The rapid and accurate identification of feeding targets is a crucial guarantee for intelligent feeding assistance robots. Balancing segmentation accuracy and operational efficiency is a key aspect of ensuring the comprehensive performance of algorithms and a significant challenge for recognition methods. To address the issue of matching segmentation accuracy and efficiency in existing methods for identifying cow feeding targets, this paper proposes a real-time feeding target instance segmentation (RTFIS) based on Segment Anything Model (SAM) optimization. Building on the SAM-det architecture, the method introduces lightweight parameter designs for the image encoder and object detector, along with a parallelized buffer queue approach to balance the operational efficiency of each module, significantly improving inference speed. The use of HQ-token enhances feature space decoding capability, optimizes the design of the mask decoder, and employs a phased training strategy tailored to feeding targets, thereby improving segmentation accuracy. Research and experimental results show that the proposed method ensures segmentation efficiency while enhancing segmentation accuracy. In the task of cow feeding target recognition, the method achieves a segmentation accuracy of 98.7% for cows, 96.4% for feed, and a processing speed of 52.9 FPS, meeting the application requirements for cow feeding target recognition in complex environments.

Aiming at the problems of low classification accuracy and weak generalization ability of existing methods for acoustic scene classification, this paper proposed an acoustic scene classification method based on reducing high frequency reverberation and frequency domain residual shrinkage network for multi-scale attention(RF-DRSN-EMA). Firstly, the underlying principles for reducing sound reverberation were presented, along with a proposed algorithm specifically reducing high-frequency reverberation. This algorithm effectively attenuated only the high-frequency reverberation while preserving essential frequency information in other bands. As a result, speech intelligibility was enhanced, and the impact of speech distortion was minimized. Secondly, based on the deep residual shrinkage network and combined with the improved frequency domain self-calibration algorithm and the multi-scale attention module, a frequency domain residual shrinkage network for multi-scale attention(RF-DRSN-EMA) was proposed. The model used RF self calibration block, whose internal long-distance and short-distance residual structure can alleviate feature collapse. In order to achieve efficient collection of frequency domain information, multi-scale attention module was used in the output of the unit, which can further focus on the effective information of the unit at the output layer, thus strengthening the representation ability of the model. Finally, the experiments were carried out based on ESC-10, Urbansound8K and DCASE2020 Task 1A data datasets. The experimental results showed that the speech enhancement method to reduce high-frequency reverberation can reduce the impact of background noise such as high frequency reverberation and eliminate redundant features, and the sound quality damage is small, thereby showing a better classification performance. At the same time, RF-DRSN-EMA realizes the typical feature denoising and efficient information collection in the frequency domain of the network, and the best classification accuracy of the model can reach 98.00%, 93.42% and 72.80%, respectively, which verifies the effectiveness and generalization of network.

Aiming at the problems of high operational energy consumption and low efficiency of the air-conditioning system in the electric drive workshop, a dynamic energy-saving optimization method considering the mutual constraints of each device of the air-conditioning system is proposed based on the IBK-IPS algorithm. Firstly, the influence mechanism between each device of the air-conditioning system is analyzed, the mathematical model of energy consumption and constraints of each device is established, and the optimization objective function of system operation energy consumption is constructed. Then, a method based on the improved black-winged kite and particle swarm (IBK-IPS) algorithm is proposed to optimise the operating parameters such as water temperature, flow rate and air volume of each device of the air-conditioning system, in order to improve the precision and effect of the control of the operating parameters of the air-conditioning system. Secondly, the Simulink platform is used to establish the energy consumption simulation model of the cooling water system and chilled water system of the air-conditioning system, and simulation experiments are carried out to verify the effect and accuracy of the optimization of the operating parameters. Finally, the method is practically applied in an electric drive room to verify the practical effect and feasibility of the proposed method. The simulation and application verification results show that: 1) the operational energy consumption of the system is effectively reduced, and the energy saving rate reaches 11.23%~34.68%; 2) the operational energy efficiency of the system is effectively optimised, and the energy efficiency is improved by 11.53%~40.78%; 3) the energy saving of the IBK-IPS algorithm is better than that of the PS, BK, and BK-PS algorithms, and the algorithm's performance indexes improve 27.27%~40.78%, respectively, compared to the remaining three algorithms by 27.27%, 61.90%, and 69.23%; 4) In real application tests, the energy saving rate of the optimised system under five different loads is 22.62%, 17.24%, 16.94%, 14.97%, and 12.64%, respectively. In summary, the energy-saving optimization method proposed in the paper can effectively solve the problems of high energy consumption and low efficiency of the operation of the air-conditioning system of the electric drive workshop air-conditioning system, which has good energy-saving effect and practicability, and can provide new ideas for the research of energy-saving optimization of air-conditioning system.

Proton exchange membrane fuel cells (PEMFC) are highly valued in fields such as vehicles, ships, and aerospace due to their advantages of being pollution-free, highly efficient, and low-noise. However, reliability issues have hindered the large-scale commercial promotion of fuel cells. To further enhance the reliability of fuel cells, this paper proposes a fault prediction method based on deep learning. First, data preprocessing methods such as standardization are used to eliminate the influence of different dimensions among monitoring parameters. Feature parameters are selected based on mechanistic knowledge to reduce the dimensionality of the original data and improve the computational efficiency of the fault prediction model. Then, a time series prediction model based on Long Short-Term Memory (LSTM) networks is constructed to predict the future operating state of the fuel cell. Finally, the predicted state data is input into a fault identification model based on Convolutional Neural Networks (CNN) to achieve fuel cell fault prediction. The proposed method is validated using experimental fault data from fuel cells. The experimental results show that the proposed fault prediction model can predict faults 10 time steps in advance with an accuracy of 96.8%.

Studying the effect of Ozone (O₃) on polycyclic aromatic hydrocarbon (PAH) during the combustion process of biodiesel can provide new insights for reducing soot emissions. A skeletal reaction mechanism of biodiesel surrogates coupled with an O₃ reaction mechanism and a PAH reaction mechanism was constructed for modeling the effect and mechanism of O₃ on poly-aromatic hydrocarbon (PAH) formation in a counterflow flame of biodiesel surrogates. The final mechanism consists of 138 species and 608 reactions. The results indicate that the maximum concentration of PAH initially increasing and subsequently decreasing with the increase of initial O₃ volume fraction. When initial O₃ volume fraction increases to 4%, the maximum concentrations of major PAH such as benzene (A₁), naphthalene (A₂), anthracene (A₃), and phenanthrene (A₄) are 4.57, 6.76, 16.16, 12.38 times that at O₃ volume fraction of 0%, respectively. The addition of O₃ has a significant impact on the concentration of PAH, and has the greatest impact on A₃. At the same time, the pathway of benzene (A₁) generation changes, the main reaction for generation of A₁ changes from C₂H₂ ＋ C₄H₅ ＝ A₁ ＋ H to C₂H₃ ＋ C₄H₄ ＝ A₁ ＋ H. And When initial O₃ volume fraction increases to 12%, they respectively 0.88, 0.357, 0.375, 0.143 times that at O₃ volume fraction of 0%. It is because that the C₂H₃ radicals are oxidized, thereby inhibiting the production of A₁.

Accurately predicting the NOx emission concentration at the outlet of the selective catalytic reduction (SCR) denitrification system in the waste incineration process is of great significance for enhancing data quality and optimizing ammonia injection. However, the waste incineration process exhibits significant nonlinearity, multivariate coupling, and time-series characteristics. These factors pose substantial challenges to achieving accurate predicting of NOx emissions. This paper presents a prediction model for NOx emission concentration at the outlet of the SCR denitrification system by integrating maximum information coefficient (MIC), principal component analysis (PCA), and long short-term memory (LSTM) neural networks. First, MIC is employed to assess the nonlinear correlations among variables, selecting input variables that exhibit the strongest correlation with NOx emission concentration while eliminating redundant variables based on the principle of maximum redundancy. Then, PCA is utilized to address the coupling characteristics among variables and to reduce information redundancy. Finally, a predicting model of NOx emission at the outlet of the SCR denitrification system is developed based on the LSTM model. The results indicate that the proposed model exhibits high accuracy and generalization ability, achieving an average absolute percentage error of 6.33%, a root mean square error of 4.71mg/m3, and 0.90 of R2. It outperforms both the back propagation neural network (BPNN) model and the support vector machine (SVM) model, thereby laying a foundation for achieving intelligent control of the SCR denitrification system in the waste incineration process.

Near-infrared optical tracking systems can restore the movement of tracked objects in real time based on the reflective marker balls attached to the tracked objects. They have been widely used in many fields. This paper proposes a real-time tracking method for multiple near-infrared targets that is robust to target loss. First, according to the imaging characteristics of reflective marker balls in near-infrared cameras, the geometric center of each reflective marker ball is extracted using the grayscale centroid method. Then, the SORT algorithm is used as a multi-target tracking method in each monocular camera to match each marker point between frames. The matching relationship of the image points of the reflective marker balls in each camera is determined based on the principle of epipolar geometry combined with the weighted bipartite graph matching method, and the three-dimensional spatial coordinates of each tracked reflective marker ball are calculated in real time based on the triangulation method. Secondly, the reflective marker balls are grouped according to the spatial position relationship between the reflective marker balls during the movement process, and the reflective marker balls belonging to the same object are identified. The appearance feature vector of the tracked object and the reflective marker ball is established based on the Euclidean distance between the reflective marker balls in the same group as the matching basis for the object loss and reappearance. The tracked object that is completely lost and then reappears is rematched using the cosine distance of the appearance feature vector. Finally, the proposed algorithm is experimentally verified. The experiment shows that the tracking accuracy of the proposed algorithm can reach 0.5mm at a speed of not less than 60 frames. In addition, the lost reproduced objects and reflective marker balls can be correctly re-matched.

Gear is an important component in mechanical transmission, and its fatigue performance affects the reliability and safety of mechanical equipment. However, in the current research on contact fatigue problems, the joint mechanism of residual stresses and inhomogeneities is unknown, and the effect on contact fatigue life is difficult to predict. In this paper, a numerical algorithm model is proposed to analyze the distribution of common residual stresses and inhomogeneities in gears. To reach this goal, the classical distributions of the residual stress in the gears are obtained from references. Then the gear loading tooth contact analysis of straight tooth cylindrical gears is carried out to find out the contact load at a series of key contact positions and the size of the equivalent contact radius. Finally, a microscopic contact calculation model combined with the residual stresses and the inhomogeneities is established. In the establishment of the model, the inhomogeneities inside the gear are transformed into inclusions containing eigenstrains using the equivalent inclusion method, and the coupling effect of the residual stresses and inhomogeneities is considered in the equivalent equilibrium equations. The stress distribution under the joint action of different residual stress fields and inhomogeneities is computed by the presented model. On this basis, a gear risk assessment model is established. The equivalent stress is calculated using Dang Van's criterion, which is then incorporated into the Lundberg-Palmgren life model to find the minimum number of cycles. In this way, the risk assessment of contact fatigue inside the gear is completed, and the effects of different meshing positions, residual stresses and inhomogeneities on the contact fatigue of the gear are analyzed. The model was verified by the finite element method, and the results agree well. The impact of inhomogeneities on the contact fatigue life of gears far exceeds the influence of residual stress, and they predominantly determine the earliest point of contact fatigue on the gears. Under the combined effects of residual tensile stress and inhomogeneities, the maximum stress value in the sub-surface layer increases and shifts towards the material surface, making gears more prone to contact fatigue.

The multi-stage flow channel magnetorheological damper (MFC-MRD) in the hydraulic actuator can effectively improve the underdamping characteristics of the hydraulic valve-controlled cylinder system and improve the stability of the system. It has potential application value in hydraulic equipment such as legged robots and excavators. However, the high power density design of MFC-MRD will inevitably increase the effective length of the damping channel, thereby increasing the damping force of the system, resulting in the existing dynamic model can not accurately describe the nonlinear hysteresis characteristics of MFC-MRD. In order to improve its nonlinear hysteresis characteristics, based on the mechanical performance test, and by analyzing the multi-stage flow channel structure and nonlinear hysteresis curve characteristics, the hyperbolic tangent curve is segmented and reorganized, and then an improved hyperbolic tangent model conforming to the hysteresis characteristics is proposed. Secondly, in the process of parameter identification, in order to avoid the problem of model parameters falling into local optimum and ' premature ', the selection operator of genetic algorithm is improved, and then a three-level stepwise approximation selection operator is proposed to improve the identification accuracy of model parameters. According to the mechanical experimental data, the relationship between model parameters and current is accurately obtained. Finally, through the model comparison, it can be seen that compared with the Bouc-Wen model and the data-driven model, the accuracy of the improved hyperbolic tangent model established in this paper is increased by up to 75%. It can accurately describe the nonlinear hysteresis characteristics of MFC-MRD, and verify the superiority and accuracy of the improved model.

To achieve rapid post-earthquake restoration of structures, a chevron braced steel frame structure with replaceable energy dissipation joints is proposed with using the design concept of controllable damage and replaceable energy dissipation joints. To explore the seismic performance of the structure and post-earthquake repair performance of structure, quasi-static tests were conducted on a 1/2 scale single story single span substructure specimen and the repaired specimen. The hysteretic curves, skeleton curves, stress curves and ductility of the specimens were studied an compared in the initial and repaired specimen. The results showed that the spindle shaped hysteretic curves is quite plump. The damage of this structure is mainly concentrated on replaceable energy dissipation joints, and the main frame is basically remained in elasticity. After the 0.83% inter-storey drift, the loading is completed. The residual inter-storey drift ratio of the structure is 0.28%, and the mechanical properties of the structure after replacing the energy dissipation joints are similar to original structures. The simplified analytical modes of chevron braced steel frames with replaceable energy dissipation joints was established, the calculation method of the elastic lateral stiffness of the structure was derived based on the deformation coordination relationship and the calculation formula of the bearing capacity of the structure when the double U-shaped metal energy dissipator yields was proposed. The maximum error between the theoretical calculation result of the elastic lateral stiffness of the structure and the experimental result was 3.72%, and the maximum error between the theoretical calculation result and the experimental result when the double U-shaped metal energy dissipator yields was 9.41%.

This paper aims to develop a theoretical model for thin cylindrical shells made of composite materials with thermal expansion and hygroscopic expansion coefficients. The analytical expression of strain energy under complex environmental conditions is derived. Based on the principle of minimum potential energy, the effects of environmental parameters on the strain energy, twist curvature, and secondary stable principal curvature of T700/Epoxy, T300/5028 Graphite-Epoxy, and AS7/M21 carbon fiber/epoxy resin-based composite cylindrical shells are studied. A finite element model of the cylindrical shell structure is established to numerically simulate the bistable deformation process of the shell, and the numerical results are compared with the theoretical results. The results show that the twist curvature of T700/Epoxy and AS7/M21 is relatively stable within the temperature range of 20°C to 120°C and humidity range of 0% to 1%, with maximum strain energy reductions of 32.7% and 9.1% respectively. The strain energy increment of T300/5028 Graphite-Epoxy cylindrical shells is as high as 914.6%, indicating that high temperature and high humidity have a significant impact on the bistable performance of the structure. Through quantitative analysis of the mechanical properties of composite materials under different temperature and humidity conditions, this study provides a scientific basis for the selection of materials and optimization of the application environment for bistable structures, contributing to the improvement of structural design reliability and durability.

As an extreme case of adjacent construction of underground engineering, closely-attached underpass will cause significant disturbance deformation of the existing structure and directly affect the normal service performance of the structure. Aiming at the freezing excavation underpassing the operating subway station, the structure-stratum dynamic finite element model was established by means of joint simulation and verified by field test. By transforming the existing line irregularity caused by the closely-attached underpass construction into the wheel-rail dynamic excitation and applying it to the simulation model, the dynamic transmission characteristics between the track and the tunnel were calculated and analyzed, and then the evolution process of the station vibration source intensity in the key construction stage was simulated, which can provide theoretical guidance for similar projects. The results show that due to the dynamic transmission distance, the vibration level of the tunnel side wall is lower than that of the track bed, but the response amplitudes of the two are increased in different degrees in the stage of the line irregularity formed by the closely-attached excavation. The vibration source intensity of the subway station is dominated by the dynamic response below 200Hz, in which the main influence frequency band of the irregularity caused by construction is within 40Hz, and the source intensity is positively correlated with the irregularity amplitude between 8-40Hz, while the vibration of the tunnel side wall is amplified within 8Hz.

The penetration of automated vehicles (AVs) will gradually increase in the future, and the consequent method of adding Dedicated lanes for AVs to existing roads has become an effective countermeasure to improve roadway efficiency and driving safety. Although the horizontal and vertical alignment are not easy to change due to the constraints of manned vehicles, the width of Dedicated lanes for AVs can be redesigned and optimized. However, there is a lack of standards and calculation basis for the design of self-driving lanes. Autonomous vehicle trajectory offset is an important basis for the design of the width of lanes dedicated to AVs. This study conducts corresponding research on the complex horizontal alignment combination design, which significantly affects the driving trajectory. The typical lateral motion control algorithm and the longitudinal motion control algorithm of AVs, as applied to the PreScan-Simulink simulation platform, are used to consider three complex horizontal alignment combination designs: the oval curve, the convex curve, and the C-shape curve. The objective was to construct simulation vehicle models of car and truck models and road scenarios, and to obtain a characterization of the influence law of horizontal alignment combination design on AV trajectory offset. Additionally, trajectory offset models of car and truck were constructed. This study shows that the feature point with the largest offset for the AV is HY1 on the oval curve and the C-shape curve, but on the convex curve the feature point with the largest offset is HH. The design speed is significantly correlated with the trajectory offsets of AV on each horizontal alignment combination design: the offsets of the feature points with the largest offsets on each design are about [9cm, 16cm] for AV at 60-130 km·h^-1; the magnitude of the trajectory offsets varies greatly with the change in design speed, and the offsets of the feature points with the largest offsets on each horizontal alignment combination design are about [13cm, 23cm] for AV at 140-150 km·h^-1. The correlation model between design speed and trajectory offset is a quadratic polynomial regression model, and the R² of the model is greater than 0.95, so the model fit meets the prediction requirements. The research method and research results of this thesis can provide a reference basis for the calculation of dedicated lane width.

When a fire occurs in a high-speed train compartment, this can lead to the rupture of windows, thus forming side openings, which significantly affects the fire combustion state and temperature distribution. Therefore, a combination of 1:8 scaled model experiment and numerical simulation is used to study the effect of different opening positions on the fire evolution in the compartment, and the longitudinal decay law of the temperature inside the high-speed train compartment is quantitatively investigated under the joint influence of the opening position and the power of the fire source. The results show that: with the increase of the fire source power at each opening position, the fire evolution inside the compartment and the smoke/flame at the opening go through the stage of full combustion of the fuel inside the compartment, the stage of anoxic combustion, and the stage of continuous overflow in turn; the influence of the fire source power and the opening position on the flame moving speed inside the compartment is discussed, and a prediction formula for the flame moving speed is put forward; the maximum temperature inside the compartment and the temperatures at the left and right sides of the opening are investigated, and the longitudinal attenuation law for the temperature inside the high-speed train is established. The maximum temperature in the compartment and the temperature decay law of the left and right sides of the opening are studied, and the temperature decay prediction model of the left and right sides of the opening is established when different openings of the high-speed train compartment fire, and the results of the research have certain reference value for the prevention and mitigation of the high-speed train compartment fire.