The plasma samples of Alzheimer disease（AD） patients are collected using Fourier transform infrared-attenuated total reflection （FTIR-ATR） spectroscopy technology. Based on the FTIR-ATR spectral data of the plasma membrane samples， the spectral data are encoded into two-dimensional images by utilizing the Gram angular field （GAF） and Markov transition field （MTF）. Meanwhile， a neural network model based on the deep residual networks and attention mechanism is combined to conduct the screening and classification research on Alzheimer disease. The experimental results show that the GAF-MTF-CNN model can effectively improve the accuracy of spectral feature extraction. Additionally， the method of combining two-dimensional data with deep learning has better classification accuracy compared with traditional classification methods. Encoding spectrum into images using GAF and MTF techniques， and combining them with an improved residual neural network， effectively enhances the generalization ability and diagnostic accuracy of AD screening models， optimizing the screening performance.

In response to the problem of low detection precision caused by inconsistent size， high background noise， and large-scale changes in X-ray image prohibited item， the optimization is performed based on RT-DETR-R18 and an X-ray image prohibited item detection algorithm named X-ray-RTDETR is proposed. Firstly， the algorithm employs CSPRepResNet embedded with efficient multi-scale attention as the backbone network to enhance feature extraction capabilities. Secondly， the simplified fast spatial pyramid pooling module is introduced after the three features maps output by the backbone network to improve the robustness and generalization ability of the model. Finally， the SPoolFormer encoder is applied to high-level feature maps with richer semantic concepts for intra-scale feature interaction. The experimental results show that the detection accuracy of X-ray-RTDETR achieves 74.6% on PIDray test set， surpassing RT-DETR-R18 by 8.5%， while reducing the number of parameters and n_FLOP by 1.67×10⁶ and 2.24×10⁹， respectively. Compared to the state-of-the-art object detection algorithms at the same scale shows that X-ray-RTDETR not only has higher detection accuracy， but also has less number of parameters and n_FLOP. At the same time， its inference speed reaches 85.47 frames per second on RTX2070 Max-Q GPU.

Aiming at the secondary control issues in islanded DC microgrids， a distributed secondary control strategy based on a dynamic event-triggered mechanism is proposed for DC microgrid containing multiple distributed generators. Building upon distributed secondary control， dynamic event-triggered conditions are designed for both the output average voltage and proportional current of each DG. This approach not only mitigates deviations between DC bus voltage and the nominal value caused by droop gains but also maintains current sharing accuracy while significantly conserving communication resources and reducing redundancy. The designed event-triggered conditions incorporate a dynamic term that adaptively adjusts event-triggered thresholds， effectively decreasing controller update frequency. The feasibility of this event-triggered mechanism is rigorously proven through Lyapunov stability theory， with the absence of Zeno behavior being guaranteed. Finally， a simulation model developed in MATLAB/Simulink validates the effectiveness of the proposed control strategy.

In order to obtain high-performance microwave frequency combs， an all-optical scheme for generating a tunable microwave frequency comb using three lasers in cascading injection is proposed. This scheme does not require external modulation from a microwave signal source， making it a fully optical generation device. The experimental setup and working principle are introduced. The impacts of various factors， namely the pump current， injected optical power， feedback optical power， and central frequency detuning between two lasers， on the number of comb lines， bandwidth， peak maximum power， and frequency spacing of the generated microwave frequency comb signals are analyzed by experiments. The results indicate that when the power attenuation values of VOA1， VOA2， and VOA3 are 13， 13 and 21.5 dB， respectively， the pump current of SL is 56.1 mA， and the frequency detuning between ML and IL is 70.625 GHz， a microwave frequency comb signal with a bandwidth of 11.87 GHz can be obtained.

When the integral method is used to solve the direct exchange area （DEA） of radiation in the zone method， the relative error is large due to the existence of “singularities”， and it usually requires a lot of computing time and has little effect to improve the calculation accuracy at “singularities”. Therefore， based on Duffy transformation method， the Duffy transformation integral formula is derived to eliminate the influence of the singularity on the calculation of DEA. The improved method and the direct Gaussian integral （DGI） method are used to solve the DEA of radiation in a two-dimensional square cavity. The results show that the completeness verification errors of the gas-zone and the surface-zone can be decreased from 3.73% and 6.70% to （4.88×10^-4）% and （2.98×10^-5）% respectively in the same calculation time by applying Duffy transform. In the case of the same calculation accuracy， the maximum difference in calculation time between them is 13 693 times. Therefore， the improved algorithm has significant advantages in computation speed and accuracy.

The process of treating spent dry barrier from aluminum electrolysis with alkali and acid is proposed. Spent dry barrier sample from an aluminum company was taken as the raw material， and its main phases are NaF， NaAlSiO₄， CaF₂， α-Al₂O₃， Na₃AlF₆， Si and β-Al₂O₃. Firstly， the NaF with mass fraction of 98.84% is obtained by water leaching separation of the raw materials. Secondly， the optimal process condition for alkali leaching treatment of water leaching residue were obtained by single factor method as follows： temperature of 90 ℃， liquid-solid ratio of 5 cm³/g， alkali-slag mass ratio of 0.25， and reaction time of 100 min， and under this condition， the leaching rate of Na₃AlF₆ was 94.04%. After that， the optimal process conditions for acid leaching treatment of alkali leaching residue were obtained by single factor method as follows： temperature of 58 ℃， reaction time of 45 min， acid concentration of 0.6 mol/L， and liquid-solid ratio of 12 cm³/g， and under this condition， NaAlSiO₄ and CaF₂ can be leached. Subsequently， the alkali leaching solution was dropped into the acid leaching solution to remove silicon in the acid leaching solution. It was found that when the pH was about 3， the precipitation effect of silica sol in the acid leaching solution was the best， and the product obtained by heating the silica sol which was obtained through filtering was mainly SiO₂·0.2Al₂O₃， with a mass fraction of 97.20%. The acid leaching solution after filtering to remove the silica sol was added to the alkali leaching solution to precipitate and recover the fluoride salts， and it was found that when the pH was 9， the recovery rates of CaF₂ and Na₃AlF₆ were the highest， which were 95.91% and 92.44%， respectively， and the mass fractions were 48.41% and 25.14%， respectively. In addition， the precipitate contains 25.32% （mass fraction） of Al（OH）_3.

The preparation of magnesium-aluminum layered double hydroxides （MgAl-LDHs） from industrial bischofite by-products in salt lake region was achieved via electrochemical deposition method， with systematic investigation of deposition parameters. The structural and properties of the products under different conditions were investigated utilizing X-ray diffraction （XRD）， scanning electron microscopy （SEM）， Fourier transform infrared spectroscopy （FT-IR）， and other analytical methods. The results show that MgAl-LDHs materials with regular morphology can be prepared by electrochemical deposition method. The product under the condition of large current density will undergo severe agglomeration. Under lower electrolyte concentration conditions， the interlayer spacing of products increases， which is beneficial for crystal growth. Excessive Mg and Al mass ratio is not conducive to product formation. The preparation process of products involved layer-by-layer detachment near electrode plates， where interfacial supersaturation influenced crystallinity.

In order to solve the problems of traditional sinter quality prediction model， such as using single feature selection method and having no background of process mechanism， which results in low model prediction accuracy and lack of interpretability， a GA-BiLSTM prediction model with feature optimization is proposed. First， the optimal feature set is selected through various feature selection methods and combined with the sintering process mechanism， then GA is used to optimize BiLSTM， and finally the optimal feature set is used as the input of the GA-BiLSTM model to predict the FeO content in sinter. The GA-BiLSTM model with feature optimization was compared with other models. The results show that the prediction error of the established model is low， and the prediction accuracy for FeO mass fraction in sinter is as high as 94% within the allowable error range of ±0.5%， which may provide a new guiding direction for improving the quality of sinter.

In order to improve the removal depth consistency of nickel-cobalt functional gradient materials （NiCo-FGM）， an adaptive grinding force control system was constructed to carry out constant force zonal grinding experiments on five types of NiCo-FGM with different mass fractions of IN718 to investigate the trend and extent of the influence of the process parameters on the removal depth and surface roughness of the materials. The feasibility of transfer learning was then analyzed and the accuracy of the removal depth modelling was compared with that of empirical formulas. Finally， comparing the removal depth prediction results of constant force and variable force grinding. The results showed that the normal force has the most significant effect on the removal depth and surface roughness of the materials. The average error in the prediction of transfer learning is reduced by 4.07%， and the efficiency is higher. The maximum difference in removal depth between the remaining content of IN718 and 50%IN718 under constant force grinding is 8.955 μm， and the maximum removal depth difference between 100%IN718 and 0%IN718 is 15.619 μm， whereas the removal depth consistency can be improved by variable force grinding.

Through experimental research on the influencing factors of mechanical properties of FeCoNiCrAl_0.5 and FeCoNiCrTi_0.5 high-entropy alloy formed specimens prepared by the selective laser melting method， the changes of density and hardness of the high-entropy alloy with laser power， scanning speed and scanning spacing， as well as the effects of powder type and element type on the tensile properties were investigated， and the tensile fracture morphology of the high-entropy alloy specimens was analyzed. The experimental results showed that the density and hardness of the specimens increase at first and then decrease with the increase of laser power， scanning speed and scanning spacing. Compared with the mixed powder， the tensile properties of FeCoNiCrAl_0.5 specimen prepared by alloyed powder are better， FeCoNiCrTi_0.5 high-entropy alloy has higher tensile strength and stronger toughness than FeCoNiCrAl_0.5 high-entropy alloy. The tensile fracture of FeCoNiCrAl_0.5 formed specimen prepared by mixed powder and alloyed powder is mostly brittle and ductile fracture， while the FeCoNiCrTi_0.5 formed specimen prepared by mixed powder are mainly plastic fracture.

In order to suppress the possible resonance of the servo drive system when flexible manipulators move， and reduce the speed fluctuation of the motor output end， a method of designing PI controllers by the pole assignment method is proposed. Based on the assumed mode method and Lagrange dynamic equation， a dynamic model of the flexible manipulator driven by the servo motor is established. The transfer function of the system is obtained by using its state equation， and the transfer characteristics of the servo drive system are analyzed. The parameters of the PI controller are designed by the pole assignment method with the same real part， and the influence of pole natural frequency ratio and damping coefficient on the system evaluation index is discussed. Finally， the simulation experiment is carried out. The results show that the appropriate increase of damping coefficient is beneficial to the stability of the system， but the length and rotational inertia of the flexible load should not be too large. Compared with the traditional Ziegler-Nichols method， the effectiveness of the proposed method is proved.

A feature extraction and trajectory planning strategy for planar fillet welds based on 3D point cloud was proposed to solve the automatic identification of weld seams and automatic robot tracking welding. Firstly， the workpiece to be welded was extracted based on the difference point cloud segmentation method， and the point cloud pre-processing was performed. Secondly， in order to obtain the feature points of the weld seam， the workpiece structure segmentation feature extraction algorithm was proposed. Then a path fitting method based on NURBS curves was fitted. Finally， a robot position estimation method for welding points was proposed to obtain the position of each path point for welding. This strategy is applicable to the weld seams of straight lines and various planar curves. The experimental results showed that the strategy can accurately extract the position of the fillet weld seam and generate the required position of track points， with the maximum error of each axis controlled within 1 mm and the total time consumed no more than 18 s， which provides a valuable reference for efficient automated welding.

To address the challenges faced by conventional steel-sheet pile cofferdams when driving into hard soil layers and the poor water-stopping effect of interlocking steel-tubular piles， a prefabricated steel-tubular and steel-sheet pile cofferdam was proposed， which fully utilized the advantages of high stiffness of steel-tubular piles and good water-stopping effect of steel-sheet piles. The numerical analysis of the structure using Abaqus shows that the steel-tubular piles at the corners greatly improve the peeling of the steel-sheet pile cofferdam corners and the high dependence on support. Increasing the number of steel-tubular piles can greatly improve the structural stress condition. Three supports， especially the bottom sealing concrete， are the most effective in suppressing the development of steel-sheet pile deformation. At the same time， this structure can effectively suppress the deformation of the riverbed soil. The maximum horizontal and vertical deformation of the soil occur mainly during the initial pumping stage and the dredging stage， accounting for 89.9% and 65.2% of the maximum deformation， respectively. The three supports and bottom sealing concrete of the structure can effectively suppress deformation during the pumping and dredging stages.

Mixed matrix membranes （MMMs） have an extensive application prospect in the field of CO₂ separation with good processing property and high efficiency for separation. To prepare mixed matrix membranes， ZIF-8 was prepared through solvothermal methods. ZIF-8 and ［Bmim］［PF₆］ were used as the dispersed phase， and polyvinylidene fluoride （PVDF） was served as a continuous phase. The physical and chemical properties of MMMs were characterized by BET， SEM， TEM， FT-IR， TG and mechanical property testing. The effects of ZIF-8 content and transmembrane pressure on the CO₂ separation performance of the membranes were studied， and the long-time stability of membranes was investigated. The results show that the best CO₂ permeability of MMMs is 679.26×10^-7 cm³·cm·cm^-2·s^-1·MPa^-1 with the CO₂/N₂ selectivity of 40.33 when mass fraction of ZIF-8 loaded on the PVDF is 15%， and it shows excellent CO₂ separation performance.

Foam concrete severs as a deformation reserve layer between the initial support and the secondary lining， effectively resisting the rheological deformation of the surrounding rock in deeply buried tunnels with high-ground stress. The incorporation of fibers enhances its compressive performance and ductility， addressing the issue of low compressive strength under specific fiber characteristics. Orthogonal test and Box-Behnken design method of response surface were used to systematically investigate the effects of various characteristics such as fiber mass fraction， length， and types， on the compressive strength and elastic modulus of foam concrete. A characteristic regression model was developed to optimize the mix ratio. The results show that the regression model established by the response surface method demonstrates high accuracy and reliability. Among the various fiber characteristics， the fiber mass fraction has the greatest impact on both the compressive strength and elastic modulus of foam concrete. Meanwhile， the interaction among multiple characteristics significantly influences compressive strength while slightly impacts elastic modulus. Particularly， the interaction between fiber length and fiber type has the most obvious impact on compressive strength. Furthermore， by maximizing the compressive strength and minimizing the elastic modulus， the optimized mix ratio result derived from the model shows that the absolute values of the relative errors are less than 5%. The small relative errors indicate that the proposed model can provide support for the multi-objective optimization of foam concrete with different fiber characteristics in the project application.

In mineral identification， there may sometimes be cases of missed identification or incorrect identification when identifying associated minerals. To address these issues， the intelligent mineral identification methods under the microscope were developed. Firstly， an intelligent identification method was constructed by improving the SSD network and incorporating image transformation. Secondly， the method was applied to microscope images of minerals in polished sections from the iron ore in Liaoning Province， China， and its accuracy was demonstrated through tests. Finally， the effects of learning rate and batch size on the loss function were determined， and the accuracy was further improved by using the gradient descent method. In the tests， the identification accuracy exceeds 90%， reaching up to 100%， with the minimum value of the loss function was approximately 0.008. The results indicate that the proposed method has strong mineral identification capabilities.

The flotation separation behaviors of ilmenite and titanaugite with benzohydroxamic acid （BHA）， salicylhydroxamic acid （SHA）， octylhydroxamic acid （OHA） and butoxylate acid （BFXA） were studied by mineral flotation test. The interaction mechanisms of four hydroxamic acids with different structures on ilmenite and titanaugite surfaces were analyzed by these means of contact angle， Zeta potential， Fourier transform infrared spectra and density functional theory （DFT）. The results indicate that the separation effect for ilmenite with four hydroxamic acids is listed in the order： BFXA>OHA>SHA>BHA. The adsorption of hydroxamic acids on the surface of ilmenite is stronger than that of titanaugite， and the hydrophobicity of ilmenite is significantly improved by adding hydroxamic acid. The two oxygen atoms within oxyoxime groups of hydroxamic acid collectors are their reaction sites. Compared with the other three hydroxamic acids， BFXA has the highest HOMO orbital energy and thus exhibits stronger adsorption performance and collection capability.

The laminar premixed flame propagation characteristics of hydrogen-enriched gas with different hydrogen mixing ratios in an O₂/CO₂ atmosphere were systematically studied using a Bunsen burner-schlieren experimental system and the Premix model. The results show that GRI3.0 mechanism can provide accurate predictions of the laminar flame propagation velocity of mixtures under high hydrogen mixing ratios and high O₂ concentrations atmosphere. With the increase of hydrogen mixing ratios， the laminar flame propagation velocity of mixtures increase significantly at O₂/CO₂ atmosphere. The peak values of laminar flame propagation velocity slightly shift to the fuel-rich side. Meanwhile， the adiabatic flame temperature and laminar flame propagation velocity of mixtures gradually increase as the O₂ concentration increases. With the increase of O₂ concentrations， the large number of H radicals enhance the dominant promotion of H+O₂=O+OH， which contributes to the increase of the laminar flame propagation velocity of mixtures.