Study On Three-axis Digital Fluxgate Sensor And Its Intelligent Calibration Method

Fluxgate is a kind of device with good comprehensive performance for measuring weak magnetic field.It has the advantages of high resolution,large measuring range,good temperature stability and small perming error,etc.The three-axis fluxgate sensor,as a vector magnetic field sensor,is simple in structure,reliable in performance and low in cost,which is widely used in many fields including aviation,aerospace,marine,medical,robotics and geological exploration.With the development of application fields,a three-axis fluxgate sensor with high precision,high resolution and high temperature stability is urgently needed.For the three-axis traditional fluxgate sensor,fluxgate probe is assembled manually,which has the shortcomings of large volume,high power consumption,and complicated manufacturing process,etc.Analog circuit is adopted for signal processing and signal output,of which the parameter is affected easily by temperature and adjusted inconveniently.Calibration process is rather time-consuming,the performance of the calibration algorithm is prominently limited,and the stability of the calibration algorithm is relatively poor.In this paper,a three-axis digital fluxgate sensor and its intelligent calibration algorithms are studied.The n-type fluxgate probe based on printed circuit board is utilized to reduce volume,power consumption and machining difficulty on the premise of satisfying performance requirements.The digitalization of signal processing and signal output is used instead of analog circuit to improve temperature stability.The intelligent calibration algorithm is proposed to improve calibration efficiency and accuracy.The main research achievements and innovations are as follows:(1)A n-shaped low-power fluxgate probe structure based on printed circuit board is proposed.In order to achieve higher sensitivity,lower noise,lower power consumption,lower volume and more efficient manufacturing process,combined with PCB technique and winding technique,a n-shaped low-power fluxgate probe structure based on printed circuit board is proposed.In the fluxgate probe,a n-shaped soft magnetic core with different cross-sectional area in each part is adopted,and two printed circuit boards are used to make skeleton.Excitation coil is wound around the skeleton corresponding to large cross-sectional area of the magnetic core,and induction coil is wound around the skeleton corresponding to small cross-sectional area at both sides of the magnetic core.Both the excitation coil and induction coil are led out by corresponding pads on the printed circuit board.Contracted core structure of the fluxgate probe is beneficial for reducing excitation current,which results in less power consumption.Opening structure of the n-shaped magnetic core reduces volume of the fluxgate probe and improves the manufacturing efficiency while low power consumption is also taken into account.For a given voltage excitation source in practical applications,the performance of fluxgate probe is simulated and analyzed by finite element method from the aspects of iron core,coil and excitation signal,and the interaction among those parameters is analyzed.The parameters of fluxgate probe are optimized to reduce power consumption further.(2)A design method and intelligent phase synchronization method of the multi-even harmonic reference signal are proposed.The output signal of fluxgate probe is mainly an Alternating-Current signal composed of harmonics,in which the even harmonic components contains the external magnetic field information.The signal detection of fluxgate probe in digital fluxgate sensors is performed by the phase-sensitive detection(phase-sensitive demodulation)method.The method using the second harmonic component as reference signal ignores the influence of other even harmonic components on phase detector,the method using the multiple output signals of fluxgate probe to synthesize reference signal is difficult to effectively remove the effect of odd harmonic components,which reduces the sensitivity of digital fluxgate sensors.In this paper,a design method of multi-even harmonic reference signal is proposed for three-axis digital fluxgate sensors.Besides,when multi-even harmonic reference signal or square wave reference signal are used,an intelligent phase synchronization method is proposed,which effectively improves the demodulation efficiency.The results show that the multi-even harmonic reference signal design method suggests an effective way to improve the sensitivity of fluxgate sensor.(3)An improved ellipse fitting calibration algorithm based on weighted approximate distance and a nonlinear calibration algorithm based on harmonic decomposition are proposed.The traditional ellipse fitting calibration algorithm does not need external reference,and is often used to compensate zero bias error,scale factor error,non-orthogonal error,soft iron error and hard iron error,etc.In this algorithm,the least square method of simplified distance from sample data to ellipse is adopt to realize the optimal estimation of ellipse coefficients,which leads to large deviation in heading determination applications,In addition,the nonlinearity of errors is neglected,which results in a limited calibration precision.In this paper,an improved ellipse fitting algorithm based on weighted approximate distance is proposed.In the algorithm,the objective function is a least square estimation based on weighted approximate distance,which is obtained by the Taylor expansion of the two-order elliptic equation,and the weighting coefficient is related to the heading.Meanwhile,taking into account the nonlinearity of errors,a nonlinear calibration algorithm besed on harmonic decomposition is proposed.In the algorithm,harmonic decomposition method is used to establish nonlinear calibration model,the linear compensation coefficients and nonlinear compensation coefficients are determined iteratively.The results show that the proposed calibration algorithms can effectively improve the accuracy of fluxgate sensors.(4)An improved ellipsoid fitting calibration algorithm for three-axis fluxgate sensor with limited sample data is proposed.After the three-axis fluxgate sensor is mounted on vehicles,calibration is essential because of the interference of surrounding ferromagnetic materials.However,due to the limitation of vehicle maneuverability,the sample data for calibration is usually limited in a certain attitude aera.Traditional ellipsoid fitting algorithms generally require the sampled data to be distributed in a uniform attitude range,and the algorithms are unstable in process of calculating ellipsoid coefficients caused by the singularity of constraint matrix.In this paper,an intelligent calibration method for limited sample data is proposed.In the method,an improved ellipsoid fitting calgorithm based on approximate distance is presented,in which the objective function is represented by the mean square value of approximate distances,and the constraint is related to the sample data.The ellipsoid coefficients are calculated by constrained least square method.In addition,for the calculation process of ellipsoid coefficients,a method based on matrix decomposition is proposed to overcome the algorithm instability.The proposed calibration method does not require any external reference,and only need three simple rotation operations of vehicle in the restricted area to obtain sample data,which simplifies the calibration process.In this paper,a three-axis digital fluxgate sensor with the resolution of 0.01°,the precision of 0.41°(pitch angle: 0°?±60°)and the operating temperature range of-40??85? has been developed,which has been successfully applied to UAVs and high-altitude balloons.Furthermore,based on the research in the paper,the improved small three-axis fluxgate sensor has been used in While Drilling Measurement systems.