Downhole tool accelerometer fault detection method and detection device

Abstract

The invention discloses a fault detection method and a fault detection device for an accelerometer of a downhole tool. Taking an accelerometer and a gyroscope as state quantities, establishing a generalized nonlinear system for the downhole tool sensor model, and performing linearization processing to obtain a generalized linear variable parameter system; based on a T-N-L observer architecture, constructing a system state estimation error equation and a residual error generation equation, establishing a fault sensitivity and interference robustness evaluation system on the basis of the system state estimation error equation and the residual error generation equation, and determining observer parameters; and the sensor obtains a system state value, an estimated value is obtained through the observer, a residual value of a current system state quantity is calculated to judge an occurrence event, and a threshold value of sensor fault occurrence under the current event is calculated to judge a fault occurrence condition. A nonlinear model is adopted, the system dynamic state is described more accurately, the model is linearized, it is guaranteed that the model and the actual working condition have the high matching degree, and meanwhile a corresponding linear variable parameter system is obtained to reduce the calculated amount.

Description

Technical Field

[0001] The present invention belongs to the technical field of drilling tools, and in particular relates to a method and a device for detecting faults of an accelerometer of a downhole tool. Background Art

[0002] In oil drilling projects, to accurately monitor and optimize drilling operations, a multi-source, heterogeneous sensing system is required to collect key drilling parameters (including drilling pressure, torque, and rotational speed) in real time. Accelerometers, as one of the core sensors, capture dynamic parameters such as drill string vibration, downhole impact, and attitude changes in real time, providing irreplaceable dynamic sensing support for intelligent drilling status diagnosis, drill bit performance evaluation, and drilling parameter optimization decisions. Their high-frequency sampling capability for multi-dimensional vibration signals is irreplaceable in monitoring complex downhole conditions.

[0003] Due to the complex operating environment, particularly in the oil drilling field, sensors are exposed to harsh downhole conditions such as high temperatures, strong vibrations, and mud erosion, resulting in a persistently high failure rate for accelerometers. Downhole instrument failures can impact drilling accuracy at best, or even lead to significant financial losses and accidents at worst, requiring repairs and rig recovery. Therefore, timely detection of downhole accelerometer failures at their earliest stages is crucial for improving drilling tool reliability and reducing drilling costs.

[0004] To perform fault detection on downhole accelerometers, accurate modeling must first be performed. If the model is not accurate enough, the dynamic characteristics of the fault detection model will deviate substantially from the actual system, which will directly reduce the sensitivity of the detection and may even cause detection failure. Due to the multi-degree-of-freedom coupled dynamic characteristics of the drilling system, downhole accelerometers exhibit strong nonlinear time-varying characteristics and multi-state coupling characteristics, making it difficult to achieve complete physical modeling. Based on the above complex working conditions and modeling challenges, in order to improve drilling reliability and reduce operating costs, research on online fault detection of downhole accelerometers has been carried out. This has significant engineering value in breaking through the bottleneck of state perception under extreme working conditions and ensuring the safety of deep well drilling.

[0005] Chinese patent publication number CN115467651A, published on December 13, 2022, is titled "Method for Detecting Intermittent Faults in Accelerometers of Rotary Steerable Drilling Tool Systems." The application discloses a method for detecting intermittent faults in accelerometers of rotary steerable drilling tool systems. However, its shortcomings include: 1. The accelerometer system modeling fails to account for the uncertainty between the model and the actual system; 2. Fault sensitivity is not considered in the fault detection design.

[0006] Chinese patent publication number CN118551138A, published on August 27, 2024, is titled "A Method for Isolating Accelerometer Faults in a Rotary Steerable Drilling Tool Attitude Measurement Device Under Strong Noise." This application discloses a method for isolating accelerometer faults in a rotary steerable drilling tool system. However, its drawback is that the use of an EKF (Extended Kalman Filter) to process measurement noise requires knowledge of the noise's probability distribution, which is difficult to determine in practice.

[0007] The paper (Yang Y, Geng Y, Wang W. Sensor fault detection and isolation based on zonotopic Kalman filter for accelerometer system in drilling tools. Measurement. 2023 Feb 15;207:112329.) uses a ensemble Kalman filter (ZKF) to determine the bounds of parameter uncertainty and measurement noise for an accelerometer system. Finally, fault detection is achieved using a time-varying residual bound, which combines residuals with a dynamic threshold. However, this method uses a ZKF to obtain residual and boundary information, and only considers system robustness. This can lead to insufficient sensitivity for different systems, potentially resulting in missed fault reports. Summary of the Invention

[0008] In response to the above problems, the purpose of the present invention is to provide a method and device for detecting faults in accelerometers of rotary steerable drilling tools. First, a system model is constructed for the spatial layout of the downhole accelerometer to improve the fitting accuracy with the actual system and provide a guarantee for subsequent fault detection; the interference robustness and fault sensitivity of the system are designed to improve the sensitivity of fault detection; a fault occurrence judgment mechanism is designed to promptly alarm for accelerometer faults. In response to the constraint of low computing power in the high-temperature environment downhole, the accelerometer online fault detection device developed in conjunction with the device adopts a dual-circuit board architecture, which shares the computing pressure of the microcontroller through functional decoupling, and combines the fault information hierarchical reporting mechanism to obtain more fault information while ensuring real-time performance.

[0009] The present invention achieves the above-mentioned purpose through the following technical solutions: A method for detecting faults in a downhole tool accelerometer, comprising: Using the accelerometer and gyroscope as state variables, a generalized nonlinear system is established for the downhole tool sensor model, and linearization is performed to obtain a generalized linear variable parameter system. Based on the TNL observer architecture, the system state estimation error equation and residual generation equation are constructed. On this basis, a fault sensitivity and interference robustness evaluation system is established to determine the observer parameters. The sensor obtains the system state value, obtains the estimated value through the observer, calculates the residual value of the current system state quantity to determine the event, and calculates the threshold of sensor failure under the current event to determine the fault situation.

[0010] Furthermore, the method for constructing the generalized nonlinear system is to establish a geodetic coordinate system of the north-east and carrier coordinate system , the carrier coordinate system The axis coincides with the borehole axis, indicating the axial direction of the drilling tool. The plane formed by the axes is the cross section of the drilling tool; the accelerometer xyz axes are sequentially distributed on the xyz axes of the carrier coordinate system. When the accelerometer xyz axes coincide with the established carrier coordinate system, the vertical downward gravity acceleration is Decompose into x-axis components and the y-axis component on the circular plane at the bottom of the well and the z-axis component , which gives the following formula: ; in is the well inclination angle, is the gravity tool face angle, and the derivative of the gravity tool face angle is: ; in The angular velocity is measured by the gyroscope, and the system state vector is , the state equations of the accelerometer and gyroscope are established as: ; in: , ; The measurement equation of the system is as follows: ; In the formula and For the unit matrix, is the fault distribution matrix, is the measurement noise of the sensor, is the fault vector of the sensor; The state space expression of the downhole tool is as follows: ; in, 、 、 、 They are The system state vector, system output vector, measurement noise vector, and sensor fault vector at time , for The derivative of the system state vector at time , is a constant matrix of appropriate dimension and satisfies ,in for The dimension of for A polynomial nonlinear function of is the output matrix, is the noise distribution matrix, is the fault distribution matrix.

[0011] Furthermore, the linearization processing method is: for and ,definition ,in is a set of integers, is the Kronecker product. Based on the Kronecker operation rule and the generalized nonlinear system model, we get: ; vector , for dimensional real number set, defining operations and ,in ; For nonlinear functions exist Performing Taylor series expansion at , we get: ; in is the Taylor coefficient matrix, is the order of the Taylor series expansion, is a set of non-negative integers, , is the Lagrange residual coefficient matrix, , , , System status exist Estimated value of the moment.

[0012] Furthermore, the Lagrange residual is processed: ; in ,definition , ,in Operation, If is a vector, find the modulus of the vector. If is a matrix, find the maximum singular value of the matrix. is a constant, , , is the corresponding generator matrix, and finally we get: ; in , is an unknown uncertainty and ;Will Expanding it, we get: ; in is the variable parameter state matrix, is the unknown uncertainty after linearization.

[0013] Furthermore, the method of obtaining the generalized linear variable parameter system based on the downhole tool linearization system is to process the output equation: ; in 、 and They are the corresponding dimension matrices respectively; the linearized system of downhole tools is expanded and defined , , , ,in , , , , , , This results in the following augmented system: ; in , , , , , , , ; The Euler method is used to discretize the augmented system and the corresponding generalized linear variable parameter system model is obtained: ; in , , is the sampling time, 、 、 、 They are The system state vector, system output vector, measurement noise vector, sensor fault vector at the moment, for The system state vector at time t, is the system matrix, is the uncertainty term after linearization, is the output matrix, is the noise distribution matrix, is the fault distribution matrix; is a variable parameter matrix, expressed in the following form: ; in is included The polyhedron, for The number of vertices, for The variable parameter matrix corresponds to the vertices, is the weighting function, ,satisfy: .

[0014] Furthermore, the method for establishing the TNL observer is as follows: For the obtained generalized linear system, a TNL observer is established; The residual system equations are divided into the residual equations of the fault subsystem and the residual equations of the interference subsystem; Finite frequency domain The sensitivity design of the fault subsystem is carried out by using the sensitivity index; The radius index is used to conduct robust design of the interference subsystem and establish robustness index; The multi-objective optimization problem of sensitivity index and robustness index is transformed into the solution of linear matrix inequality, and the observer parameter matrix is ​​calculated. 、 、 .

[0015] Furthermore, for the obtained generalized linear system, the TNL observer equation is established to obtain Observation value of the system state vector at time and The observed value of the output vector at time The calculation formula of state estimation error is defined as , residual ,in for The observed value of the system state vector at time t can be obtained based on the obtained generalized linear system and TNL observer equation The state estimation error at time and residuals The calculation formula is: divide the residual equation into the fault subsystem and the interference subsystem, and obtain the subsystem The error of time and The residual at time 、 and 、 ; Using finite frequency domain index Conduct sensitivity design for fault subsystems, The sensitivity of the reaction system to failure; using the set-based The radius index is used to design the robustness of the interference subsystem. According to the interference subsystem, the ellipsoidal beam of the interference subsystem error can be obtained as , and satisfy: ; in 、 They are Moment and time The center vector of 、 They are Moment and time The generator matrix of for The parameter matrix composed of the maximum values ​​of each element in; for The state vector at the moment is the center vector of the ellipsoidal beam, for The generator matrix of the ellipsoidal beam of the state vector at the moment; for The generation matrix of the ellipsoidal beam of the momentary noise variation; for The generation matrix of the noise ellipsoidal beam is measured at every moment; for The generation matrix of the ellipsoidal beam of the momentary fault variation; Finally, the multi-objective optimization problem of sensitivity and robustness indicators is transformed into the solution of linear matrix inequalities to obtain the corresponding observer parameter matrix.

[0016] Furthermore, the ellipsoidal beam of the interference residual subsystem is reduced in dimension by the following method: dimensional ellipsoidal beam and integers , for -dimensional hypercube, and satisfies ,definition For the general The matrix obtained by arranging the column vectors of the matrix in descending order of the Euclidean norm is ,in for Before the Matrix List, For the remaining columns, is a diagonal matrix, and ,in express No. elements, express The absolute value of elements, , .

[0017] Furthermore, the fault occurrence evaluation mechanism is to first calculate , and satisfy: ; in Interference subsystem The moment residual ellipsoidal bundle, for The center vector of for The generation matrix of ; the fault occurrence mechanism judgment method is: ; in , for Moment residual No. A quantity, for The absolute value of for The absolute value of the maximum boundary, It can be calculated by the following formula ; in for Moment Matrix No. OK, To calculate the vector The 1-norm of .

[0018] A downhole tool accelerometer fault detection device includes a drill collar body, in which a first circuit board and a second circuit board are installed, the first circuit board includes a first microprocessor and an auxiliary gyroscope, the second circuit board includes a communication interface and a second microprocessor, the accelerometer measurement value is transmitted to the first microcontroller and the second microcontroller via the communication interface; the first microcontroller of the first circuit board obtains the observation value of each state based on the auxiliary gyroscope and accelerometer data through a TNL observer and an LPV system model obtained after linearization; the observation value obtained by the first microcontroller is transmitted to the second circuit board via SPI communication; the second microcontroller calculates the residual value of the current system state quantity based on the observation value calculated by the first circuit board and the measurement value obtained by the sensor to determine the occurrence of an event, and calculates the threshold value of the sensor fault under the current event to determine the occurrence of the fault; the fault monitoring device sends a fault alarm signal to the ground, and the fault alarm signal is the residual information that exceeds the multiple of the threshold.

[0019] The beneficial effects of the present invention are as follows: The present invention provides a method and device for detecting faults in downhole tool accelerometers. By constructing a generalized nonlinear system model using the accelerometer and gyroscope as state variables, the system accurately characterizes the sensor's dynamic characteristics and improves the model's compatibility with actual operating conditions. The generalized nonlinear system is linearized to ensure a high degree of compatibility between the model and actual operating conditions, while also generating a corresponding linear variable parameter system to reduce computational complexity. An ellipsoidal beam set characterization method is used to quantitatively describe the boundaries of uncertain information, such as quantization error and noise, to obtain more accurate thresholds. A finite frequency domain sensitivity index is introduced to optimize the ellipsoidal beam, improving fault detection sensitivity. Finally, a fault detector is designed to enable online fault detection.

[0020] The detection device utilizes a hierarchical fault information upload mechanism and includes two sets of data processing boards. The first and second data processing boards work in tandem, respectively, to provide fault detection and model linearization functions. This reduces the computing power requirements of the microcontroller, meeting the low-power, high-reliability requirements of underground high-temperature environments. This dual-processor collaboration reduces computing power pressure during linearization, facilitating online deployment during drilling. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of a method for detecting a fault of a downhole tool accelerometer according to the present invention; Figure 2 Schematic diagram of drilling tool posture parameters; Figure 3 To stabilize the platform sensor space layout diagram; Figure 4 It is the decomposition diagram of the x-axis component of the carrier coordinate system; Figure 5 It is the component decomposition diagram on the circular plane at the bottom of the well; Figure 6 This is a flow chart of the application of the present invention to a rotary steerable drilling tool attitude measurement system; Figure 7 This is a flow chart of another embodiment of the present invention; Figure 8 The size of the fault is ( ) when the accelerometer Join on the axis Constant bias fault of large and small size; Figure 9 The residual and threshold result diagram after adding the fault; Figure 10 This is the fault alarm result diagram; Figure 11 It is a structural diagram of a downhole accelerometer fault detection device; Figure 12 This is a specific module relationship diagram of the downhole accelerometer fault detection device.

[0022] Among them: 1-first end cover, 2-drill collar body, 3-second circuit board, 301-first communication protection module, 302-second microcontroller, 303-storage module, 304-power interface, 305-communication interface, 306-second communication protection module, 4-first circuit board, 401-communication protection module, 402-first microcontroller, 403-auxiliary gyroscope, 5-second end cover. DETAILED DESCRIPTION

[0023] In drilling tool fault detection, due to the complex relationship between various devices such as sensors, it is usually only necessary to use the accelerometer as the state quantity and the gyroscope as the noise to establish a linear model for processing. However, the system model established in this way is quite different from the actual device, resulting in low fault detection accuracy. How to establish a system model with high fitting accuracy and sensitivity to faults will become the focus and difficulty. The focus of this patent is a downhole accelerometer fault detection method and detection device, which uses accelerometers and gyroscopes to establish a nonlinear system for downhole tool sensors, and obtains a new accelerometer fault detection model through the linearization method of generalized nonlinear systems to improve the accuracy and reliability of fault detection.

[0024] like Figure 1 As shown, a method for detecting a downhole tool accelerometer fault includes the following steps: S101: Using the accelerometer and gyroscope as state variables, a generalized nonlinear system is established for the downhole tool sensor model. The nonlinear system is linearized to obtain a corresponding generalized linear variable parameter (LPV) system.

[0025] Gyroscopes are usually added to assist in the attitude measurement of accelerometer systems. How to establish a mathematical model that includes all sensors will be a key and difficult point in sensor fault detection.

[0026] The definition of various posture parameters of drilling tools in the well is as follows: Figure 2 As shown. The inclination is the angle between the borehole direction line and the plumb line. The inclination describes the angle at which the wellbore deviates from the vertical. A zero inclination indicates a vertical well. is the gravity tool face angle, which is the high side direction line Go clockwise to angle.

[0027] The spatial layout of the attitude measurement system sensors is as follows: Figure 3 As shown, the sensor layout includes a triaxial accelerometer and a gyroscope. A centrosymmetric sensor layout was chosen because the rotation of the drill string inevitably introduces rotational acceleration, and this layout minimizes the impact of this acceleration. Therefore, during dynamic measurements, vibration noise can be considered the only source of interference from the gravity component.

[0028] Euler angle method is used to express the attitude matrix. Establish the North-East geodetic coordinate system and carrier coordinate system . The carrier coordinate system The axis coincides with the borehole axis, indicating the axial direction of the drilling tool. The plane formed by the axes represents the cross-section of the drill string. Accelerometers are distributed along the axes of the carrier coordinate system. Both the earth coordinate system and the carrier coordinate system conform to the right-hand rule. When rotating about the coordinate axis starting from the origin, counterclockwise rotation is positive and clockwise rotation is negative. The carrier coordinate system can be obtained by converting the earth coordinate system using the current attitude angle of the drill string.

[0029] According to the measurement principle of the accelerometer with central symmetric layout, when the xyz axis of the accelerometer coincides with the established carrier coordinate system, according to the vector decomposition rule as follows Figure 4 and Figure 5 As shown, the vertical downward acceleration of gravity can be Decomposed into the x-axis component in the carrier coordinate system and the component on the circular plane at the bottom of the well, which is as follows Figure 5 Decompose into y-axis components and the z-axis component , which gives the following formula: ; Derivative of the gravity tool face angle: ; in is the angular velocity measured by the gyroscope, let , the state equations of the accelerometer and gyroscope are established as: ; in: , .

[0030] During the drilling process, steerable drilling tools often operate in harsh environments with high temperature, high pressure, and strong vibration. The measurements of accelerometers and gyroscopes often contain a lot of noise, and the sensors are easily damaged and malfunction. The above model can be used to detect accelerometer and gyroscope faults. The measurement equation of the system is given as follows: ; In the formula and For the unit matrix, is the fault distribution matrix, is the measurement noise of the sensor, is the fault vector of the sensor.

[0031] In summary, the state space expression of the attitude measurement system is as follows: ; It is a generalized nonlinear system, which needs to be linearized first for subsequent noise filtering and fault detection. 、 、 、 They are The system state vector, system output vector, measurement noise vector, and sensor fault vector at time , for The derivative of the system state vector at time , is a constant matrix of appropriate dimension and satisfies ,in for The dimension of for A polynomial nonlinear function of is the output matrix, is the noise distribution matrix, is the fault distribution matrix.

[0032] In order to realize the fault detection of the sensor, it is necessary to linearize the above-mentioned generalized nonlinear system and transform it into a linear system. This ensures that the model has a high degree of matching with the actual working conditions and obtains the corresponding linear variable parameter system to reduce the amount of calculation. and ,definition ,in is a set of integers, Expressed as a Kronecker product, then: ; Taking into account the Kronecker operation rule and the generalized nonlinear system model, we can obtain: ; For vector , for -dimensional real number set, acting on the function Operation Defined as and ,in For nonlinear functions exist Perform Taylor series expansion at , and we get: ; in is the Taylor coefficient matrix, is the order of the Taylor series expansion, The larger the value, the higher the linearization accuracy. is a set of non-negative integers, , is the Lagrange residual coefficient matrix, , , , System status exist Estimated value of the moment. Any number between 0 and 1. for the reason A diagonal matrix of arbitrary numbers, thus obtaining for and Any vector between .

[0033] In order to further improve the fitting accuracy, the Lagrange residual is processed: ; in .definition , ,in Operation, If is a vector, find the modulus of the vector. If is a matrix, find the maximum singular value of the matrix. is a constant, , , is the corresponding generator matrix, from which we get: ; The unknown vector and , for dimensional real number set, which gives the following form: ; in , is an unknown uncertainty and .

[0034] Will Expand it and we get: ; in is the variable parameter state matrix, is the unknown uncertainty after linearization, when When it reaches infinity, the uncertainty term will be zero.

[0035] In order to meet the subsequent observer design requirements, the linearized equation needs to be converted into a generalized linear variable parameter (LPV) system and the output equation needs to be processed: ; in 、 and are matrices of corresponding dimensions respectively.

[0036] Then we augment the system and define , , , ,in , , , , from which we can get the following augmented system: ; in , , , , , , , .

[0037] By discretizing the augmented system using the Euler method, we can obtain the corresponding generalized linear variable parameter (LPV) system model: ; in , , is the sampling time, 、 、 、 They are The system state vector, system output vector, measurement noise vector, sensor fault vector at the moment, for The system state vector at time t, is the system matrix, is the uncertainty term after linearization, is the output matrix, is the noise distribution matrix, is the fault distribution matrix, is the augmented constant matrix.

[0038] is a variable parameter matrix, which can be expressed in the following form: ; in is included The polyhedron, for The number of vertices, for The variable parameter matrix corresponds to the vertices, is the weighting function, ,satisfy: .

[0039] S102: Based on the TNL observer architecture, the system state estimation error equation and residual generation equation are constructed. On this basis, a fault sensitivity and interference robustness evaluation system is established to determine the observer parameters.

[0040] The TNL observer design method is as follows: for the obtained generalized linear system, the TNL observer is designed as follows: ; in, 、 They are The observed value of the system state vector and the observed value of the output vector at time t, for The observed value of the system state vector at time t, 、 、 is the observer matrix to be designed and needs to satisfy .

[0041] The state estimation error is: ; The residual is: ; The residual system equation is as follows: ; in and They are Moment and The state estimation error at time , for The system residual at time , and They are The noise variation and sensor failure variation are measured at all times.

[0042] In order to design the fault sensitivity and interference robustness of the residual system equation, the equation is divided into a fault subsystem and a interference subsystem and designed separately.

[0043] The faulty subsystem is: ; in and They are Moment and The state estimation error of the faulty subsystem at time t, for The system residual of the faulty subsystem at time instant t.

[0044] The interfering subsystems are: ; in and They are Moment and The state estimation error of the interference subsystem at time , for The system residual of the moment-to-moment interference subsystem.

[0045] Finite frequency domain The sensitivity index of the fault subsystem is designed based on the sensitivity index: ; in The sensitivity of the reaction system to failures, The larger the value, the higher the sensitivity.

[0046] Using a membership-based The radius index is used to design the robustness of the interference subsystem. First, the ellipsoidal beam of the interference residual subsystem can be obtained based on the interference subsystem: ; in represents the Minkowski sum, two sets and The Minkowski sum of is defined as . Represents a linear map, an ellipsoidal beam With a matrix The linear mapping of can be obtained by the standard matrix product operation: . Interference subsystem The moment residual ellipsoidal bundle, for The center vector of for The generator matrix of . Interference subsystem The ellipsoidal bundle of the state estimation error at the moment, for The center vector of for The generator matrix of Interference subsystem Ellipsoid bundle of the state estimation error at time instant. for The parameter matrix composed of the maximum values ​​of each element in . for The ellipsoidal bundle of the state vector at time instant, for The center vector of for The generator matrix of . for The ellipsoidal beam of the momentary noise variation, for The generator matrix of . for The ellipsoidal beam that measures the noise at all times, for The generator matrix of . for Ellipsoid beam of the moment fault variation, for The generator matrix of .

[0047] further, , and satisfy: The characterization of noise and linearization error bounds of the ellipsoidal beam is achieved.

[0048] definition of The radius is , then we can get The radius index is: ; in and Interference subsystem Moment and The error of state estimation at the moment radius. is a given scalar, is the influence of the disturbance on the state estimation error. The radius will eventually converge to .

[0049] The multi-objective optimization problem of sensitivity and robustness indicators is transformed into the solution of linear matrix inequalities: ; in and , are the weights of sensitivity index and robustness index respectively. By optimizing the above objectives, we can get: ; in 、 and Substitute the corresponding vertex The observer parameter matrix obtained by optimization.

[0050] S103: The sensor obtains the system measurement value, and calculates the residual value of the current system state quantity through the estimated value obtained by the observer to determine the event, and calculates the threshold of the sensor fault under the current event to determine the fault situation.

[0051] first, , and satisfy: .

[0052] Furthermore, the fault mechanism judgment method is as follows: According to the residual system equation, when the system operates normally, the residual system equation can be approximated to the interference subsystem, then the residual ellipsoidal beam will always contain the true residuals, i.e. If a fault occurs, Will no longer be satisfied. In view of this, a residual evaluation mechanism based on ellipsoidal beam is given: ; in , for Moment residual No. A quantity, for The absolute value of for The absolute value of the maximum boundary, It can be calculated by the following formula: ; in for Moment Matrix No. OK, is a vector The 1-norm of .

[0053] In another embodiment, the rotary steerable drilling tool is a common downhole drilling tool, such as Figure 2 As shown, a downhole tool accelerometer fault detection method is applied to a rotary steerable drilling tool, and a rotary steerable drilling tool accelerometer fault detection method is provided, comprising the following steps: S201: Based on the physical and spatial layout characteristics of the sensors of the rotary steerable drilling tool attitude measurement system, a generalized polynomial nonlinear system model of the accelerometer and gyroscope is established, and then linearized to obtain the corresponding generalized linear variable parameter (LPV) system.

[0054] Gyroscopes and accelerometers are installed in rotary steerable drilling tools to obtain drilling information such as penetration rate and gravity tool face angle.

[0055] Reference Figure 3 , the spatial layout of gyroscopes and accelerometers in rotary steerable drilling tools is as follows Figure 3 As shown, its layout is centrally symmetrical, which minimizes the impact of rotational acceleration on the accelerometer.

[0056] Based on the above sensor spatial layout, a system model for the gyroscope and accelerometer can be established. Due to the coupled state, the model is a generalized polynomial nonlinear system model. For subsequent fault detection design, the model needs to be linearized. This linearization method improves the model's accuracy in fitting the actual system, ultimately resulting in a generalized linear variable parameter (LPV) system.

[0057] The mathematical model of the accelerometer and gyroscope of the rotary steerable drilling tool attitude measurement system is: ; in is the state vector, 、 and Gravity exist 、 、 The component of gravity on the axis, is the angular velocity. 、 、 、 They are The system state vector, system output vector, measurement noise vector, and sensor fault vector at time , for The derivative of the system state vector at time , is a constant matrix of appropriate dimension and satisfies ,in for The dimension of for A polynomial nonlinear function of is the output matrix, is the noise distribution matrix, is the fault distribution matrix. The system parameters are as follows: , , , , .

[0058] After linearization of the generalized nonlinear system, we get: ; where it is defined and , , Expressed as Kronecker product, the parameters after linearization are: 、 、 , ,in , , , , are the augmentations of the original state vector, output vector, measurement noise vector, and fault vector, respectively, indicating The state vector, output vector, measurement noise vector and fault vector at time t. is a known generalized matrix. is the variable parameter matrix. Uncertainty matrix resulting from system linearization. 、 are the measurement noise distribution matrix and fault distribution matrix respectively.

[0059] Pick , then the specific forms of each parameter are as follows: , , , , , , , , , .

[0060] in , , , , . , is the uncertainty term after linearization, where , express The number of rows here is 4, ,set up is the ellipsoidal beam of the initial state vector, is the state vector The central vector of is also the estimated value of the state vector, for The generator matrix of . 、 、 and They are 0, 1, 1, 0 respectively. is an unknown matrix of suitable dimensions, only Right now The maximum singular value of is less than or equal to 1.

[0061] Using the Euler method for discretization we get: ; in is a variable parameter matrix and ,in for The vertices corresponding to the variable parameter matrix, To include variable parameters polytopes The number of vertices, , is a weighted function that satisfies . , is the sampling time, 、 、 、 They are The system state vector, system output vector, measurement noise vector, sensor fault vector at the moment, for The system state vector at time t, is the system matrix, is the uncertainty term after linearization, is the output matrix, is the noise distribution matrix, is the fault distribution matrix. From this, the specific forms of each parameter can be obtained as follows: , ,according to The specific form of the matrix shows that it is mainly affected by 、 and The influence of these three parameters, so the corresponding vertices, due to , , Each parameter is bounded, so we get the vertex and weight functions.

[0062] S202: For the generalized LPV system of the rotary steerable drilling tool, a TNL observer is established to obtain the system's state estimation error equation and residual generation equation. On this basis, a fault sensitivity and interference robustness evaluation system is established to determine the observer parameters.

[0063] Based on the TNL observer architecture, the unified state estimation error equation and residual generation equation are established. Indicators and The radius index is used to perform fault sensitivity and disturbance robustness analysis to determine the observer parameters.

[0064] The observer design method for the generalized LPV system of the rotary steerable drilling tool is as follows: the TNL observer is designed as follows: ; in, 、 They are The observed value of the system state vector and the observed value of the output vector at time t, for The observed value of the system state vector at time t, 、 、 is the observer matrix to be designed and needs to satisfy .

[0065] The state estimation error at time , residual , Measure the noise change at all times , fault variation , then the residual system equation of the rotary steerable drilling tool attitude measurement system is as follows: ; Its fault subsystem and interference subsystem are as follows: The faulty subsystem is: ; The interfering subsystems are: ; According to the finite frequency domain Indicators are used to design the sensitivity of fault subsystems: ; Furthermore, robust design is carried out using Radius index, for this purpose, the definition of a one-dimensional ellipsoidal beam is given as: ; in is the center vector of the ellipsoidal beam, is the Minkowski sum operation, is the generating matrix of the ellipsoidal beam, is the unit hypercube, is the dimension of the polyhedron.

[0066] Then, according to the interference subsystem, the corresponding ellipsoidal beam can be obtained as: ; in Interference subsystem Moment residual ellipsoidal bundle. and Interference subsystem Moment and Ellipsoid bundle of the state estimation error at time instant. for The parameter matrix composed of the maximum values ​​of each element in . for The ellipsoidal bundle of the state vector at time instant, for The center vector of for The generator matrix of . for The ellipsoidal beam of the momentary noise variation, for The generator matrix of . for The ellipsoidal beam that measures the noise at all times, for The generator matrix of . for Ellipsoid beam of the moment fault variation, for The generator matrix of .

[0067] In this embodiment, 、 、 、 is a constant and does not change with time, then: , , , .

[0068] Furthermore, the ellipsoidal beam calculation method of the interference residual subsystem can be obtained as follows: according to As can be seen from the calculation formula, there is a problem of increasing dimension in the operation. For this reason, the matrix needs to be reduced in dimension. The dimension reduction process is given below: consider dimensional ellipsoidal beam and integers , for -dimensional hypercube, and satisfies ,definition For the general The matrix obtained by arranging the column vectors of the matrix in descending order of the Euclidean norm is ,in for Before the Matrix List, For the remaining columns, is a diagonal matrix, and ,in express No. elements, express The absolute value of elements, , .

[0069] definition of The radius is , then we can get The radius index is: ; The multi-objective optimization problem of sensitivity and robustness indicators is transformed into the solution of linear matrix inequalities, which can be solved using commercial solvers: ; in and , are the weights of sensitivity and robustness indicators respectively.

[0070] The parameter matrix corresponding to each vertex is obtained by solving: ; in 、 and Vertex The corresponding observer parameter matrix.

[0071] S203: Detecting accelerometer The main target is the axis constant deviation fault. The system measurement value is obtained through the sensor, and the residual value of the current system state quantity is calculated through the estimated value obtained by the observer. Compare with the threshold to determine whether a fault has occurred.

[0072] The system measurement value is obtained through the sensor, and the estimated value generated by the observer is used to calculate the residual value of the current system state quantity. The system residual is described by the ellipsoidal beam, and the residual boundary when there is no fault can be obtained. This boundary is identified as the threshold. When the residual exceeds the threshold, a fault occurs, thereby achieving the purpose of fault detection.

[0073] The estimated value of the state vector can be obtained by the obtained observer parameter matrix and the above-mentioned observer equation of the rotary steering drilling tool attitude measurement system. The corresponding residual can be obtained. Further, when the system operates normally, the residual system equation can be approximated to the interference subsystem, then the residual ellipsoidal beam will always contain the true residuals, i.e. If a fault occurs, Will no longer be satisfied. In view of this, a residual evaluation mechanism based on ellipsoidal beam is given: ; in , for Time accelerometer The residual of the axis, for The absolute value of for Fault-free residual The absolute value of the maximum boundary, It can be calculated by the following formula: ; in for Moment generation matrix The corresponding row, for The 1-norm of .

[0074] like Figure 4 As shown, the residual calculation method and fault alarm logic are as follows: S301: Model initialization, given the observer parameters designed according to the above method, set initialization.

[0075] S302: Calculate the residual of the generalized LPV system of the needle rotary steerable drilling tool , ellipsoidal beam set of interference subsystem and .

[0076] S303: By Calculating the threshold .

[0077] S304: Calculation fault alarm signal .

[0078] like Figures 8-10 As shown, Figure 8 The size of the fault is ( ) when the accelerometer Join on the axis Constant deviation fault of different sizes. Figure 9 The residual and threshold result diagram after adding the fault. Figure 10 This is the fault alarm result diagram.

[0079] It can be seen from the test results that the proposed method achieves high-precision fitting of nonlinear systems. The designed sensor fault detection can separate the faulty ellipsoidal beam from the fault-free ellipsoidal beam. The ellipsoidal beam can well contain the actual residual signal. The correct fault alarm can be achieved through the inclusion relationship between the actual residual signal and the ellipsoidal beam interval, thus completing the design of fault detection under the set membership framework.

[0080] It should be noted that the matrix dimensions mentioned in this patent have appropriate dimensions by default. The above content is a detailed description made in combination with a specific rotary steerable drilling tool stable platform model, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions.

[0081] like Figure 11 As shown, another embodiment provides an accelerometer online fault detection device suitable for underground high-temperature, low-computing-power environments. This device utilizes a dual-circuit board architecture to reduce microcontroller load. The device consists of a drill collar body and two connecting components: a first end cap 1 connects to the front drill collar motor, a second end cap 5 connects to the rear vibration mechanism, and a second circuit board 3 and a first circuit board 4 are integrated within the drill collar body 2, forming a distributed data processing system.

[0082] The accelerometer online fault detection device proposed in this embodiment has a hardware structure consisting of a first circuit board and a second circuit board. The first circuit board 4 includes a communication protection module 401, a first microcontroller 402, and an auxiliary gyroscope 403. The communication protection module 401 is connected to the first microcontroller 402 to ensure controller security. The communication protection module can be specifically an isolation barrier that isolates external electrical devices from the first microcontroller to prevent damage to the first microcontroller. The auxiliary gyroscope 403 provides rotational speed data. The second circuit board 3 integrates the first and second communication protection modules 301, 306, the second microcontroller 302, the storage module 303, the power interface 304, and the communication interface 305. The communication interface 305 and the power interface 304 are responsible for external communication and power supply, respectively. Accelerometer data is transmitted to the second microcontroller 302 via the communication interface 305 and to the first microcontroller 402 via the SPI bus. The first and second communication protection modules ensure data exchange security and prevent damage to the electrical interfaces. The first microcontroller 402 generates estimated values ​​for each state based on the measured data from the auxiliary gyroscope and the accelerometer transmitted from the communication interface 305, as well as the TNL observer parameters calculated in the external computer. These values ​​are then transmitted to the second circuit board via SPI. The second microcontroller 302 executes the fault detection algorithm based on the input from the first circuit board 4, generating residuals and threshold boundaries, which are stored in the memory module.

[0083] The collaborative working mode of the first and second circuit boards is mainly that the first circuit board obtains rotational speed data through the auxiliary gyroscope, and the accelerometer data is obtained from the outside through the communication interface. The state estimation value is obtained through the linearized model and the given TNL observer and then transmitted to the second circuit board via SPI. The linearized model is a generalized linear variable parameter system obtained by mathematically modeling the downhole tool sensor model and performing linearization processing. The TNL observer constructs the system state estimation error equation and the residual generation equation, and establishes a fault sensitivity and interference robustness evaluation system based on the system state estimation error equation and the residual generation equation to determine the observer parameters; the second circuit board obtains the system measurement value based on the sensor, and calculates the residual value of the current system state quantity through the estimated value obtained by the first circuit board. Determine the occurrence of an event and calculate the threshold of sensor failure under the current event to determine the fault situation.

[0084] The specific data processing method of the first microcontroller and the second microcontroller is the same as the specific steps of the downhole tool accelerometer fault detection method, and will not be repeated here.

[0085] In view of the bandwidth limitation of the pulser, a graded signal upload mechanism is designed, that is, if the residual exceeds the threshold N times, the value N is uploaded, and the complete data is stored locally synchronously for the ground system to call and analyze, realizing the dual functions of fault level determination and data tracing.

[0086] To meet the needs of low-computing-power scenarios underground, an accelerometer online fault detection device was designed. It consists of two circuit boards: the first and second. The first circuit board performs state estimation and transmits the estimated value to the second circuit board, which calculates residuals and margins. The calculated residuals and margins determine whether a fault has occurred, and faults are graded and uploaded based on the residual size to obtain more fault information.

Claims

1. A method for detecting faults in downhole tool accelerometers, characterized in that include: Using the accelerometer and gyroscope as state variables, a generalized nonlinear system is established for the downhole tool sensor model, and linearization is performed to obtain a generalized linear variable parameter system. Based on the TNL observer architecture, the system state estimation error equation and residual generation equation are constructed. On this basis, a fault sensitivity and interference robustness evaluation system is established to determine the observer parameters. The sensor obtains the system state value, obtains the estimated value through the observer, calculates the residual value of the current system state quantity to determine the event, and calculates the threshold of sensor failure under the current event to determine the fault situation.

2. A downhole tool accelerometer fault detection method according to claim 1, characterized in that The method for constructing the generalized nonlinear system is to establish a geodetic coordinate system of the north-east. and carrier coordinate system , the carrier coordinate system The axis coincides with the borehole axis, indicating the axial direction of the drilling tool. The plane formed by the axes is the cross section of the drilling tool; the accelerometer xyz axes are sequentially distributed on the xyz axes of the carrier coordinate system. When the accelerometer xyz axes coincide with the established carrier coordinate system, the vertical downward gravity acceleration is Decompose into x-axis components and the y-axis component on the circular plane at the bottom of the well and the z-axis component , which gives the following formula: ； in is the well inclination angle, is the gravity tool face angle, and the derivative of the gravity tool face angle is: ； in The angular velocity is measured by the gyroscope, and the system state vector is , the state equations of the accelerometer and gyroscope are established as: ； in: , ; The measurement equation of the system is as follows: ； In the formula and For the unit matrix, is the fault distribution matrix, is the measurement noise of the sensor, is the fault vector of the sensor; The state space expression of the downhole tool is as follows: ； in, 、 、 、 They are The system state vector, system output vector, measurement noise vector, and sensor fault vector at time , for The derivative of the system state vector at time , is a constant matrix of appropriate dimension and satisfies ,in for The dimension of for A polynomial nonlinear function of is the output matrix, is the noise distribution matrix, is the fault distribution matrix.

3. A method for detecting faults in a downhole tool accelerometer according to claim 2, characterized in that The linearization processing method is: for and ,definition ,in is a set of integers, is the Kronecker product. Based on the Kronecker operation rule and the generalized nonlinear system model, we get: ； vector , for dimensional real number set, defining operations and ,in ; For nonlinear functions exist Performing Taylor series expansion at , we get: ； in is the Taylor coefficient matrix, is the order of the Taylor series expansion, is a set of non-negative integers, , is the Lagrange residual coefficient matrix, , , , System status exist Estimated value of the moment.

4. A downhole tool accelerometer fault detection method according to claim 3, characterized in that: Processing of the Lagrange residual: ； in ,definition , ,in Operation, If is a vector, find the modulus of the vector. If is a matrix, find the maximum singular value of the matrix. is a constant, , , is the corresponding generator matrix, and finally we get: ； in , is an unknown uncertainty and ;Will Expanding it, we get: ； in is the variable parameter state matrix, is the unknown uncertainty after linearization.

5. A downhole tool accelerometer fault detection method according to claim 4, characterized in that: The method of obtaining the generalized linear variable parameter system based on the downhole tool linearization system is to process the output equation: ； in 、 and They are the corresponding dimension matrices respectively; the linearized system of downhole tools is expanded and defined , , , ,in , , , , , , This results in the following augmented system: ； in , , , , , , , ; The Euler method is used to discretize the augmented system and the corresponding generalized linear variable parameter system model is obtained: ； in , , is the sampling time, 、 、 、 They are The system state vector, system output vector, measurement noise vector, sensor fault vector at the moment, for The system state vector at time t, is the system matrix, is the uncertainty term after linearization, is the output matrix, is the noise distribution matrix, is the fault distribution matrix; is a variable parameter matrix, expressed in the following form: ； in is included The polyhedron, for The number of vertices, for The variable parameter matrix corresponds to the vertices, is the weighting function, ,satisfy: 。 6. A method for detecting faults in a downhole tool accelerometer according to claim 5, characterized in that The establishment method of the TNL observer is: For the obtained generalized linear system, a TNL observer is established; The residual system equations are divided into the residual equations of the fault subsystem and the residual equations of the interference subsystem; Finite frequency domain The sensitivity design of the fault subsystem is carried out by using the sensitivity index; The radius index is used to conduct robust design of the interference subsystem and establish robustness index; The multi-objective optimization problem of sensitivity index and robustness index is transformed into the solution of linear matrix inequality, and the observer parameter matrix is ​​calculated. 、 、 .

7. A method for detecting faults in a downhole tool accelerometer according to claim 6, characterized in that For the obtained generalized linear system, the TNL observer equation is established to obtain Observation value of the system state vector at time and The observed value of the output vector at time The calculation formula of state estimation error is defined as , residual ,in for The observed value of the system state vector at time t can be obtained based on the obtained generalized linear system and TNL observer equation The state estimation error at time and residuals The calculation formula is: divide the residual equation into the fault subsystem and the interference subsystem, and obtain the subsystem The error of time and The residual at time is 、 and 、 ; Using finite frequency domain index Conduct sensitivity design for fault subsystems, where The sensitivity of the reaction system to failure; using the set-based The radius index is used to design the robustness of the interference subsystem. The ellipsoidal beam of the interference subsystem error is obtained according to the interference subsystem: , and satisfy: ； in 、 They are Moment and time The center vector of 、 They are Moment and time The generator matrix of for The parameter matrix composed of the maximum values ​​of each element in; for The state vector at the moment is the center vector of the ellipsoidal beam, for The generator matrix of the ellipsoidal beam of the state vector at the moment; for The generation matrix of the ellipsoidal beam of the momentary noise variation; for The generation matrix of the noise ellipsoidal beam is measured at every moment; for The generation matrix of the ellipsoidal beam of the momentary fault variation; Finally, the multi-objective optimization problem of sensitivity and robustness indicators is transformed into the solution of linear matrix inequalities to obtain the corresponding observer parameter matrix.

8. A downhole tool accelerometer fault detection method according to claim 7, characterized in that The ellipsoidal beam of the interference residual subsystem is reduced in dimension by: dimensional ellipsoidal beam and integers , for -dimensional hypercube, and satisfies ,definition For the general The matrix obtained by arranging the column vectors of the matrix in descending order of the Euclidean norm is ,in for Before the Matrix List, For the remaining columns, is a diagonal matrix, and ,in express No. elements, express The absolute value of elements, , .

9. A method for detecting faults in a downhole tool accelerometer according to claim 8, characterized in that The evaluation mechanism for the occurrence of a fault is to first calculate , and satisfy: ； in Interference subsystem The moment residual ellipsoidal bundle, for The center vector of for The generation matrix of ; the fault occurrence mechanism judgment method is: ； in , for Moment residual No. A quantity, for The absolute value of for The absolute value of the maximum boundary, Calculated by the following formula: ； in for Moment Matrix No. OK, To calculate the vector The 1-norm of .

10. A detection device based on the downhole tool accelerometer fault detection method according to any one of claims 1 to 9, comprising a drill collar body, characterized in that: A first circuit board and a second circuit board are installed in the drill collar body. The first circuit board includes a first microprocessor and an auxiliary gyroscope. The second circuit board includes a communication interface and a second microprocessor. The accelerometer measurement value is transmitted to the first microcontroller and the second microcontroller via the communication interface. The first microcontroller of the first circuit board obtains observation values ​​of each state based on the auxiliary gyroscope and accelerometer data through a TNL observer and a linearized LPV system model. The observation values ​​obtained by the first microcontroller are transmitted to the second circuit board via SPI communication. The second microcontroller calculates the residual value of the current system state quantity based on the observation value calculated by the first circuit board and the measurement value obtained by the sensor to determine the occurrence of an event, and calculates the threshold value of the sensor failure under the current event to determine the occurrence of the fault; The fault monitoring device sends a fault alarm signal to the ground, and the fault alarm signal is a multiple of the residual information exceeding the threshold.

Images