Magnetic field and magnetic gradient detection method and device

Abstract

The invention belongs to the technical field of magnetic anomaly detection, and particularly relates to a magnetic field and magnetic gradient detection method and device.The magnetic field and magnetic gradient detection device comprises three three-component time difference type fluxgate sensors and a reference time difference type fluxgate sensor which are arranged in a spherical shell, and the current in a spherical feedback coil is adjusted; the values of the three components of the reference time difference type fluxgate sensor approach to zero; obtaining respective output signals of the three-component time difference type fluxgate sensor and the reference time difference type fluxgate sensor; calculating a time difference through the output signal; calculating a measured magnetic field according to the time difference; and constructing a magnetic gradient tensor according to magnetic induction intensity signals in different axial directions on the three-component time difference type fluxgate sensor and the reference time difference type fluxgate sensor. The problems of signal drift, weak anti-interference capability and the like of a traditional amplitude type fluxgate sensor in measurement are thoroughly solved, and the accuracy of magnetic field measurement is greatly improved.

Description

Technical Field

[0001] The present application belongs to the field of magnetic anomaly detection technology, and specifically relates to a magnetic field and magnetic gradient detection method and device. Background Art

[0002] Magnetic gradient tensor (MGT) sensors are a type of high-precision magnetic field measurement instrument used to capture directional variations in the magnetic field. Their operating principle is to simultaneously and precisely measure the rate of change of the magnetic field in three dimensions (typically the X, Y, and Z axes) using a combination of multiple fluxgates or magnetic sensors. This allows them to obtain a magnetic field gradient tensor, providing critical data for in-depth analysis of magnetic field characteristics. Given the high precision required for sensing magnetic field direction, MGT probes are currently primarily implemented using a superconducting quantum interference device (SQUID) combination. SQUID sensors are manufactured using standard Nb / AlOx / Nb technology and are waterproof and thermally resistant. These systems have undergone numerous experiments and flight tests, successfully measuring helicopter magnetic field characteristics and conducting flight tests of single gradiometers and multiple generations of systems. However, superconducting materials face numerous challenges in practical applications. Not only are they extremely sensitive to the external magnetic field environment, but even subtle magnetic field fluctuations can disrupt the stability of the superconducting system, affecting measurement accuracy. Therefore, in high-precision measurement scenarios, comprehensive and efficient magnetic field shielding is essential to isolate them from external interference. Moreover, superconducting materials exhibit superconducting properties only in an extremely low temperature environment, and usually require the use of cryogenic cooling media such as liquid nitrogen or liquid helium to maintain the low temperature state. This undoubtedly significantly increases the operation and maintenance costs of the superconducting system, increases the complexity of the system architecture, and also brings high energy consumption problems, limiting the expansion of its application scenarios. For existing spherical feedback three-component fluxgate magnetic gradient full tensor probes, although the measurement of the magnetic gradient tensor is achieved, due to the limitations of the inductive fluxgate, the inductive magnetic sensor can only measure the alternating magnetic field and cannot measure the DC magnetic field. The adaptive ability is not strong. In many practical application scenarios such as geological exploration, magnetic anomaly detection, and military reconnaissance, the lack of DC magnetic field information greatly limits the effectiveness of this type of probe, making it difficult to meet the needs of comprehensive and accurate magnetic field measurements in complex environments. Summary of the Invention

[0003] The technical problem to be solved by this application is to provide a magnetic field and magnetic gradient detection method to solve the problem that the inductive magnetic sensor can only measure the alternating magnetic field but cannot measure the DC magnetic field and has poor adaptability.

[0004] Another aspect of the present application provides a magnetic field and magnetic gradient detection device.

[0005] A magnetic field and magnetic gradient detection method according to an embodiment of the first aspect of the present application includes: Three three-component time-difference fluxgate sensors and one reference time-difference fluxgate sensor are placed in a spherical housing to form a cross structure centered at the coordinate origin. Two of the three-component time-difference fluxgate sensors are symmetrically distributed along the X-axis, and the other three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor are symmetrically distributed along the Y-axis. A spherical feedback coil is wound in three enameled wire winding grooves carved on the surface of the spherical housing and intersecting perpendicularly along the X, Y, and Z directions, respectively. The current in the spherical feedback coil is adjusted so that the values ​​of the three components of the reference time difference type fluxgate sensor approach zero; Passing excitation signals into the three three-component time-difference type fluxgate sensors and the reference time-difference type fluxgate sensor; Obtaining respective output signals of a three-component time-difference type fluxgate sensor and a reference time-difference type fluxgate sensor; Calculate the time difference through the output signal; Calculate the measured magnetic field based on the time difference; The magnetic induction signals of different three-component time-difference fluxgate sensors and reference time-difference fluxgate sensors are obtained according to the superposition of the measured magnetic field and the excitation magnetic field. The magnetic gradient tensor is constructed according to the magnetic induction intensity signals of different axes on the three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensors.

[0006] Furthermore, the excitation signal is a triangular wave, and the generated excitation magnetic field is expressed as: , In the formula is the slope of the triangular wave excitation magnetic field, For time, is the number of cycles, is the excitation cycle, For the excitation magnetic field.

[0007] Furthermore, the formula for calculating the measured magnetic field based on the time difference is: , where is the measured magnetic field, is the magnetic field frequency, For the time difference.

[0008] Furthermore, the calculation formula for the magnetic induction intensity signal of different three-component time-difference fluxgate sensors obtained by superposition of the measured magnetic field and the excitation magnetic field is: , B is the magnetic induction intensity signal, is the vacuum permeability, is the relative magnetic permeability of the core material, is the excitation magnetic field amplitude.

[0009] Furthermore, a magnetic gradient tensor is constructed based on the magnetic induction intensity signals of different axes on the three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor, including: According to the difference in magnetic field induction intensity of the two three-component time-difference type fluxgate sensors on the X axis in the X, Y and Z directions respectively, divided by the distance between the two three-component time-difference type fluxgate sensors in the X direction, the gradient of the magnetic field component in the X direction as it changes in the X direction, the gradient of the magnetic field component in the Y direction as it changes in the X direction, and the gradient of the magnetic field component in the Z direction as it changes in the X direction are obtained respectively; According to the difference in magnetic induction intensity between the three-component time-difference fluxgate sensor on the Y axis and the reference time-difference fluxgate sensor in the X, Y, and Z directions, divided by the distance between the three-component time-difference fluxgate sensor in the Y direction and the reference time-difference fluxgate sensor, the gradient of the magnetic field component in the X direction as it changes in the Y direction, the gradient of the magnetic field component in the Y direction as it changes in the Y direction, and the gradient of the magnetic field component in the Z direction as it changes in the Y direction are obtained respectively; The gradient of the magnetic field component in the Z direction is obtained by taking the negative value of the sum of the gradient of the magnetic field component in the X direction as it changes in the X direction and the gradient of the magnetic field component in the Y direction as it changes in the Y direction.

[0010] Furthermore, the current in the spherical feedback coil is adjusted so that the values ​​of the three components of the reference time difference type fluxgate sensor approach zero, including: Obtain output data of a reference time difference type fluxgate sensor; Construct a preliminary statistical characteristic model of the noise, set the objective function of minimizing the noise, and iteratively adjust the feedback parameters with a fixed step size; In each iteration, the degree of stochastic resonance achieved under the current feedback parameter settings is evaluated by comparing the output data with the expected output; If the ideal state is not reached, the feedback parameters are adjusted according to the error signal.

[0011] Furthermore, the output signals of the three three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensor are decomposed by wavelet transform, and the high-frequency noise components are filtered out by a threshold function while retaining the key features to obtain the denoising coefficient; Reconstruct the denoising coefficient to generate a low-noise signal and input it into an untrained neural network; Start the improved particle swarm algorithm, initialize the particle speed, position and dynamic parameters, and adjust the particle state in real time during iteration; When the global fitness value accuracy is met or the maximum number of iterations is reached, the optimal particle position sequence is output as the initial weight and threshold of the neural network; Drive the neural network training until the output error meets the set conditions; The output of the neural network is the denoised output signal.

[0012] Furthermore, after calculating the time difference through the output signal, the obtained time difference is denoised, including: Isolation forest anomaly detection is performed on the time difference. An anomaly detection model is constructed through dynamic window segmentation and adaptive parameter adjustment. For each detected anomaly point, the location and degree of anomaly are recorded. For each anomaly point, the surrounding normal points are selected as reference points and the weighted mean is used to replace the anomaly point to obtain the preprocessed signal. The preprocessed signal is input into the VMD decomposition module, and the mode number K and penalty factor are adaptively selected based on the characteristics of the preprocessed signal. , decompose the preprocessed signal into K modal components and calculate the energy entropy of each component; By normalizing the entropy value, weights are assigned according to the energy entropy of each modal component and weighted fusion is performed to generate an intermediate signal with enhanced main features; The intermediate signal enhanced by the main feature is subjected to hierarchical adaptive threshold calculation, soft threshold denoising is applied layer by layer, and the processing progress is iteratively judged. After completing all layers, the time difference is reconstructed.

[0013] A magnetic field and magnetic gradient detection device according to an embodiment of the second aspect of the present application includes: Three three-component time-difference fluxgate sensors and one reference time-difference fluxgate sensor are placed in a spherical housing, forming a cross structure centered on a coordinate origin. Two of the three-component time-difference fluxgate sensors are symmetrically distributed along the X-axis, and the other three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor are symmetrically distributed along the Y-axis. The surface of the spherical shell is engraved with three enameled wire winding grooves that intersect perpendicularly along the X, Y, and Z directions respectively, for winding the spherical feedback coil; a controller for adjusting the current in the spherical feedback coil so that the values ​​of the three components of the reference time difference type fluxgate sensor approach zero; A detection circuit is used to input excitation signals into the three three-component time-difference type fluxgate sensors and the reference time-difference type fluxgate sensor; Used to obtain the output signals of three three-component time difference type fluxgate sensors and a reference time difference type fluxgate sensor; Used to calculate the time difference through the output signal; Used to calculate the measured magnetic field based on the time difference; It is used to obtain the magnetic induction signals of different three-component time-difference fluxgate sensors and reference time-difference fluxgate sensors based on the superposition of the measured magnetic field and the excitation magnetic field, and to construct a magnetic gradient tensor based on the magnetic induction intensity signals of different axes on the three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensors.

[0014] Compared with the existing technology, this application has the following advantages: it avoids the shortcomings of amplitude-type fluxgate sensors and adopts a time-difference fluxgate sensor for detection, obtaining accurate time-difference measurements. This completely solves the problems of signal drift and weak anti-interference ability in traditional amplitude-type fluxgate sensors, greatly improving the accuracy of magnetic field measurements. It is suitable for applications requiring high magnetic field measurement accuracy and for measurements in complex electromagnetic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flow chart of a magnetic field and magnetic gradient detection method provided in an embodiment of the present application; Figure 2 A diagram showing the relative positions of the time-difference fluxgate sensor provided in an embodiment of the present application; Figure 3 A block diagram of the controller, detection circuit, and acquisition module provided in an embodiment of the present application; Figure 4 This is a circuit diagram of the acquisition module provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0017] See also Figure 1 A flow chart of a magnetic field and magnetic gradient detection method is shown. The present invention provides a magnetic field and magnetic gradient detection method, comprising: placing three three-component time-difference fluxgate sensors and one reference time-difference fluxgate sensor in a spherical housing to form a cross structure centered on a coordinate origin in space, wherein two of the three-component time-difference fluxgate sensors are symmetrically distributed along the X-axis, and the other three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor are symmetrically distributed along the Y-axis; a spherical feedback coil is wound in three enameled wire winding grooves carved on the surface of the spherical housing and intersecting perpendicularly along the X, Y, and Z directions, respectively; The current in the spherical feedback coil is adjusted so that the values ​​of the three components of the reference time difference type fluxgate sensor approach zero; Passing excitation signals into the three three-component time-difference type fluxgate sensors and the reference time-difference type fluxgate sensor; Obtaining respective output signals of three three-component time-difference type fluxgate sensors and a reference time-difference type fluxgate sensor; Calculate the time difference through the output signal; Calculate the measured magnetic field based on the time difference; The magnetic induction signals of different three-component time-difference fluxgate sensors and reference time-difference fluxgate sensors are obtained according to the superposition of the measured magnetic field and the excitation magnetic field. The magnetic gradient tensor is constructed according to the magnetic induction intensity signals of different axes on the three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensors.

[0018] The three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor here have the same structure. They measure the external magnetic field based on the principle of electromagnetic induction and the nonlinear characteristics of the soft magnetic core material. When the periodically alternating excitation magnetic field acts on the soft magnetic core to reach a bidirectional oversaturation state, when the external magnetic field is 0, the time difference between the positive and negative saturation states when the soft magnetic core is magnetized does not change. When the external magnetic field is not 0, the time difference between the positive and negative saturation states when the core is magnetized changes. The time difference is linearly related to the measured magnetic field, thereby obtaining a single-direction measured magnetic field.

[0019] In the embodiment of the present application, the current in the spherical feedback coil is adjusted according to the output signal of the reference time difference type fluxgate sensor, so that the values ​​of the three components of the reference time difference type fluxgate sensor approach zero, and the current on the spherical feedback coil is adjusted to generate a feedback magnetic field to achieve offset with the measured magnetic field. When the magnetic field measured by the reference time difference type fluxgate sensor approaches zero, the magnetic field environment in which the four time difference type fluxgate sensors operate is close to a zero magnetic field. The induced signal measured in this way is the residual magnetic field generated by the measured magnetic field and the spherical feedback coil. This residual magnetic field will not be zero, but approaches zero. The smaller the value of the residual magnetic field, the more accurate the measurement result.

[0020] The output signals of the magnetic field measured by the other three three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensor, namely the induction signals, are collected to calculate the time difference and the magnetic gradient tensor.

[0021] The three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor measure the output signal of the magnetic field. The methods for obtaining the time difference may include: the peak detection method, which locates the peak moment of the induction signal pulse through the peak detection circuit, and calculates the time difference based on the time of three adjacent peak points, thereby eliminating the need to consider the impact of the change in the amplitude of the induction signal on the reading of the pulse peak moment.

[0022] In one embodiment, the excitation signal is a triangular wave, and the generated excitation magnetic field is expressed as: , where is the slope of the triangular wave excitation magnetic field, For time, is the number of cycles, is the excitation cycle, is the excitation magnetic field. When the excitation period is The excitation magnetic field When used as an excitation signal, the detection principle based on the hysteresis saturation bistable characteristics of the magnetic core is used to detect the time difference of the two-way saturation and the measured magnetic field. The relationship between the time difference is finally used to obtain the measured magnetic field. In one embodiment, a triangular wave is selected as the excitation signal, which acts on the excitation coils of the three-component time difference type fluxgate sensor and the reference time difference type fluxgate sensor to measure the magnetic field. This excitation signal has low requirements on the coercive force of the magnetic core, and the sensitivity is only determined by the amplitude of the excitation magnetic field. and magnetic field frequency Decide.

[0023] In one embodiment, the formula for calculating the measured magnetic field based on the time difference is: , where is the measured magnetic field, is the magnetic field frequency, is the time difference; the calculation formula for the magnetic induction intensity signal of different three-component time difference type fluxgate sensors obtained by superposition of the measured magnetic field and the excitation magnetic field is: , B is the magnetic induction intensity signal, is the vacuum permeability, is the relative magnetic permeability of the core material, is the excitation magnetic field amplitude.

[0024] In one embodiment, constructing a magnetic gradient tensor based on magnetic induction intensity signals on different axes of a magnetic field measured by a three-component time-difference fluxgate sensor and a reference time-difference fluxgate sensor includes: According to the difference in magnetic induction intensity of the two three-component time-difference type fluxgate sensors in the X direction in the X, Y and Z directions respectively, divided by the distance between the two three-component time-difference type fluxgate sensors in the X direction, the gradient of the magnetic field component in the X direction as it changes in the X direction, the gradient of the magnetic field component in the Y direction as it changes in the X direction, and the gradient of the magnetic field component in the Z direction as it changes in the X direction are obtained; According to the difference in magnetic induction intensity in the X, Y, and Z directions between the three-component time-difference fluxgate sensor in the Y direction and the reference time-difference fluxgate sensor, divided by the distance between the three-component time-difference fluxgate sensor in the Y direction and the reference time-difference fluxgate sensor, the gradient of the magnetic field component in the X direction as it changes in the Y direction, the gradient of the magnetic field component in the Y direction as it changes in the Y direction, and the gradient of the magnetic field component in the Z direction as it changes in the Y direction are obtained; The gradient of the magnetic field component in the Z direction is obtained by taking the negative of the sum of the gradient of the magnetic field component in the X direction as it changes in the X direction and the gradient of the magnetic field component in the Y direction as it changes in the Y direction.

[0025] Two three-component time difference fluxgate sensors are symmetrically distributed along the X axis, and one three-component time difference fluxgate sensor and a reference time difference fluxgate sensor are symmetrically distributed along the Y axis. The full tensor probe consisting of three three-component time difference fluxgate sensors and one reference time difference fluxgate sensor is composed of a magnetic gradient tensor. The form can be expressed as: , In the formula , , represents the elements in the magnetic gradient tensor, which contains 9 components. However, in the passive space (referring to a specific area without the source of external magnetic field or electric field. In this space, all electromagnetic fields are generated by boundary conditions or distant sources, rather than directly by sources inside the area), in mathematical and physical descriptions, passive space usually means that the charge density and current density in the area are zero. This leads to the establishment of two important electromagnetic field equations: Divergence Theorem. In a passive space, the divergence of the electric field is zero, that is, ∇ B=0, which means that the electric field lines are closed, with no beginning and no end.

[0026] Ampere's circuit theorem. In a passive space, the curl of the magnetic field is zero, that is, ∇ B=0, which means that the magnetic field lines are also closed and there are no magnetic monopoles.

[0027] These conditions lead to the symmetry of the magnetic gradient tensor and the linear dependence of some of its components. Since the curl of the magnetic field is zero, the off-diagonal elements of the magnetic gradient tensor satisfy = , = , = , meaning there are only three independent off-diagonal elements; in addition, the curl of the magnetic field is zero. The trace of the magnetic gradient tensor (the sum of all diagonal elements) must also be zero, i.e. + + = 0, which means that only two of the three diagonal elements are independent; therefore, only five quantities are independent in the sourceless space.

[0028] The calculation elements are: (1) , It is obtained by taking the difference in magnetic induction intensity of two three-component time-difference fluxgate sensors symmetrically distributed on the X axis in the X direction and dividing it by the distance between them. It is calculated to represent the gradient of the magnetic field component in the X direction as it changes with the X direction.

[0029] is the component of the magnetic induction intensity in the X direction of one of the two three-component time-difference type fluxgate sensors that are symmetrically distributed along the X axis, The X-axis is symmetrically distributed between two three-component time-difference fluxgate sensors, and the other one is the component of the magnetic induction intensity in the X direction. Obtain.

[0030] (2) , It is obtained by taking the difference in magnetic induction intensity in the Y direction of two three-component time-difference fluxgate sensors symmetrically distributed on the X axis and dividing it by the distance between them. To calculate, the gradient of the magnetic field component in the Y direction changes with the X direction, and the magnetic induction intensity is calculated by the time difference Obtain.

[0031] is the component of the magnetic induction intensity in the Y direction of one of the two three-component time-difference type fluxgate sensors that are symmetrically distributed along the X axis, It is the component of the magnetic induction intensity in the Y direction of the other three-component time-difference type fluxgate sensor of the two three-component time-difference type fluxgate sensors that are symmetrically distributed about the X axis.

[0032] (3) , It is obtained by taking the difference in magnetic induction intensity in the Z direction of two three-component time-difference fluxgate sensors symmetrically distributed on the X axis and dividing it by the distance between them. To calculate, the magnetic field component in the Z direction changes with the gradient of the X direction, and the magnetic induction intensity is calculated by the time difference Obtain.

[0033] is the component of the magnetic induction intensity in the Z direction of one of the two three-component time-difference type fluxgate sensors that are symmetrically distributed along the X axis, It is the component of the magnetic induction intensity in the Z direction of the other three-component time-difference type fluxgate sensor of the two three-component time-difference type fluxgate sensors that are symmetrically distributed about the X axis.

[0034] (4) , It is the difference in magnetic induction intensity between the reference time difference type fluxgate sensor on the Y axis and the three-component time difference type fluxgate sensor in the X direction, divided by the distance between them , which represents the gradient of the magnetic field component in the X direction as it changes with the Y direction.

[0035] is the component of the magnetic induction intensity of the reference time difference type fluxgate sensor on the Y axis in the X direction, is the component of the magnetic induction intensity of the three-component time difference type fluxgate sensor on the Y axis in the X direction.

[0036] (5) , It is the difference in magnetic induction intensity between the reference time difference type fluxgate sensor and the three-component time difference type fluxgate sensor on the Y axis, divided by the distance between them It is calculated to represent the gradient of the magnetic field component in the Y direction as it changes in the Y direction.

[0037] is the component of the magnetic induction intensity of the reference time difference type fluxgate sensor on the Y axis in the Y direction, is the component of the magnetic induction intensity of the three-component time difference type fluxgate sensor on the Y axis in the Y direction.

[0038] (6) , It is the difference in magnetic induction intensity between the reference time difference fluxgate sensor on the Y axis and the three-component time difference fluxgate sensor in the Z direction, divided by the distance between them It is calculated to represent the gradient of the magnetic field component in the Z direction as it changes with the Y direction.

[0039] is the component of the magnetic induction intensity of the reference time difference type fluxgate sensor on the Y axis in the Z direction, is the component of the magnetic induction intensity of the three-component time difference type fluxgate sensor on the Y axis in the Z direction.

[0040] (7) , It is obtained by taking the difference in magnetic induction intensity in the Z direction of two three-component time-difference fluxgate sensors symmetrically distributed on the X axis and dividing it by the distance between them. It is calculated to represent the gradient of the magnetic field component in the Z direction as it changes with the X direction.

[0041] is the component of the magnetic induction intensity in the Z direction of one of the two three-component time-difference type fluxgate sensors that are symmetrically distributed along the X axis, It is the component of the magnetic induction intensity in the Z direction of the other three-component time-difference type fluxgate sensor of the two three-component time-difference type fluxgate sensors that are symmetrically distributed about the X axis.

[0042] (8) , It is the difference in magnetic induction intensity between the reference time difference fluxgate sensor on the Y axis and the three-component time difference fluxgate sensor in the Z direction, divided by the distance between them To calculate, it represents the gradient of the magnetic field component in the Z direction as it changes with the Y direction, but due to the reference time difference type fluxgate sensor output is zero.

[0043] is the component of the magnetic induction intensity of the reference time difference type fluxgate sensor on the Y axis in the Z direction, is the component of the magnetic induction intensity of the three-component time difference type fluxgate sensor on the Y axis in the Z direction.

[0044] (9) , The essence is to calculate the gradient of the magnetic field in the Z direction, which is affected by the changes in the magnetic field in the X and Y directions. It is the result of the combined effect of the magnetic field changes in the X and Y directions. In addition, since the curl of the magnetic field is zero, the trace of the magnetic gradient tensor (the sum of all diagonal elements) must also be zero, that is, ,so A minus sign is required.

[0045] In one embodiment, adjusting the current in the spherical feedback coil so that the values ​​of the three components of the reference time difference type fluxgate sensor are zero includes: Obtain output data of a reference time difference type fluxgate sensor; Construct a preliminary statistical characteristic model of the noise, set the objective function of minimizing the noise, and iteratively adjust the feedback parameters with a fixed step size; In each iteration, the degree of stochastic resonance achieved under the current feedback parameter settings is evaluated by comparing the output data with the expected output; If the ideal state is not reached, the feedback parameters are adjusted according to the error signal.

[0046] The measured magnetic field is measured based on the output signal of the reference time-difference fluxgate sensor. The current in the spherical feedback coil is adjusted based on the output signal to generate a canceling magnetic field. The ultimate goal is to drive the output signal of the reference time-difference fluxgate sensor close to zero. It is understandable that the measured values ​​of the remaining three three-component time-difference fluxgate sensors in the same environment as the reference time-difference fluxgate sensor also approach zero. However, experiments have shown that these values ​​cannot be zero, as a residual magnetic field is present.

[0047] The statistical features used in constructing a preliminary statistical feature model of noise may include: Probability distribution: Noise signals usually have certain statistical characteristics, and the changes in their amplitude and phase can be described by probability distribution. Common probability distributions include Gaussian distribution, uniform distribution, Poisson distribution, etc. Mean: The mean of a noise signal reflects the average level of the noise signal. Under normal circumstances, the mean of a noise signal is close to zero. Variance and standard deviation: Variance and standard deviation are used to measure the fluctuation range of the noise signal amplitude and are important parameters for measuring noise intensity. The standard deviation is the positive square root of the variance. Correlation function and power spectral density: The correlation function of noise describes the degree of correlation of noise signals at different time points, while the power spectral density describes the power distribution of the noise signal in the frequency domain. Select an appropriate preliminary statistical feature model, such as a Gaussian model, a Poisson model, etc.

[0048] The objective function used can be mean squared error or absolute error, and the feedback parameters are iteratively adjusted to gradually optimize the value of the objective function. The basic idea of ​​iterative adjustment is that in each iteration, the gradient (or derivative) of the objective function is calculated based on the current parameter value, and then the feedback parameters are updated in the opposite direction of the gradient to gradually reduce the value of the objective function.

[0049] The fixed step size (also known as the learning rate) is an important parameter in the iterative adjustment process. It determines the magnitude of the feedback parameter update at each iteration.

[0050] In another embodiment, the following scheme may be used to adjust the current in the spherical feedback coil so that the values ​​of the three components of the reference time difference type fluxgate sensor are zero: Select a feedback structure (the odd-order, cubic, and higher-order feedback structures described below), set the feedback parameters, input the excitation signal, and output it through Runge-Kutta iteration. A particle swarm algorithm is then introduced to dynamically solve the coefficients of the odd-order polynomial. Meanwhile, a cross-correlation analysis is performed on the excitation signal. By analyzing multiple sets of data, the closer the correlation between the excitation signal and the output signal approaches 1, the more likely the feedback system is resonating. The data at this point is recorded and stored, and the optimal feedback parameters are found and reset to resonate the feedback system, including the reference time-difference fluxgate sensor. This process concludes, achieving the goal of actively adapting the three three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensor to external noise. The above process automatically adjusts the feedback parameters through software calculations.

[0051] This application studies the conditions under which a bistable system generates stochastic resonance. Based on the equivalent mathematical model of the bistable system, an adaptive control strategy is designed to achieve stochastic resonance. The bistable characteristics of the time-difference fluxgate sensor can be described by the Langevin equation: , where is the input signal, and , is the input signal, is a Gaussian white noise signal. , , is a positive number. After taking its derivative, the Langevin equation shows an obvious odd characteristic. Therefore, the odd function characteristics of the bistable system are used. Considering that the selection of the feedback form can directly affect the sensitivity of the time-difference type fluxgate sensor, the odd, cubic, and high-order odd feedback structures are selected for comparison. Experiments have shown that under different noise intensities, the shape of the noise response characteristic curve is basically the same. When a high-order odd feedback structure is used, the time-difference type fluxgate sensor can achieve maximum sensitivity within a certain noise range; cubic feedback can adjust the sensitivity peak position so that it actively adapts to external noise. Therefore, the present application combines linear and high-order odd feedback structures, which can effectively adjust the coercive force and sensitivity peak position, and can greatly improve the sensitivity of the time-difference type fluxgate sensor and reduce the impact of noise when the noise characteristics are uncertain.

[0052] Considering the unknown characteristic values ​​of environmental noise, it is difficult to change system parameters (coefficients of odd-order functions) one by one to achieve random resonance. Therefore, it is necessary to design a feedback parameter adaptive adjustment algorithm to dynamically solve the coefficients of odd-order polynomials. In combination with the need for real-time magnetic field measurement by time-difference fluxgate sensors, a particle swarm algorithm is introduced to dynamically solve the coefficients of odd-order polynomials. At this time, the particles represent the feedback coefficient values ​​to be solved. By initializing the feedback coefficient set, the effect of using the optimal coefficient feedback on improving sensor sensitivity is judged based on the fitness function value. The fitness function is used to evaluate a potential feasible solution, that is, the optimal value of the feedback coefficient. Therefore, this application uses the mean square error as the fitness function, which can measure the gap between the model prediction value and the actual value. When the fitness value reaches a certain value, the time difference is maximized, and the time-difference fluxgate sensor reaches the optimal resonance state. At this time, the output value of the particle swarm algorithm is the optimal coefficient combination of the feedback function. By calculating the fitness value of the new position of each feedback coefficient particle, updating the individual and global optimal values ​​according to the fitness value of each feedback coefficient, updating the speed and position of each feedback coefficient particle, checking whether the termination condition of the optimal fitness is met, if not, returning to the step loop of calculating the fitness value of the new position of each feedback coefficient particle, and outputting the optimal solution of the feedback coefficient if it is met, the sensitivity of the time difference type fluxgate sensor is indirectly improved by reaching a resonant state.

[0053] In one embodiment, the output signals of the three three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensor are decomposed by wavelet transform, and high-frequency noise components are filtered out by a threshold function while retaining key features to obtain a denoising coefficient; Reconstruct the denoising coefficient to generate a low-noise signal and input it into an untrained neural network; Start the improved particle swarm algorithm, initialize the particle speed, position and dynamic parameters, and adjust the particle state in real time during iteration; When the global fitness value accuracy is met or the maximum number of iterations is reached, the optimal particle position sequence is output as the initial weight and threshold of the neural network; Drive the neural network training until the output error meets the set conditions; The output of the neural network is the denoised output signal.

[0054] Among them, the denoising coefficient is reconstructed to generate a low-noise signal and input into the untrained neural network by using inverse wavelet transform for reconstruction. The inertia weight in the traditional particle swarm algorithm is Usually it is a fixed value or a simple linear decrease, which cannot be dynamically adapted according to the search process. The embodiment of the present application adopts a linear decrease strategy: , in, and are the initial and final values ​​of the inertia weight respectively; is the current iteration number; is the maximum number of iterations. Set High weight, this is the early weight, the value is set larger, enhance the global exploration, set It is a low weight, which is the later weight. The value is set to a smaller value, focusing on local development.

[0055] The speed update of the traditional particle swarm algorithm has no compression factor constraint, which may lead to excessive speed, particle oscillation or divergence. The improved speed update formula is as follows: , in, It is a particle In dimension The speed of the t+1th iteration on (the speed to be updated), It is a particle In dimension The speed of the t-th iteration on It is a particle In dimension The best historical position of the t-th iteration of the particle swarm is in dimension d. is the global optimal position of the tth iteration, is the compression factor, which is usually calculated based on the learning factor and The sum and the dimension of the particle swarm , dimension It may not appear directly in the calculation formula of the compression factor, but it will affect the performance of the algorithm and the choice of parameters.

[0056] The difference between traditional and improved speed updates is: New compression factor Limit the velocity amplitude to prevent particles from getting out of control.

[0057] Among them, the compression factor The calculation formula is: When the compression factor Force compression speed to ensure convergence stability: , in, yes and A function of the sum and their product, generally set >4 to ensure that the square root in the formula is meaningful and the factors are compressed Is a positive value.

[0058] In one embodiment, after calculating the time difference from the output signal, denoising the obtained time difference includes: Isolation forest anomaly detection is performed on the time difference. An anomaly detection model is constructed through dynamic window segmentation and adaptive parameter adjustment. For each detected anomaly point, the location and degree of anomaly are recorded. For each anomaly point, the surrounding normal points are selected as reference points and the weighted mean is used to replace the anomaly point to obtain the preprocessed signal. The preprocessed signal is input into the VMD decomposition module (Variational Mode Decomposition), and the mode number K and penalty factor are adaptively selected based on the characteristics of the preprocessed signal. , decompose the preprocessed signal into K modal components and calculate the energy entropy of each component; By normalizing the entropy value, weights are assigned according to the energy entropy of each modal component and weighted fusion is performed to generate an intermediate signal with enhanced main features; The intermediate signal enhanced by the main feature is subjected to hierarchical adaptive threshold calculation, soft threshold denoising is applied layer by layer, and the processing progress is iteratively judged. After completing all layers, the time difference is reconstructed.

[0059] Since the high-frequency random noise (frequency band 10kHz~100kHz) generated by irreversible domain wall jumping during the magnetization process of the soft magnetic core is distributed near the peak of the induced voltage pulse, it easily causes time difference fluctuations. Therefore, the time difference signal is denoised based on the isolation forest anomaly detection and VMD decomposition module, which effectively suppresses signal distortion and enhances denoising accuracy.

[0060] The modal number K of traditional VMD decomposition needs to be preset manually or calculated by fixed empirical formula, which is susceptible to noise interference and leads to insufficient or excessive decomposition. The combined rule with energy entropy avoids over-decomposition caused by high-frequency noise interference. The modal number K is as follows: , in is the signal length to avoid over-decomposition under high-frequency noise interference.

[0061] Traditional VMD decomposition cannot adapt to different noise scenarios and may cause signal details to be lost or noise to remain. Adaptive adjustment penalty factor : When there is high noise Increase, approach 90, strengthen bandwidth constraint to suppress noise; penalty factor for low noise Decrease, approaching 80, to preserve signal details. The formula is as follows: , Signal variance in high noise scenarios Increase, penalty factor Approaching 90, strengthening the modal bandwidth constraint. In low noise scenarios Approaching 80, it preserves signal details while maintaining modal orthogonality.

[0062] Traditional VMD decomposition uses the alternating direction multiplier method (ADMM), but does not dynamically adjust iterative parameters based on signal characteristics, resulting in poor convergence and adaptability. The present embodiment improves VMD decomposition by iteratively updating the alternating direction multiplier method, including modal updates, center frequency updates, and Lagrange multiplier updates. Here, Modal Update: Update in the frequency domain : , It means the At the first iteration, Modes in the frequency domain The updated value at For the Modes in the frequency domain The value at Represents the frequency domain representation of the original signal (input signal). Indicates that except modes, other modes in the frequency domain The weight and It represents the Lagrange multiplier, which is used to constrain the decomposition conditions. Representative The center frequency of a mode.

[0063] Center frequency update: , | | indicates the At the first iteration, The energy density (square of the modulus) in the frequency domain after the mode is updated is calculated by weighted integration (with The ratio of the modal energy to the energy integral is used to calculate the center frequency, so that the modal energy is concentrated near the center frequency. Indicates the Iterations, updated center frequency.

[0064] Lagrange multiplier update: ,in, It means the The updated value of the Lagrange multiplier at iteration . Indicates the The Lagrange multiplier of the iteration. is the update step size (learning rate), which controls the amplitude of the Lagrange multiplier update.

[0065] After decomposition, weights are assigned according to modal energy entropy, and wavelet denoising is integrated: high energy entropy modes (containing more noise) are assigned low weights, and low energy entropy modes (containing more effective signals) are assigned high weights.

[0066] , , In the formula It means the modal signals in the time domain energy. Indicates the The entropy of the modal signal in each time domain. The high entropy value mode (including complex noise) is given a low weight to suppress the noise reconstruction contribution. The low entropy value mode (including effective signal) is given a high weight. In terms of multi-scale threshold denoising, the modal signal state of each time domain is Perform adaptive wavelet threshold processing: use a low threshold for high-frequency modes (to preserve details) and a high threshold for low-frequency modes (to suppress baseline drift). represents the weight, Indicates the The entropy of the modal signal in the time domain.

[0067] In one embodiment, the wavelet transform is improved by applying an adaptive soft thresholding process to the wavelet coefficients at each scale. The improved method is more adaptable than the traditional soft thresholding process and can effectively remove noise while preserving the signal characteristics as much as possible. The following is a public statement: The adaptive soft thresholding process is to apply a soft thresholding process to the wavelet coefficients at each scale: , Represents the updated wavelet coefficients after threshold processing. Indicates the threshold value corresponding to scale j. Different scales j use different , is the wavelet coefficient, the threshold value increases for high noise scale (to suppress noise), and decreases for low noise scale (to preserve details).

[0068] See also Figure 2 The relative position diagram of the time difference type fluxgate sensor shown in the figure and the reference Figure 3 The embodiment of the present application shown in the figure provides a structural block diagram of a magnetic field and magnetic gradient detection device, which includes: Three three-component time-difference fluxgate sensors and one reference time-difference fluxgate sensor are placed in a spherical housing, forming a cross structure centered on a coordinate origin. Two of the three-component time-difference fluxgate sensors are symmetrically distributed along the X-axis, and the other three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor are symmetrically distributed along the Y-axis. The surface of the spherical shell is engraved with three enameled wire winding grooves that intersect perpendicularly along the X, Y, and Z directions respectively, which are used to wind the spherical feedback coil; a controller for adjusting the current in the spherical feedback coil so that the values ​​of the three components of the reference time difference type fluxgate sensor approach zero; A detection circuit is used to input excitation signals into the three three-component time-difference type fluxgate sensors and the reference time-difference type fluxgate sensor; Used to obtain the output signals of three three-component time difference type fluxgate sensors and a reference time difference type fluxgate sensor; Used to calculate the time difference through the output signal; Used to calculate the measured magnetic field based on the time difference; It is used to obtain the magnetic induction signals of different three-component time-difference fluxgate sensors and reference time-difference fluxgate sensors based on the superposition of the measured magnetic field and the excitation magnetic field, and to construct a magnetic gradient tensor based on the magnetic induction intensity signals of different axes on the three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensors.

[0069] The spherical housing is constructed of a non-conductive material, which can range from polycarbonate plastic, fiberglass, polyurethane foam, phenolic plastic, or mica. The enameled wire winding grooves are of equal width and spacing, but not equal depth. Inside the spherical housing, a three-component time-difference fluxgate sensor and a reference time-difference fluxgate sensor are secured with bolts using a cross-shaped bracket.

[0070] The three-component time difference type fluxgate sensor has the same structure as the reference time difference type fluxgate sensor, which includes a sensitive unit and a detection circuit. The sensitive unit consists of a soft magnetic core and a coil to sense the measured magnetic field. The three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor are based on the principle of electromagnetic induction and the nonlinear characteristics of the soft magnetic core material to achieve the measurement of the external magnetic field. (Periodic change diagram) When the soft magnetic core reaches a bidirectional supersaturation state through the periodic alternating excitation magnetic field, when the external magnetic field is 0, the time difference between the positive and negative saturation states when the core is magnetized does not change. When the external magnetic field is not 0, the time difference between the positive and negative saturation states when the core is magnetized changes. This time difference is linearly related to the measured magnetic field, thereby obtaining a single-direction measured magnetic field. The detection circuit includes an excitation signal generation module, an induction signal processing module, and a time difference reading and processing module. The excitation signal generation module is connected to the soft magnetic core and is used to generate a triangular wave. The induction signal processing module is used to pre-process the induction signal and denoise it. The denoising process is to perform wavelet transform decomposition, filter out high-frequency noise components through a threshold function, and retain key features to obtain a denoising coefficient. Reconstruct the denoising coefficient to generate a low-noise signal and input it into an untrained neural network; Start the improved particle swarm algorithm, initialize the particle speed, position and dynamic parameters, and adjust the particle state in real time during iteration; When the global fitness value accuracy is met or the maximum number of iterations is reached, the optimal particle position sequence is output as the initial weight and threshold of the neural network; Drive the neural network training until the output error meets the set conditions; The output of the neural network is the denoised output signal.

[0071] The time difference of the denoised output signal is calculated and read by the time difference reading and processing module. The measured magnetic field is calculated according to the time difference, and the magnetic induction signals of different three-component time difference type fluxgate sensors and the reference time difference type fluxgate sensor are obtained according to the superposition of the measured magnetic field and the excitation magnetic field. The magnetic gradient tensor is constructed according to the magnetic induction intensity signals of different axes on the three-component time difference type fluxgate sensor and the reference time difference type fluxgate sensor.

[0072] The time difference reading and processing module is further used to denoise the time difference obtained after calculating the time difference through the output signal, including: Isolation forest anomaly detection is performed on the time difference. An anomaly detection model is constructed through dynamic window segmentation and adaptive parameter adjustment. For each detected anomaly point, the location and degree of anomaly are recorded. For each anomaly point, the surrounding normal points are selected as reference points and the weighted mean is used to replace the anomaly point to obtain the preprocessed signal. The preprocessed signal is input into the VMD decomposition module, and the mode number K and penalty factor are adaptively selected based on the characteristics of the preprocessed signal. , decompose the preprocessed signal into K modal components and calculate the energy entropy of each component; By normalizing the entropy value, weights are assigned according to the energy entropy of each modal component and weighted fusion is performed to generate an intermediate signal with enhanced main features; The intermediate signal enhanced by the main feature is subjected to hierarchical adaptive threshold calculation, soft threshold denoising is applied layer by layer, and the processing progress is iteratively judged. After completing all layers, the time difference is reconstructed.

[0073] See also Figure 4 As shown, in one embodiment, the controller further includes an acquisition module connected to the reference time-difference fluxgate sensor. The acquisition module is connected to the input of the DSP module via an operational amplifier U1 in series with a resistor R1. After processing by the DSP module, the DAC module (digital-to-analog conversion module) converts the analog signal into an analog signal. Finally, the analog signal outputs current to the spherical feedback coil via the current output module. Through feedback regulation, the values ​​of the three components of the reference time-difference fluxgate sensor approach zero. The processing process of the digital signal processing module includes: Obtain output data of a reference time difference type fluxgate sensor; Construct a preliminary statistical characteristic model of the noise, set the objective function of minimizing the noise, and iteratively adjust the feedback parameters with a fixed step size; In each iteration, the degree of stochastic resonance achieved under the current feedback parameter settings is evaluated by comparing the output data with the expected output; If the ideal state is not reached, the feedback parameters are adjusted according to the error signal.

[0074] The acquisition module can also include a voltage follower to isolate the reference time-difference fluxgate sensor from subsequent circuitry and improve its output drive capability. The DSP module, after running a digital integration algorithm, filtering, and denoising, outputs the processed digital signal to the DAC module. However, the output digital signal has low power and cannot directly drive the spherical feedback coil. Therefore, the output of the DAC module is connected to the current output module, which uses an operational amplifier U2. Because the current required by the spherical feedback coil can reach over 30mA, and ordinary operational amplifiers cannot provide sufficient output current, a push-pull buffer Q consisting of two PNP transistors is added after the current output module to enhance the output's forward and reverse current capabilities. The output terminals of the two PNP transistors are connected to the inverting input of the current output module. When using a power amplifier to drive the spherical feedback coil, excessive current may occur. Connecting resistor R2 to the output terminals of the two PNP transistors limits the current, protecting the current output module, the spherical feedback coil, and other circuit components from overload damage. The current is then output to the spherical feedback coil.

[0075] When a magnetic field and magnetic gradient detection device of an embodiment of the present application performs magnetic gradient tensor detection, in order to achieve effective compensation for the ambient magnetic field, the maximum magnetic field intensity to be compensated in a certain direction is set to 50000nT as a design example, the radius R of the spherical shell is set to 16cm, and the output current I of the current output module is required to be greater than 30mA. When the number of slots wound around the spherical feedback coil is 12, the slot spacing is set to 0.8cm, the slot width is 0.2cm, and the distance from the enameled wire winding slot to the center of the spherical shell is 16cm. At the same time, enameled wire with a diameter of 0.1mm is used, and the number of winding turns per slot is 25.

[0076] according to , where , B is the magnetic induction intensity, is the dielectric constant, Feedback coil current, is the radius of the feedback coil. When the distance between the position to be calculated and the spherical feedback coil is =0, the number of turns is derived ,at this time Turns / m, then calculate the formula based on the number of turns , M is the number of circles, L is the length, and N can be obtained lock up.

[0077] By calculating the maximum magnetic field generated by the spherical feedback coil: B= , we know that B 50000nT.

[0078] In the embodiments of this application, magnetic gradient tensor measurement primarily relies on the difference between two three-component time-difference fluxgate sensors. Therefore, errors caused by magnetic field inhomogeneity have a negligible impact on tensor output accuracy. Based on this characteristic, the baseline distance between the two three-component time-difference fluxgate sensors is set to 12 cm. The same design approach and method are used for the other two directions. Ultimately, the three time-difference fluxgate sensors and a reference time-difference fluxgate sensor are arranged in a cross-shaped configuration with a spacing of 12 cm.

[0079] The embodiments of the present application can effectively measure the magnetic gradient tensor. The time-difference fluxgate sensor operates stably near zero magnetic field, reducing nonlinear errors and improving measurement accuracy. It should be noted that this does not mean that the magnetic field is completely offset to absolute zero. In practical applications, a certain residual magnetic field will always exist, which will cause the time-difference fluxgate sensor to produce a corresponding output. This not only ensures high stability and accuracy, but also avoids the effects of crosstalk.

[0080] Compared with traditional fluxgate sensors, the embodiments of the present application use a time-difference fluxgate sensor based on a new time-difference measurement principle, which completely solves the problems of signal drift and weak anti-interference ability in the measurement of traditional amplitude-type fluxgate sensors, greatly improving the accuracy of magnetic field measurement, and showing great advantages in high-end detection fields with extremely high requirements for magnetic field accuracy. In addition, with the help of the magnetic field feedback link, the spherical feedback coil shows extremely high performance advantages. Compared with other types of feedback coils, the spherical feedback coil can generate a more uniform and wider magnetic field. While maintaining the same magnetic field uniformity, the spherical feedback coil is smaller in size and has higher space utilization. By integrating four time-difference fluxgate sensors into the same three-axis spherical feedback coil, a stable zero-magnetic field working environment is effectively created, fundamentally avoiding the mutual interference caused by the presence of the feedback coil when the sensors are fed back individually, greatly simplifying the system calibration process, and significantly shortening the baseline distance.

[0081] Compared with the prior art, the embodiments of the present application avoid the defects of amplitude-type fluxgate sensors in terms of detection principle, adopt time-difference fluxgate sensors for detection, and obtain accurate time-difference measurement values ​​through method optimization, thereby providing higher measurement accuracy. High-resolution operation parameters are set according to requirements, effectively reducing errors and processing fluxgate sensor signals more accurately. It is particularly suitable for occasions with high requirements for magnetic field measurement accuracy. For measurements in complex electromagnetic environments, feedback parameters can be adaptively adjusted to ensure that they can actively adapt to the external environment in different environments, thereby improving adaptability and robustness. In addition, a complete signal denoising method for three-component time-difference fluxgate sensors is implemented by wavelet transform decomposition and particle swarm algorithm collaborative optimization of neural networks. Wavelet domain feature extraction, dynamic parameter particle swarm optimization and neural network parameter initialization are deeply integrated: on the one hand, signal complexity is reduced by wavelet transform, and on the other hand, the compression factor constraint and weight reduction strategy of the improved particle swarm are used to achieve automatic and accurate adaptation of network parameters. The high-frequency component changes corresponding to the time difference of the three axes on which the three-component time-difference fluxgate sensor measures the magnetic gradient tensor can be accurately captured, which can significantly improve the denoising effect and make the measured magnetic gradient tensor value more accurate. The high permeability and low coercivity of the soft magnetic core of the time-difference fluxgate sensor result in a "narrow pulse and high amplitude" characteristic for the induced signal. The wavelet transform of this application matches the nonlinear characteristics of the core magnetization process, preserving not only the peak pulses during the saturation phase but also utilizing its frequency-domain localization characteristics to accurately filter out the Barkhausen noise during the magnetization of the soft magnetic core of the time-difference fluxgate sensor. Compared to traditional single-method denoising, this method enhances the fitting capability of the induced signal, effectively suppresses signal distortion, and improves denoising accuracy. In response to sudden changes in the full tensor magnetic gradient caused by time-difference anomalies in magnetic gradient measurements due to external boundaries, a dynamic window-based isolated forest anomaly detection method is designed to address dynamic parameter coupling, adaptive adjustment of VMD modal parameters and wavelet thresholds, and cross-domain collaborative noise reduction (time-domain anomaly processing → frequency-domain modal separation → multi-scale wavelet refinement). Considering the complexity of the measured magnetic field and the reliance of the full magnetic gradient tensor measurement on the multi-scale characteristics of time differences, an entropy-weighted adaptive VMD decomposition module was constructed to address the entropy-weighted decision fusion mechanism (quantifying the value of modal information using energy entropy to avoid feature dilution caused by equal weighting). Furthermore, the uniform magnetic field generated by the spherical feedback coil is stronger than that of other types of feedback coils, and under the same uniformity conditions, the spherical feedback coil occupies the smallest space. By arranging multiple time-difference fluxgate sensors within a three-axis spherical feedback coil, these sensors can operate stably in a near-zero magnetic field environment, avoiding potential mutual interference between the individual time-difference fluxgate sensors when providing independent feedback.This application not only facilitates system calibration, but also significantly shortens the distance between baselines, thereby improving measurement accuracy. It is suitable for scenarios with strict requirements on magnetic field precision, such as scientific research experiments, high-end industrial testing, and geological exploration.

[0082] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A magnetic field and magnetic gradient detection method, characterized in that: include: Three three-component time-difference fluxgate sensors and one reference time-difference fluxgate sensor are placed in a spherical housing to form a cross structure centered at the coordinate origin. Two of the three-component time-difference fluxgate sensors are symmetrically distributed along the X-axis, and the other three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor are symmetrically distributed along the Y-axis. A spherical feedback coil is wound in three enameled wire winding grooves carved on the surface of the spherical housing and intersecting perpendicularly along the X, Y, and Z directions, respectively. The current in the spherical feedback coil is adjusted so that the values ​​of the three components of the reference time difference type fluxgate sensor approach zero; Passing excitation signals into the three three-component time-difference type fluxgate sensors and the reference time-difference type fluxgate sensor; Obtaining respective output signals of a three-component time-difference type fluxgate sensor and a reference time-difference type fluxgate sensor; Calculate the time difference through the output signal; Calculate the measured magnetic field based on the time difference; The magnetic induction signals of different three-component time-difference fluxgate sensors and reference time-difference fluxgate sensors are obtained according to the superposition of the measured magnetic field and the excitation magnetic field. The magnetic gradient tensor is constructed according to the magnetic induction intensity signals of different axes on the three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensors.

2. The magnetic field and magnetic gradient detection method according to claim 1, characterized in that: The excitation signal is a triangular wave, and the generated excitation magnetic field is expressed as: , where is the slope of the triangular wave excitation magnetic field, For time, is the number of cycles, is the excitation cycle, For the excitation magnetic field.

3. The magnetic field and magnetic gradient detection method according to claim 1, characterized in that: The calculation formula for calculating the measured magnetic field based on the time difference is: , where is the measured magnetic field, is the magnetic field frequency, For the time difference.

4. The magnetic field and magnetic gradient detection method according to claim 1, characterized in that: The calculation formula for the magnetic induction intensity signal of different three-component time difference type fluxgate sensors obtained by superposition of the measured magnetic field and the excitation magnetic field is: , B is the magnetic induction intensity signal, is the vacuum permeability, is the relative magnetic permeability of the core material, is the excitation magnetic field amplitude.

5. The magnetic field and magnetic gradient detection method according to claim 3, characterized in that: The magnetic gradient tensor is constructed based on the magnetic induction intensity signals of different axes on the three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor, including: According to the difference in magnetic field induction intensity of the two three-component time-difference type fluxgate sensors on the X axis in the X, Y and Z directions respectively, divided by the distance between the two three-component time-difference type fluxgate sensors in the X direction, the gradient of the magnetic field component in the X direction as it changes in the X direction, the gradient of the magnetic field component in the Y direction as it changes in the X direction, and the gradient of the magnetic field component in the Z direction as it changes in the X direction are obtained respectively; According to the difference in magnetic induction intensity between the three-component time-difference fluxgate sensor on the Y axis and the reference time-difference fluxgate sensor in the X, Y, and Z directions, divided by the distance between the three-component time-difference fluxgate sensor in the Y direction and the reference time-difference fluxgate sensor, the gradient of the magnetic field component in the X direction as it changes in the Y direction, the gradient of the magnetic field component in the Y direction as it changes in the Y direction, and the gradient of the magnetic field component in the Z direction as it changes in the Y direction are obtained respectively; The gradient of the magnetic field component in the Z direction is obtained by taking the negative value of the sum of the gradient of the magnetic field component in the X direction as it changes in the X direction and the gradient of the magnetic field component in the Y direction as it changes in the Y direction.

6. The magnetic field and magnetic gradient detection method according to claim 1, characterized in that: The current in the spherical feedback coil is adjusted to make the values ​​of the three components of the reference time difference type fluxgate sensor approach zero, including: Obtain output data of a reference time difference type fluxgate sensor; Construct a preliminary statistical characteristic model of the noise, set the objective function of minimizing the noise, and iteratively adjust the feedback parameters with a fixed step size; In each iteration, the degree of stochastic resonance achieved under the current feedback parameter settings is evaluated by comparing the output data with the expected output; If the ideal state is not reached, the feedback parameters are adjusted according to the error signal.

7. The magnetic field and magnetic gradient detection method according to claim 1, characterized in that: The output signals of the three three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensor are decomposed by wavelet transform, and the high-frequency noise components are filtered out by a threshold function while retaining the key features to obtain the denoising coefficient. Reconstruct the denoising coefficient to generate a low-noise signal and input it into an untrained neural network; Start the improved particle swarm algorithm, initialize the particle speed, position and dynamic parameters, and adjust the particle state in real time during iteration; When the global fitness value accuracy is met or the maximum number of iterations is reached, the optimal particle position sequence is output as the initial weight and threshold of the neural network; Drive the neural network training until the output error meets the set conditions; The output of the neural network is the denoised output signal.

8. The magnetic field and magnetic gradient detection method according to claim 1, characterized in that: After calculating the time difference through the output signal, the obtained time difference is denoised, including: Isolation forest anomaly detection is performed on the time difference. An anomaly detection model is constructed through dynamic window segmentation and adaptive parameter adjustment. For each detected anomaly point, the location and degree of anomaly are recorded. For each anomaly point, the surrounding normal points are selected as reference points and the weighted mean is used to replace the anomaly point to obtain the preprocessed signal. The preprocessed signal is input into the VMD decomposition module, and the mode number K and penalty factor are adaptively selected based on the characteristics of the preprocessed signal. , decompose the preprocessed signal into K modal components and calculate the energy entropy of each component; By normalizing the entropy value, weights are assigned according to the energy entropy of each modal component and weighted fusion is performed to generate an intermediate signal with enhanced main features; The intermediate signal enhanced by the main feature is subjected to hierarchical adaptive threshold calculation, soft threshold denoising is applied layer by layer, and the processing progress is iteratively judged. After completing all layers, the time difference is reconstructed.

9. A magnetic field and magnetic gradient detection device, characterized in that: include: Three three-component time-difference fluxgate sensors and one reference time-difference fluxgate sensor are placed in a spherical housing, forming a cross structure centered on a coordinate origin. Two of the three-component time-difference fluxgate sensors are symmetrically distributed along the X-axis, and the other three-component time-difference fluxgate sensor and the reference time-difference fluxgate sensor are symmetrically distributed along the Y-axis. The surface of the spherical shell is engraved with three enameled wire winding grooves that intersect perpendicularly along the X, Y, and Z directions respectively, for winding the spherical feedback coil; a controller for adjusting the current in the spherical feedback coil so that the values ​​of the three components of the reference time difference type fluxgate sensor approach zero; A detection circuit is used to input excitation signals into the three three-component time-difference type fluxgate sensors and the reference time-difference type fluxgate sensor; Used to obtain the output signals of three three-component time difference type fluxgate sensors and a reference time difference type fluxgate sensor; Used to calculate the time difference through the output signal; Used to calculate the measured magnetic field based on the time difference; It is used to obtain the magnetic induction signals of different three-component time-difference fluxgate sensors and reference time-difference fluxgate sensors based on the superposition of the measured magnetic field and the excitation magnetic field, and to construct a magnetic gradient tensor based on the magnetic induction intensity signals of different axes on the three-component time-difference fluxgate sensors and the reference time-difference fluxgate sensors.

Images