GMI磁传感器及磁异信号检测技术研究

发布时间：2019-06-24 19:31

【摘要】：与目前常用的基于磁通门(Fluxgate)、霍尔效应(Hall)和磁电阻效应等的磁传感器相比较,基于巨磁阻抗效应(Giant Magneto-Impedance,简称GMI)的GMI磁传感器还具有灵敏度高、体积小、响应速度快、磁滞小以及功耗低等优点,在微弱磁场测量等领域具有广阔的应用前景。与此同时,以地球磁场为背景的磁异检测(Magnetic Anomaly Detection,简称MAD)技术,具有无源被动检测、隐蔽性能好以及抗干扰能力强等特点,可广泛应用于资源勘探、交通运输和军事目标探测与跟踪等领域。因此,开展GMI磁传感器及磁异信号检测技术方面的研究具有重要的理论意义和工程应用价值。本文围绕着GMI磁传感器的设计以及铁磁性目标所引起的磁异信号的检测问题等内容展开深入的研究,其主要包括:(1)简要介绍了GMI效应及其与驱动电流频率之间的关系等内容,研究了所选用非晶丝材料的频率特性,并据此选取了敏感探头的驱动电流频率和线性工作区间;同时,利用直流线圈提供偏置磁场的方式,设计了一个差分式结构的敏感探头,提出了利用具有强噪声抑制能力的正交锁定差分电路代替峰值检测电路作为GMI磁探头的信号调理电路,并详细分析了该信号调理电路的工作原理以及该GMI磁传感器的各项性能;另外,根据第二代电流传输器(second generation Current Conveyor,简称CCII)和跨导运算放大器(Operational Trans-conductance Amplifier,简称OTA)等有源模块的工作原理,利用现有商业芯片AD844,提出了一个振荡条件和谐振频率独立可调节的双输出正交正弦振荡器,并详细介绍了该振荡器的工作原理以及进行了非理想条件分析。(2)建立了磁异检测的数学模型,详细推导了基于三个正交基函数的磁异信号的表达式,并针对目标信号的特征时间已知的条件下,根据正交匹配滤波器组理论,提出并详细阐述了适用于白噪声情形下的基于正交基函数匹配滤波器组的磁异检测系统,并以此为基础,构建并详尽分析了可应用于有色噪声情形下的基于LMS自适应AR模型的白化滤波器和正交基函数匹配滤波器组相结合的磁异检测系统。同时,针对1/f分形噪声为外界磁场环境和电路等噪声的主要来源之一的情形,通过利用仿真验证的方式,详细分析了文中所设计的自适应白化滤波器的预白化效果和磁异检测系统的性能,同时利用实测数据进一步验证了该检测系统的可行性和优越性。(3)以自回归-滑动平均ARMA模型理论为基础,简要介绍了可应用于不平稳时间序列情形下的求和自回归-滑动平均(Auto-Regressive Integrated Moving Average,简称为ARIMA)模型及其建立流程,并针对磁异目标信号的特征时间未知且背景噪声为非平稳时间序列的情形,根据ARMA模型的L步预测公式,提出了基于ARIMA模型预测的磁异检测算法,且利用实测数据和仿真分析相结合的手段,详细分析了该检测算法的性能并指出其存在的局限性。同时,以灰色关联度理论为基础,进一步提出了基于ARIMA模型残差与灰色关联度相结合的磁异检测算法,并通过仿真验证的方式对其性能进行了详细分析。(4)简要介绍了离散小波变换与逆变换等理论,针对快速离散小波变换的分解过程中存在数据逐级减少等问题,引出了非抽样离散小波变换的概念;利用小波变换具有时域-频域空间局部化特性,根据小波阈值消噪的原理,通过分析传统硬阈值法和软阈值法中存在的不足之处,提出了基于加权模糊函数的小波阈值消噪的磁异检测算法,并分别通过仿真验证和实测数据分析的方式,详细分析了该检测算法的性能并指出该算法的优越性;此外,利用小波变换对1/f噪声具有预白化作用的特性,提出了基于非抽样离散小波变换的小波系数能量的磁异检测算法,不但详细推导了该检测算法的检测概率、虚警概率等理论表达式,而且通过一系列仿真验证实验,详尽分析了该检测器的性能。
[Abstract]:Compared with the current magnetic sensors based on Fluxgate, Hall effect (Hall) and magnetic resistance effect, the GMI magnetic sensor based on the giant magnetic impedance effect (GMI) also has the advantages of high sensitivity, small volume, high response speed, Has the advantages of low magnetic hysteresis, low power consumption and the like, and has wide application prospect in the fields of weak magnetic field measurement and the like. At the same time, the magnetic anomaly detection (MAD) technology based on the earth's magnetic field has the characteristics of passive passive detection, good hiding performance, strong anti-interference ability and the like, and can be widely applied to the fields of resource exploration, transportation and military target detection and tracking. Therefore, the research of GMI magnetic sensor and magnetic field signal detection technology is of great theoretical significance and engineering application value. In this paper, the research on the design of the GMI magnetic sensor and the detection of the magnetic anomaly signal caused by the ferromagnetic target has been carried out in-depth. The main contents of this paper are as follows: (1) The GMI effect and its relation with the driving current frequency are briefly introduced. the frequency characteristic of the selected amorphous wire material is studied, the driving current frequency and the linear working range of the sensitive probe are selected according to the frequency characteristic, A signal conditioning circuit using an orthogonal lock difference circuit with strong noise suppression capability instead of a peak detection circuit as a GMI magnetic probe is proposed, the working principle of the signal conditioning circuit and the performance of the GMI magnetic sensor are analyzed in detail; in addition, According to the working principle of the active module such as the second generation current (CCII) and the transconductance operational amplifier (OTA), the present commercial chip AD844 is used to propose a dual-output quadrature sinusoidal oscillator with independent adjustable oscillation and resonant frequency. The working principle of the oscillator and the non-ideal condition analysis are introduced in detail. (2) The mathematical model of the magnetic anomaly detection is established, the expression of the magnetic anomaly signal based on the three orthogonal basis functions is derived in detail, and under the condition that the characteristic time of the target signal is known, according to the theory of the orthogonal matched filter bank, The magnetic anomaly detection system based on the orthogonal base function matching filter bank in the case of white noise is presented and described in detail, and based on this, The combination of the whitening filter and the orthogonal base function matching filter bank of the LMS adaptive AR model, which can be applied in the case of colored noise, is constructed and analyzed in detail. At the same time, for the case of 1/ f fractal noise as one of the main sources of the external magnetic field environment and the circuit, the pre-whitening effect of the adaptive whitening filter and the performance of the magnetic anomaly detection system are analyzed in detail by means of simulation and verification. At the same time, the feasibility and superiority of the detection system are further verified by using the measured data. (3) Based on the theory of the autoregressive-sliding average ARMA model, the auto-regressive integrated moving average (ARIMA) model which can be applied to the case of non-stationary time series and its establishment flow are introduced. and aiming at the situation that the characteristic time of the magnetic anomaly target signal is unknown and the background noise is a non-stationary time sequence, according to the L-step prediction formula of the ARMA model, a magnetic anomaly detection algorithm based on the ARIMA model prediction is proposed, The performance of the detection algorithm is analyzed in detail and its limitations are pointed out. At the same time, based on the theory of grey degree of association, a magnetic anomaly detection algorithm based on the combination of the residual of ARIMA model and the gray correlation degree is proposed, and its performance is analyzed in detail by means of simulation and verification. (4) The theory of discrete wavelet transform and inverse transform is introduced in brief. The concept of non-sampling discrete wavelet transform is derived for the problem of step-by-step reduction in the decomposition of the fast discrete wavelet transform. The local characteristics of time-domain and frequency-domain space are obtained by using the wavelet transform. Based on the principle of wavelet threshold de-noising, by analyzing the shortcomings of the traditional hard-threshold method and the soft-threshold method, a magnetic anomaly detection algorithm based on the wavelet threshold de-noising based on the weighted fuzzy function is proposed, The performance of the detection algorithm is analyzed in detail and the superiority of the algorithm is pointed out. In addition, using the characteristic of the wavelet transform to prewhiten the 1/ f noise, a magnetic anomaly detection algorithm based on the wavelet coefficient energy of the non-sampled discrete wavelet transform is proposed. The detection probability, false alarm probability and other theoretical expressions of the detection algorithm are derived in detail, and the performance of the detector is analyzed in detail through a series of simulation and verification experiments.
【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN911.23;TP212.13

【相似文献】