transitory software programs, non-transitory computer executable programs and modules as a kind of non-transitory computer readable storage medium. The processor 1101 executes various functions and data processing of the server by running the non-transitory software programs, instructions and modules stored in the memory 1102, so as to implement the program instructions / modules corresponding to the method provided by any one or more of the above embodiments in the embodiments of the present application.

[0154] The memory 1102 can include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required by a function; the data storage area can store data created according to the use of the electronic device, etc. In addition, the memory 1102 can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory 1102 can optionally include a memory disposed remotely with respect to the processor 1101, and these remote memories can be connected to the electronic device through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0155] Note that the computer-readable medium described herein can be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. The computer-readable medium can be, for example but not limited to, a system, a device, or a computer program product embodied in one or more computer readable media embodying computer readable instructions, data structures, program modules, or other data. Computer-readable storage media include, at least, volatile memory, non-volatile memory, removable or non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer-readable storage media include, but are not limited to, phase-change memory, static random access memory, dynamic random access memory, other types of random access memory, read-only memory, electrically erasable programmable read-only memory, flash memory or other memory technology, compact disc read-only memory, digital versatile discs or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. Computer-readable storage media can be used to store program instructions, data structures, program modules and other data for use by a computer.

[0156] Computer-readable media includes permanent and non-permanent, removable and non-removable media, and can be implemented by any method or technology for information storage. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change memory, static random access memory, dynamic random access memory, other types of random access memory, read-only memory, electrically erasable programmable read-only memory, flash memory or other memory technology, read-only optical disc, digital versatile disc or other optical storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage device, or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0157] Computer program code for carrying out operations of the present application can be written in one or more programming languages or combinations of languages including object-oriented, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as the C programming language or similar programming languages. Program code can be executed entirely on a user's computer, partially on a user's computer, as a standalone software package, partially on a user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any kind of network, including a local area network or a wide area network, or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0158] In the above-described embodiments, all or part of the embodiments can be implemented by software, hardware, firmware, or any combination thereof. For example, application specific integrated circuits, general purpose computers or any other similar hardware devices can be used. In some embodiments, the software programs of the present application can be executed by a processor to implement the above steps or functions. Similarly, the software programs of the present application (including related data structures) can be stored in a computer-readable recording medium, such as a RAM memory, a magnetic or optical drive or a soft disk and the like. In addition, some steps or functions of the present application can be implemented by hardware, such as a circuit cooperating with a processor to perform the respective steps or functions.

[0159] The computer program product provided by the embodiments of the present application includes one or more computer programs / instructions, which, when executed by a processor, generate all or part of the processes or functions described in the embodiments of the present application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be magnetic media (such as floppy disk, hard disk, magnetic tape), optical media (such as DVD), or semiconductor media (such as solid state disk) and the like.

[0160] The flowchart or block diagram in the drawings illustrates the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram can represent a module, a segment, or a portion of code, which contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks noted in succession can in fact be executed substantially concurrently, or the blocks can sometimes be executed in reverse order, depending on the functionality involved. It will also be noted that each block of the block diagrams and / or flowchart illustrations, or combinations of blocks in the block diagrams and / or flowchart illustrations, can be implemented by special-purpose hardware-based systems that perform the specified functions or operations, or combinations of special-purpose hardware and computer instructions.

[0161] The scope of this application is defined by the appended claims rather than the foregoing description, and is therefore intended to encompass all variations falling within the meaning and scope of equivalents of the claims. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a device claim may also be implemented by a single unit or device in software or hardware. Terms such as "first," "second," etc., are used only for distinguishing descriptions and do not indicate any particular order, nor should they be construed as indicating or implying relative importance.

[0162] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily made by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims, and the above embodiments should be regarded as exemplary and non-limiting.

Claims

1. A method of monitoring an exit flow of a drilling fluid against disturbances, characterized in that, The method comprises the steps of: a millimeter wave radar liquid level meter is installed above the outlet buffer tank of the drilling fluid, millimeter wave radar waves are emitted to the liquid surface in the buffer tank, and reflected echo signals are received; the echo signals are filtered in time domain and frequency domain, the fusion weight of the time domain filter output and the frequency domain filter output is dynamically adjusted according to the signal-to-noise ratio, and a filter signal in which environmental interference echoes are filtered out is obtained; the filter signal is input into a pre-trained BP neural network model, and phase information affected by a liquid surface disturbance interference factor is output; the phase information and feature data representing environmental parameters, radar parameters and medium characteristic parameters are input into a pre-trained physical information neural network model, and a radar wave attenuation coefficient corresponding to the phase information is output; the radar wave attenuation coefficient and the phase information are input into a constructed liquid surface fluctuation and radar echo attenuation quantification model, and a liquid surface height change amount is obtained; the current liquid surface height is obtained by combining the liquid surface height change amount with a radar ranging formula, and the relative flow change of the drilling fluid in the outlet buffer tank is determined according to the current liquid surface height.

2. The method of monitoring the flow of a drilling fluid from an outlet against disturbances as claimed in claim 1, wherein, The step of filtering the echo signals in time domain and frequency domain, dynamically adjusting the fusion weight of the time domain filter output and the frequency domain filter output according to the signal-to-noise ratio, and obtaining a filter signal in which environmental interference echoes are filtered out, comprises the steps of: the echo signals are segmented and windowed, and then converted into frequency domain signals by fast Fourier transform; an adaptive weight update is performed on the frequency domain signals by using a frequency domain least mean square algorithm, and a frequency domain filter signal is obtained; a frequency domain error signal is calculated according to the difference between the expected response signal and the frequency domain filter signal; the frequency domain error signal is converted and fused with the output of the time domain adaptive filter to obtain an error-compensated time domain filter signal; the fusion weight is adaptively calculated according to the signal-to-noise ratio estimation result, the time domain filter signal and the frequency domain filter output are weighted and fused according to the fusion weight, and the filter signal in which environmental interference echoes are filtered out is obtained.

3. The method of monitoring the flow of a drilling fluid from an outlet against disturbances as claimed in claim 2, wherein, The step of converting the frequency domain error signal and fusing the output of the time domain adaptive filter to obtain an error-compensated time domain filter signal comprises the steps of: the frequency domain error signal is converted into a time domain error signal by inverse fast Fourier transform; based on the time domain error signal, the weight parameters of the time domain adaptive filter are updated by using a normalized least mean square algorithm; the input signal vector composed of the continuous sampling values of the echo signals in a preset sampling window is filtered by using the time domain adaptive filter with updated weight parameters, and the time domain filter signal is obtained.

4. The method of monitoring the flow of a drilling fluid from an outlet against disturbances as claimed in claim 1, wherein, The BP neural network model is trained, specifically comprising: experimental simulation data and radar measured data under different working conditions are collected as a first data training set, and the phase information corresponding to the calibration data of the laboratory simulated multiphase flow fluctuation is taken as the first label sample corresponding to the first data training set; first multi-dimensional signal feature parameters representing amplitude-frequency characteristics and phase characteristics are extracted from the first data training set for filtering and preprocessing, and a first input feature vector is formed; forward propagation calculation is performed based on the first input feature vector, and a first prediction output is obtained; calculating a first error value between the first predicted output and the first label sample, the first error value being a difference between predicted phase information and calibrated phase information; performing back propagation to update weight parameters of the BP neural network model when the first error value exceeds a preset first error threshold, until a convergence condition is met, to complete training of the BP neural network model.

5. The method of monitoring the flow of a drilling fluid from an outlet against disturbances as claimed in claim 1, wherein, training the physical information neural network model, specifically comprising: collecting radar wave attenuation data measured in a laboratory microwave darkroom and radar measured data under different working conditions as a second data training set, and taking radar wave attenuation data calibrated by the radar wave attenuation data as second label samples corresponding to the second data training set; combining the second data training set with corresponding environmental parameters, radar working parameters and medium boundary conditions, extracting second multi-dimensional signal feature parameters for representing liquid surface echo propagation characteristics and phase response characteristics to form a second input feature vector; embedding a wave equation derived from a Maxwell equation set and a medium boundary condition as a physical constraint in the physical information neural network model, and performing forward propagation calculation based on the second input feature vector to obtain a second predicted output; calculating a loss value between the second predicted output and the second label sample, and a residual value of the physical constraint wave equation, and obtaining a second error value by weighted summation of the two; performing back propagation to update weight parameters of the physical information neural network model when the second error value exceeds a preset second error threshold, until a convergence condition is met, to complete training of the physical information neural network model.

6. The method of monitoring the flow of a drilling fluid from an outlet against disturbances as claimed in claim 5, wherein, The medium boundary condition is derived by taking a limit of an integral form of the Maxwell equation set at the interface, and is used to respectively establish continuity constraints of tangential components of electric field and magnetic field, and continuity constraints of normal components of electric displacement vector and magnetic induction intensity vector at the interface between air and drilling fluid, so as to establish a corresponding relationship between macroscopic electromagnetic field distribution and microscopic characteristics of dielectric constant and magnetic permeability of drilling fluid medium.

7. The method of monitoring the flow of a drilling fluid from an outlet against disturbances as claimed in claim 1, wherein, The step of inputting the radar wave attenuation coefficient and the phase information into the constructed liquid surface wave and radar echo attenuation quantization model to obtain a liquid surface height change amount comprises: obtaining a liquid surface phase change amount by calculating an inverse tangent value based on a ratio of a real part and an imaginary part of the phase information; multiplying the liquid surface phase change amount by a light speed correlation coefficient to obtain an initial change value of the liquid surface height; performing nonlinear correction on the initial change value based on a tank damping coefficient, the radar wave attenuation coefficient and a liquid surface wave speed to weaken measurement errors caused by an interference factor, to obtain a corrected liquid surface height change amount, wherein the liquid surface wave speed is obtained by performing frequency spectrum analysis on the echo signal.

8. An electronic device, comprising: The electronic device comprises: one or more processors; and a memory storing computer program instructions which, when executed, cause the processors to perform the drilling fluid outlet flow disturbance-resistant monitoring method of any one of claims 1-7.

9. A computer readable storage medium having stored thereon a computer program and / or instructions, characterized in that, The computer program and / or instructions, when executed on the processor, implement the method of monitoring the resistance of the drilling fluid outlet flow to disturbances as claimed in any of claims 1 to 7.

10. A computer program product comprising computer programs and / or instructions, characterized in that, The computer program and / or instructions, when executed on the processor, implement the method of monitoring the resistance of the drilling fluid outlet flow to disturbances as claimed in any of claims 1 to 7.

Images