Spurious Peaks Research Articles

A refined output-only modal identification technique for structural health monitoring of civil infrastructures

The increasing installation of structural health monitoring systems has raised the need for efficient and rapid data interpretation algorithms. Operational modal analysis is currently one of the most used techniques for extracting modal properties, and some attempts have been made to automate this procedure. However, automated techniques often require manual calibration of hyperparameters and show inconsistencies across different case studies. This paper proposes a refined version of the automated frequency domain decomposition (AFDD) using the modal assurance criterion (MAC) to obtain natural frequencies and mode shapes. Sensitivity analyses were conducted on the ambient vibration response of the Yonghe cable-stayed bridge in China, accounting for influential factors including noise levels, acceleration record length, sensor layouts. A novel approach is then introduced to interpret the results of the sensitivity analyses. The approach consists in plotting the result of each analysis in a stabilization diagram and then using a Gaussian mixture model that clusters the poles into core and outliers. This allows to identify regions where the MAC thresholds are optimal. By comparing the results of all the sensitivity analyses it was possible to define a single optimal MAC threshold, avoiding the need for fixing a value based on the user’s experience. Three substantially different case studies were analyzed to extensively test the methodology: the Yonghe cable-stayed bridge, the PolyU footbridge in Hong Kong, and Moletta Tower in the Circus Maximus archeological site in Italy. The analysis compared the proposed AFDD algorithm to the traditional frequency domain decomposition and covariance-driven stochastic subspace identification. Specifically, the efficiency in identifying close frequencies, weakly excited modes, spurious peaks, and complex modes was evaluated for each method, which highlighted the robustness of the proposed optimized AFDD. The analysis showcases peculiar characteristics and drawbacks of each method when trying to identify complex vibrational modes of the specific case studies. It was found that the proposed AFDD procedure performs better than traditional methods despite it may misidentify complex modes due to the constraints of a narrower modal domain and similar geometries.

Characterizing the features of the low-amplitude peaks in delta Scuti stars with TESS

Abstract The presence of low-amplitude peaks over the noise in the power spectra of δ Scuti stars is frequently disregarded. These seemingly insignificant peaks, collectively referred to as grass, might contain valuable information about the origin of these stars and the reasons behind the occurrence or absence of a plateau. It is crucial to systematically parameterize the grass phenomenon throughout a comprehensive sample that covers the entire δ Scuti star parameter range. Thus, we conduct a quantitative study of long-duration, high-duty-cycle TESS light curves, leading to improved detection methods for plateaus and a deeper understanding of their nature. This approach minimizes the impact of unresolved peaks caused by mode variations over time. Additionally, we present appropriate analysis techniques to mitigate window effects and identify and eliminate spurious peaks. We demonstrate here that the grass can be effectively parameterized based on peak density. With such parameterization two distinct regimes are found: the sparse grass regime, characterized by low peak density and the absence of a plateau in the power spectra, and the dense grass regime, characterized by high peak densities and the presence of an observable plateau. Our study is the first rigorous quantification of the emergence of such a plateau in the power spectra of δ Scuti stars. Since the grass might be related with fractality, mode variability, and stellar rotation rate, its parameterization opens a new way to analyze these stars.

D-BJH: The Intrinsic Model for Characterizing the Pore Size Distribution of Porous Materials.

A novel approach, the differential Barrett-Joyner-Halenda model (D-BJH), is proposed to address the limitations of the traditional BJH model in determining the pore size distribution (PSD). This method integrates multilayer adsorption and capillary condensation processes using advanced numerical techniques, notably, a Galerkin-based framework for solving the differential BJH equations. The D-BJH model offers a precise analytical framework, establishing itself as a benchmark in adsorption theory. It enables accurate PSD calculations and provides robust theoretical support for interpreting adsorption isotherms. Additionally, a simplified D-BJH method allows for direct point-by-point PSD calculation, reducing computational complexity and eliminating cumulative errors inherent in traditional methods like BJH, DH, CI, and VBS. The D-BJH model demonstrates that PSD is directly related to the differential function of the isotherm and pore size concerning relative pressure. Comparative analyses with BJH, DH, and NLDFT models show that D-BJH effectively explains fundamental adsorption mechanisms, such as the "spurious peak" observed in desorption branches. This work expands the applicability of the BJH model, proving that both the D-BJH model and its simplified method are accurate and comprehensive for characterizing the pore structures of microporous and mesoporous materials.

Quantum nature of reactivity modification in vibrational polariton chemistry.

In this work, we present a mixed quantum-classical open quantum system dynamics method for studying rate modifications of ground-state chemical reactions in an optical cavity under vibrational strong-coupling conditions. In this approach, the cavity radiation mode is treated classically with a mean-field nuclear force averaging over the remaining degrees of freedom, both within the system and the environment, which are handled quantum mechanically within the hierarchical equations of motion framework. Using this approach, we conduct a comparative analysis by juxtaposing the mixed quantum-classical results with fully quantum-mechanical simulations. After eliminating spurious peaks that can occur when not using the rigorous definition of the rate constant, we confirm the crucial role of the quantum nature of the cavity radiation mode in reproducing the resonant peak observed in the cavity frequency-dependent rate profile. In other words, it appears necessary to explicitly consider the quantized photonic states in studying reactivity modification in vibrational polariton chemistry (at least for the model systems studied in this work), as these phenomena stem from cavity-induced reaction pathways involving resonant energy exchanges between photons and molecular vibrational transitions.

Quality Control of Doppler Spectra from a Vertically Pointing, S-Band Profiling Radar

Abstract This study describes a novel combination of methods to remove spurious spectral peaks, or “spurs,” from Doppler spectra produced by a vertically pointing, S-band radar. The University of Massachusetts S-band frequency-modulated, continuous-wave radar (UMass FMCW) was deployed to monitor the growth of the CBL over northern Alabama during the VORTEX–Southeast field campaign in 2016. The Doppler spectra contained spurs caused by high-voltage switching power supplies in the traveling wave tube amplifier. In the original data-processing scheme for this radar, a median filtering method was used to eliminate most of the spurs, but the largest ones persisted, which significantly degraded the quality of derived radar moments (e.g., reflectivity, Doppler velocity, and spectrum width) and hindered further analysis of these data (e.g., hydrometeor classification and boundary layer height tracking). Our technique for removing the spurs consists of three steps: (i) a Laplacian filter identifies and masks peaks in the spectra that are characteristic of the spurs in shape and amplitude, (ii) an in-painting method then fills in the masked area based on surrounding data, and (iii) the moments data (e.g., reflectivity, Doppler velocity, and spectrum width) are then recomputed using a coherent power technique. This combination of techniques was more effective than the median filter at removing the largest spurs from the Doppler spectra and preserved more of the underlying Doppler spectral structure of the scatterers. Performance of both the median-filter and the in-painting methods is assessed through statistical analysis of the spectral power differences. Downstream products, such as boundary layer height detection, are more easily derived from the recomputed moments. Significance Statement This manuscript describes a novel combination of image and signal processing techniques used to recover meteorological observations from corrupted Doppler radar spectra. This successful recovery of meteorologically significant information illustrates the importance of retaining Doppler spectra when practical. In seeking solutions to data quality issues, the atmospheric science community should remain cognizant of promising techniques offered by other disciplines. We present this data rescue study as an example to the meteorological community.

Reconstructing lattice QCD spectral functions with stochastic pole expansion and Nevanlinna analytic continuation

The reconstruction of spectral functions from Euclidean correlation functions is a well-known, yet ill-conditioned inverse problem in the fields of many-body and high-energy physics. In this paper, we present a comprehensive investigation of two recently developed analytic continuation methods, namely stochastic pole expansion and Nevanlinna analytic continuation, for extracting spectral functions from mock lattice QCD data. We examine a range of Euclidean correlation functions generated by representative models, including the Breit-Wigner model, the Gaussian mixture model, the resonance-continuum model, and the bottomonium model. Then, we apply the two methods to reconstruct spectral functions from charmonium correlation functions computed using lattice QCD simulations at a finite temperature. Our findings demonstrate that the stochastic pole expansion method, when combined with the constrained sampling algorithm and the self-adaptive sampling algorithm, successfully recovers the essential features of the spectral functions and exhibits excellent resilience to noise of input data. In contrast, the Nevanlinna analytic continuation method suffers from numerical instability, often resulting in the emergence of spurious peaks and significant oscillations in the high-energy regions of the spectral functions, even with the application of the Hardy basis function optimization algorithm. Published by the American Physical Society 2024

S-method based on modified group delay for time-frequency analysis

A noise robust time-frequency representation based on the S-method (SM) and modified group delay (MGD) (S-MGD) is proposed in this article. The SM is free from cross-terms as it is the sum of individual Wigner distributions of different signal components and provides high-energy concentration both in time and frequency. The MGD reduces the Gibbs ripple and spurious spectral peaks because of noise, without using any frequency smoothing, and thus maintains the best frequency resolution as that of rectangular window. Hence,a combination of SM and MGD provides the desirable features of both the methods. The proposed method is tested for different signals like time varying random process, frequency shift keying, synthesized car engine sound/pressure signals using a solo combustion and a real-time echolocation pulse emitted by a big brown bat (Eptesicus fuscus) which indicate that the new S-MGD has improved noise immunity and frequency resolution compared to the method that uses frequency smoothing to suppress noise and Gibbs effects.

Mechanical Fault Sound Source Localization Estimation in a Multisource Strong Reverberation Environment

Aiming at the sound source localization of mechanical faults in a strong reverberation scenario with multiple sound sources, this paper investigates a mechanical fault source localization method using the U-net deep convolutional neural network. The method utilizes the SRP-PHAT algorithm to calculate the response power spectra of the collected multichannel fault signals. Through the utilization of the U-net neural network, the response power spectra containing spurious peaks are transformed into “clean” estimated source distribution maps. By employing interpolation search, the estimated source distribution maps are processed to obtain location estimations for multiple fault sources. To validate the effectiveness of the proposed method, this paper constructs an experimental dataset using mechanical fault data from electromechanical equipment relays and conducts sound source localization experiments. The experimental results show that the U-net network under 0.2 s/0.5 s/0.7 s reverberation time can effectively eliminate spurious peak interference in the response power spectrum. As the signal-to-noise ratio decreases, it can still distinguish the sound sources with a distance of 0.2 m. In the context of multifault source localization, the method is capable of simultaneously locating the positions of four fault sources, with an average localization error of less than 0.02 m. The method in this paper effectively eliminates spurious peaks in the response power spectra under conditions of multisource strong reverberation. It accurately locates multiple mechanical fault sources, thereby significantly enhancing the efficiency of mechanical fault detection.

Q-space imaging based on Gaussian radial basis function with Laplace regularization.

To introduce the diffusion signal characteristics presented by spherical harmonics (SH) basis into the q-space imaging method based on Gaussian radial basis function (GRBF) to robustly reconstruct ensemble average diffusion propagator (EAP) in diffusion MRI (dMRI). We introduced the Laplacian regularization of the signal into the dMRI imaging method based on GRBF, and derived the relevant indicators of microstructure imaging and the orientation distribution function (ODF) providing fiber bundle direction information based on EAP. In addition, this method is combined with a multi-compartment model to calculate the diameter of fiber bundle axons. The evaluation of the results included qualitative comparisons and quantitative assessments of the signal fitting. The results show that the proposed method achieves the more significant accuracy improvement in reconstructing signal. Meanwhile, ODFs estimated by the proposed method show the sharper profiles and less spurious peaks, even under the sparse and noisy conditions. In the 36 sets of axon diameter estimation experiments, 34 and 30 sets of results showed that the proposed method reduced the mean and SD of axon diameter estimates, respectively. Moreover, compared with the current state-of-the-art method, the mean and SD of axon diameter estimated by the proposed method are mostly lower, with 32 and 29 of 36 groups. The proposed method outperforms the GRBF regarding signal fitting and the estimation of the EAP and ODF with multi-shell sparse samples. Moreover, it shows the potential to recover important features of microstructures with less uncertainty by using proposed method together with multi-compartment models.

Impact of image registration errors on the quality of hyperspectral images in imaging static Fourier transform spectrometry

Imaging static Fourier transform spectrometry (isFTS) is used for pushbroom airborne or spaceborne hyperspectral remote sensing. In isFTS, a static two-wave interferometer imprints linear interference fringes over the image of the scene, so that the spectral information is multiplexed over several instantaneous images, and numerical reconstruction is needed to recover the full spectrum for each pixel. The image registration step is crucial since insufficient accuracy leads to artefacts on the images and the estimated spectra. In order to investigate these artifacts, we performed a theoretical study and designed a simulation program. We established that registration errors create crenellated spatial patterns, the magnitude of which depends on the radiance gradient of the scene, the amplitude of the registration error, and the wavelength. In the case of sinusoidal perturbations, which may correspond for instance to mechanical vibrations of the carrier, we established that spurious peaks appear on the spectrum, similarly to what happens in dynamic FTS, but with spatial patterns specific to static interferometers.

A defect location method for power cable based on Burg power spectral

AbstractThe frequency‐domain reflection (FDR) has been demonstrated to be a trustworthy technique in solving the defect location of power cable by field experiments. However, the location spectrum of the FDR requires manual window smoothing and can be disturbed by spurious peaks. Aiming at these shortcomings of FDR, a new method of cable defect location based on Burg power spectral (BPS) is introduced in this paper. The idea of this method is to use linear difference variance to fit the distribution of reflection coefficient spectrum and build an auto‐regressive (AR) model. The Burg algorithm is employed to estimate the coefficients model and calculate the power distribution of the AR model. Then, the cable defects will be located by BPS with high precision and resolution. In this method, the fast Fourier transform (FFT) with windowed function is replaced by an AR model without windowed function. This suppressed the impact of spurious peaks or spectrum leakage in FFT on the localization defects, and the localization resolution is higher. Finally, we validate the feasibility and effectiveness of BPS through experiments conducted on a 500 m laboratory cable and a 9.6 km submarine cable.

Plasma-Induced Interfacial Processes in Metal Halides FTIR Gas Cell Windows

Fourier transform infrared spectroscopy (FTIR) is one of the most widely used vibrational diagnostic techniques to investigate gas-phase reactive oxygen and nitrogen species (RONS). However, the technique carries intrinsic challenges, particularly in relation to interfering peaks in the spectral data. This study explores the interfacial processes that occur when reactive oxygen and nitrogen species generated by a non-equilibrium air plasma interact with the metal halide windows of an FTIR gas cell, leading to the appearance and evolution of spurious absorption peaks which complicate spectral interpretation. Raman spectroscopy, X-ray photoelectron spectroscopy, time of flight secondary ion mass spectrometry and attenuated total reflectance-FTIR spectroscopy were used to elucidate the origin of spurious absorption peaks spanning the 1400–1300 cm−1 spectral range as a result of KBr exposure to plasma generated species. It was found that plasma exposed KBr contained a lower atomic fraction of Br which was replaced by the NO3 nitrate group, the main absorbance peak of which progressively evolved with plasma exposure and affected the window transparency over the corresponding FTIR region. A correlation was revealed between KNO3 formation, plasma power and exposure time to a growth and change in molecular vibrational energies corresponding to asymmetric NO3 stretching vibrations in the KNO3 structure.

Region of interest selection for GC×GC–MS data using a pseudo fisher ratio moving window with connected components segmentation

Comprehensive two-dimensional gas chromatography mass spectrometry (GC×GC–MS) data present several challenges for analysis largely because chemical factors drift along the chromatographic modes across different chromatographic runs, and there is frequently a lack of reliable molecular ion measurements with which to align data across multiple samples. Tensor decomposition techniques such as Parallel Factor Analysis (PARAFAC2/PARAFAC2×N) allow analysts to deconvolve closely eluting signals for quantitative and qualitative purposes. These techniques make relatively few assumptions about chromatographic peak shapes or the relative abundance of noise and allow for highly accurate representations of the underlying chemical phenomena using well-characterized and scrutinized principles of chemometrics. However, expert intervention and supervision is required to select appropriate Regions of Interest (ROI) and numbers of chemical components present in each ROI. We previously reported an automated ROI selection algorithm for GC–MS data in Giebelhaus et al. where we observed the ratio of the first and second eigenvalues within a moving window across the entire chromatogram. Here, we present an extension of this work to automatically detect ROIs in GC×GC–MS chromatograms, while making no assumptions about peak shape. First, we calculate the probabilities of each acquisition being in a ROI, then apply connected components segmentation to discretize the regions of interest. For sparse chromatograms we found the algorithm detected spurious peaks. To address this, we implemented an iterative ROI selection process where we autoscaled the moving window to the standard deviation of the noise from the previous iteration. Using three user-defined parameters, we generated informative ROIs on a wide range of GC×GC-TOFMS chromatograms.

High accurate detection method for aortic valve opening of seismocardiography signals

Accurately detecting the aortic valve opening (AO) peaks of seismocardiography (SCG) is a challenging problem due to interference of other morphological inflection points. In this paper, a high accurate method is proposed to extract AO peaks based solely upon the SCG data. The raw SCG signal is purified by a simple first order interference cancellation method, resulting in a reduced number of modes to be decomposed. The purified SCG signal is then decomposed into a series of quasi-orthogonal modes by successive variational mode decomposition (SVMD) without any prior about the number of modes. Considering the pulsatile nature of the AO signal, a waveform factor criterion is proposed to reconstruct the AO signal based on the pulsatile level of each mode. A seventh power law detector is designed to amplify the AO peak and suppress spurious peaks. The publicly available combined measurement of electrocardiogram (ECG), breathing and seismocardiograms (CEBS) database is exploited to verify the performance of the proposed method. We show that the average sensitivity of our technique is 99.02%, the prediction rate is 99.06%, and the detection accuracy is 98.10%, which is superior to several state-of-the-art methods. In addition, compared with the ECG reference value, the instantaneous heart rate extracted by the proposed method is in good agreement with that of the ECG, e.g., the maximum average absolute error percentage is as low as 2.11 and the maximum average value relative error is about 0.03, further demonstrating that the proposed method can achieve accurate estimates of heart rate with an accelerometer alone.

A Sparse Bayesian Learning Method for Direction of Arrival Estimation in Underwater Maneuvering Platform Noise

The underwater maneuvering platform generates self-noise when sailing, which shows spatial directionality to the arrays fixed on the platform. In this paper, it is called spatially colored noise (SCN). The direction of arrival (DOA) estimation results are often influenced by this self-noise, leading to a decrease in estimation accuracy and to the appearance of spurious peaks. To resolve this problem, a sparse Bayesian learning (SBL) method adapted to underwater maneuvering platform noise is proposed in this paper. The SBL framework with unknown SCN is established first. Then, the SCN covariance matrix is estimated by projecting the received data covariance matrix into the noise subspace, and the DOA estimation results are finally obtained through multiple iterations. The simulation results show that the proposed method avoids spurious peaks, and compared to the existing methods, the proposed method achieves a higher accuracy in the case of low SNRs and small snapshot numbers. The sea trial data processing results show that the proposed method provides lower and flatter noise spectrum levels without spurious peaks.

Incorporation of antimony-bearing mining wastes into clinker Portland raw feed: The difficulty of Sb analysis in calcium silicates.

Within the global trend to valorise various mineral wastes as substituents in Portland cement clinker raw feed, mining wastes are promising candidates. However, they might still contain high levels of metallic elements. Their fate in the kiln is not always understood as well as their incorporation within the various clinker's phases. This is especially the case for antimony. Its in situ microanalysis by the means of energy-dispersive spectroscopy is tricky since several of its L lines (Lα1 = 3604,72eV, Lα2 = 3595,32eV) overlap with calcium lines (Kα1 = 3691,68eV, Kα2 = 3688,09eV). Hence, at low concentrations, it is not possible to visualise its characteristics peaks. Increasing the counting rate by increasing the acceleration current results in the generation of spurious sum peaks, rendering the analysis not viable. Wavelength dispersive spectroscopy (i.e. electron microprobe) allows a better spectral resolution and quantification of Sb in the clinker phases. In a Portland cement clinker doped with 1% of Sb-bearing mining waste, the Sb2O3 content in belite and alite is in the 0.2-0.4wt.% range as well as for C3A; in C4AF the content is higher, from 1.4wt.% to 2wt.%. However, there are microstructural evidence that Sb forms blebs from less than 1 μm in diameter up to 10 μm, included in calcium silicates. Hence, its incorporation in the lattice of calcium silicates is dubious. However, for ferrites and aluminates there is no microstructural evidence for remaining Sb-bearing phases, suggesting a direct incorporation in crystal lattices.

Deconvolution without calibration measurement of multi-peak charge distributions in dielectric materials

The electrical and mechanical properties of a dielectric material are significantly influenced by its charge distribution, necessitating an understanding of how charge accumulation occurs in order to determine the suitability of use for specific applications. This research presents a novel approach to deconvolving charge distributions with multiple peaks while dispensing with calibration measurement required by the non-contact pulsed electro-acoustic technique. Unlike traditional deconvolution methods that rely on reference signal measurements, the proposed approach reconstructs the charge distribution through parameters directly estimated from the piezoelectric transducer output voltage, effectively eliminating the need for such measurements altogether. Simulations show that even when closely spaced peaks exist within multi-peak charge distributions, our method can accurately reconstruct their profiles. In addition, the charge distributions obtained from experimental data of PTFE, FEP, and PEEK indicate that the proposed approach outperforms traditional methods in predicting penetration depth while reducing fluctuations and spurious peaks. We expect this method to be a valuable tool for the design and optimization of electronic systems in harsh environments through offering a simple, fast, and accurate alternative to conventional deconvolution methods.

Parameter Estimation in Spectral Resolution Enhancement Based on Forward-Backward Linear Prediction Total Least Square Method.

In a Fourier transform infrared (IR) spectrometer, the Michelson interference signal extrapolation method based on linear prediction is often used to improve spectral resolution. In this method, an autoregressive (AR) model is established for the Michelson interference signal in the spectrometer. Once the AR model parameters are determined, the AR process is predictable. The interference signal can be used to figure out the AR model's parameters. Based on this, the AR model can be used to extrapolate the interference signal to improve the spectral resolution. In this paper, the forward-backward linear prediction total least squares (FB-TLS) method is proposed to estimate the parameters of the AR model. The parameters that are estimated are used to improve the IR spectral resolution. By simulating different order and signal-to-noise ratio situations, the effects of the Burg, the least square, and the FB-TLS parameter estimation methods on spectral resolution enhancement are studied. The simulation results demonstrate that the FB-TLS parameter estimation method can effectively suppress noise and avoid spurious peaks. The experimental results demonstrate that the FB-TLS parameter estimation method is effective for spectral resolution enhancement technology based on linear prediction. When the FB-TLS method is used to enhance NH3 IR spectral resolution from 2 cm-1 to 1 cm-1, the spectral prediction error in the NH3 characteristic band is only 0.21% compared with the measured NH3 spectrum, whose spectral resolution is 1 cm-1.

In‐house manufactured benchtop XRF spectrometer

AbstractThe growing efficiency of agribusiness depends on the development of technology. A key challenge regards the determination of plant mineral content, which help farmers to make a decision about fertilizer use. Currently, such nutritional status diagnoses are made based on wet chemical methods; they are usually slow and generate chemical wastess. In this context, an in‐house manufactured XRF benchtop for agri‐samples is proposed in this study. This XRF system was furnished with a manual primary filter changer, vacuum chamber, dual interlock radiological protection system, which were installed in the equipment cover, and light signals. Analytical capabilities were also evaluated. The equipment presented spurious peaks of Ti, Fe, Ni, and Zn, which is not ideal for some agronomical applications; this issue might be overcome in the future by means of additional primary filters. The total x‐ray spectrum counting rate is roughly one magnitude order lower than two other commercial XRF systems, but the signal‐to‐noise ratio is quite similar. The vacuum chamber under partial vacuum improves the Al, Si, P and S signals from 1.5 to 2.2. The limits of detection (LOD, mg kg−1) for P (261), S (168), K (68), and Ca (50) macronutrients and Mn (14), Fe (15), Cu (9), and Zn (8) micronutrients are close to the ones reported in the literature. The limits of detection for P, S, K, and Ca are suitable for plant analysis.

Determination of four aflatoxins in feeds by high throughput automated immunoaffinity magnetic beads purification-ultra performance liquid chromatography

Aflatoxin (AFT) is an extremely toxic and highly toxic carcinogenic substance. This is particularly problematic due to the risk of aflatoxin contamination in raw feed materials and products during production, transportation, and storage. In this study, immunoaffinity magnetic beads (IMBs) were prepared for the purification of four aflatoxins (aflatoxin B1 (AFB1), aflatoxin B2 (AFB2), aflatoxin G1 (AFG1) and aflatoxin G2 (AFG2)). The aflatoxin contents were then determined rapidly and accurately using ultra performance liquid chromatography (UPLC). More specifically, the coupling ratio of magnetic beads (MBs) to the aflatoxin monoclonal antibody was initially optimized, wherein an MB volume of 1 mL and an antibody content of 2.0 mg was found to meet the purification requirements of this method. The magnetic properties of the MBs and the IMBs were then investigated using a vibrating sample magnetometer (VSM) at room temperature. As a result, the maximum saturation super magnetizations of the MBs and the IMBs were determined to be 28.61 and 23.22 emu/g, respectively, indicating that the saturation magnetization intensity of the IMBs was reduced by coupling with a non-magnetic antibody. However, the saturation magnetization intensity remained sufficiently high to permit magnetic separation from the solution. In addition, the appearance of the IMBs was examined using a biomicroscope, and it was clear that the magnetic cores were wrapped in agarose gel. Furthermore, the reaction time between the IMBs and the aflatoxins was investigated, and the optimal reaction time for meeting the purification requirements was determined to be 2 min. The stability of the IMBs was then evaluated under refrigerated storage conditions at 4 ℃. It was found that the prepared IMBs maintained a high aflatoxin enrichment capacity for at least eight months. Through the examination of three different extraction solutions, a mixture of acetonitrile and water (70∶30, v/v) was found to be optimal for the extraction of aflatoxins from the feed samples. Moreover, five sample dilutions and purification effects were also examined, and phosphate-buffered saline (containing 0.5% Tween-20) was selected as the preferred sample dilutant. With the optimized conditions, the effectiveness of using IMB for the purification of different feed samples was investigated. The resulting UPLC chromatogram showed no spurious peaks close to the target peaks, demonstrating a good purification performance. Following matrix spiking (5, 20, and 40 μg/kg, calculated based on AFB1) of the four feed samples (i. e., soybean meal, distillers dried grains with solubles, pig feed, and chicken feed), the spiked recoveries of the four aflatoxins ranged from 91.1% to 119.4% with a relative standard deviation (RSD) of <6.9%. In addition, the inter-day precision was 4.5% to 7.5%, and the method exhibited a good reproducibility. Subsequently, the developed method was used to detect AFB1 using reference materials. The test value was 18.6 μg/kg with an accuracy of 110.3%, thereby constituting satisfactory results. Upon testing 21 randomly purchased feed samples using this method, four of these samples contained AFB1, and the test results obtained using the developed method and stable isotope dilution LC-MS/MS were comparable. It was therefore apparent that the IMB purification method combined with UPLC analysis exhibited a good accuracy for aflatoxin determination. Thus, an automatic purification system was established to facilitate the operation and use of IMBs. This system was able to purify 24 samples simultaneously in 30 min. An IMB purification kit for was also designed and produced for aflatoxin detection in feed samples. The kit contained the sample dilutant, IMBs, the washing solution, and the eluent. After extraction of the feed sample, the extraction solution was added to the sample wells provided in the kit, and the purification system automatically completed the steps of aflatoxin enrichment, impurity washing, and elution of the target toxin. It should be noted that the purification process does not require the operator to manually add the solution, thereby simplifying operation. Overall, the purification method established in this study achieved the high-throughput and automatic purification of the four aflatoxins in feed samples.