It is also necessary to consider the impact damage, the loosening of components and the vibration of the structure caused by the impact load factors in the study of the vibration state of the equipment. In this paper, an impact load identification algorithm based on half cosine function is designed. The proper interval is determined by genetic algorithm, and the dimensions of beams, thin plates and trusses are determined by data method. The numerical simulation results show that the error of SCFF fitting method is lower than that of Tikhonov and Chebyshev orthogonal polynomial fitting (COPF), and the SCFF recognition advantage is more obvious with the increase of noise. The peak error of less than 10% is obtained, and the minimum value is reached under the parameter optimization, which indicates that the parameter optimization has good applicability. The test results show that the low-frequency vibration state of the cantilever beam caused by the impact load is inferred by analyzing the spectrum data of the response signal, and the correction of the model by the first four modes meets the feasibility requirements. When the SCFF method is used for identification, it can form a good agreement with the actual load and obtain a smaller peak error.

In order to solve the problems of incomplete balance equation, inaccurate balance equation and large error of cam contour caused by the influence of piston deflection on cam dynamic moment in existing cam construction of pedal brake unit, the structure of pedal brake cam and output force stability were studied. Firstly, the concept of optimum compensation angle is introduced, and the rotating moment of the thrust caused by piston deflection in two directions to the cam articulation point is included in the new balance equation, which makes the balance equation more complete, accurate and has less error. Based on this, a new cam mechanism is designed and constructed. Then, on the pedal brake unit routine tester, a comparative experiment is carried out on the stability of output force value between the original cam mechanism and the newly constructed cam mechanism with optimized compensation angle. Finally, the above data and the curves of action travel and output force are comprehensively compared and analyzed. The results show that the output force value stability of cam with compensation angle is better than that of cam without compensation angle. In actual operating stroke, the maximum difference of output force between two times is (-0.8, +0.6) kN with compensation angle cam error, and (-1.2, +1.2) kN without compensation angle cam error. The error with compensation angle cam is also less than that without compensation angle cam error.

In the railway container yard, there are few mature intelligent anti-lifting solutions available for train flatbed loading and unloading operations due to the poor detection accuracy or speed of traditional detection methods. This paper proposes a fast anti-lifting detection method for trains based on an improved back propagation (BP) neural network. By acquiring weight data from the four locks of the hoist, a flatbed lifting detection model is established using a BP neural network. During weight adjustment, a momentum factor and an adaptive learning rate are incorporated to optimize the model's performance. Through practical tests, this method demonstrates that this model achieves a high detection rate and fast detection speed, making it suitable for providing intelligent safety protection for automated rail mounted gantry in the railway container yard.

With the wide application of target detection technology in unmanned construction and other scenarios, traditional target detection algorithms face challenges such as low recognition accuracy, large computational volume, and slow processing speed in complex engineering environments. Based on these challenges, this paper proposes a target detection method based on improved YOLOv8 for engineering scenes, which improves the C2f structure by introducing the star block in YOLOv8, significantly reducing the number of model parameters and computational volume while ensuring the detection accuracy. Based on this, this paper introduces a lightweight shared and detail-enhanced convolutional detection head, which further improves the detail-capturing ability of the detection head and significantly reduces the computational burden. The experimental results show that compared with YOLOv8n on Roboflow-based engineering scene dataset, the mAP@0.5 and mAP@0.5: 0.95 of the improved model improves by 0.3% and 2.0%, while the number of parameters and computational volume decreases by 36.7% and 34.6%, and improves the frames per second (FPS) by 23.3% accordingly, which verifies the superiority of the improved algorithm in terms of lightweight and detection accuracy.

In order to explore the factors influencing the slip rate between tracked combine harvester and paddy soil, a coupled simulation model of tracked combine harvester and paddy soil is established by the RecurDyn software. Selecting the mass, driving speed, steering angular speed of the tracked combine harvester and the moisture content, positive slope, and oblique slope of the paddy soil as test factors, simulations under different conditions are implemented. Through a six-factor two-level PB (Plackett-Burman) test was conducted to screen out the slip rate significant factors that is soil moisture content, positive slope, and steering angle velocity. Then, a three-factor three-level BB (Box-Behnken) test was conducted on the three selected factors. The results show that soil moisture content, positive slope, and steering angle velocity have a very significant impact on slip rate, and the order of significance is steering angle velocity>soil moisture content>positive slope. As the positive slope and steering angle velocity increase, the slip rate increases significantly, while as the soil moisture content increases, the slip rate slowly decreases. The soil moisture content has a certain interaction with the positive slope and steering angular velocity. However, there is basically no interaction between positive slope and steering angular velocity. This study has important reference value and scientific significance for improving the path tracking accuracy and stability of tracked combine harvester for autonomous harvesting operation in paddy soil.