Addressing the issue of the unstable output and low accuracy issues of single-rod hydraulic cylinders under heavy load, high stiffness, and dynamic disturbances, a nonlinear model of the hydraulic system is established based on the analysis of force-bearing process and structural characteristics. Then an adaptive control method based on an asymmetric barrier Lyapunov function is proposed to performance steady tracking under output force constraints. The controller integrates adaptive parameter, extended state observer, and dynamic surface control and handles the system's parameter uncertainties, unknown states estimation and time-varying disturbances, and complexity explosion caused by high-order derivatives. The output force boundaries are constrained by the constructed asymmetric barrier Lyapunov function. Lyapunov-based analysis proves the system's asymptotic stability. Co-simulation verifys control effectiveness. The results show that the proposed method can accurately estimate and compensate for uncertainties, ensure the output remains within safe boundaries during loading while achieving high-precision actuator's position tracking.

A finite time prescribed performance neural network control strategy is proposed to address the need for control strategies that balance both transient and steady-state performance in the electro-hydrostatic actuator (EHA). A nonlinear mathematical model of the EHA is established, and a barrier Lyapunov function is constructed by incorporating a finite time prescribed performance function for the tracking error. Based on the backstepping control framework, a neural network-based position tracking controller is designed. The stability and theoretical performance of the controller are rigorously proven using Lyapunov analysis. A co-simulation model is built using MATLAB and AMESim, and comparative simulations are conducted with a PI controller and a neural network controller without prescribed performance. The results demonstrate that the proposed controller achieves significantly higher tracking accuracy. Compared to the PI controller and the neural network controller without prescribed performance, it improves sinusoidal trajectory tracking accuracy by 85% and 47%, and point-to-point trajectory tracking accuracy by 85% and 55.9%, respectively. Furthermore, the tracking error converges below the predefined steady-state bound within a finite time and remains within the prescribed performance constraints throughout the operation.

The gas volume regulation of compressor is greatly affected by complex working conditions, and single fault identification cannot guarantee the accuracy of abnormal identification. In order to further improve the precision of compressor gas volume control, a self-healing control method of compressor gas volume regulation oriented to fault anomaly recognition is designed. Based on the analysis of the electro-hydraulic actuator system, the self-healing control strategy of gas volume regulation is given. A stepless gas coupling control model is established to eliminate the fault problem by actively controlling the load, and to ensure that the gas regulating system can complete the goal of online self-healing under the unstable condition. The results show that the error of network training results is less than 2% after adding two levels of pressure, and the ideal prediction target is reached. The self-healing control curve of the system is in good agreement with the simulation results, and the simulation process meets the requirements. The first-stage actuator is in a relatively constant vibration phase, and the stable load can be obtained by judging the first-stage adjustment process. The research can be applied to other fields of mechanical transmission fault diagnosis and has promotional value.

In grain storage management, the curved surface structure of silo walls imposes stringent demands on the adhesion performance of wall-climbing robots. These robots must possess sufficient flexibility to adapt to curved surfaces while maintaining adequate rigidity to ensure stable support. We find that when a rigid suction cup is employed on walls with varying curvature, the limited deformation capacity of the suction cup body causes the sponge to adopt a “saddle-shaped” deformation during adhesion, which significantly diminishes the adhesion performance. To address this issue, we optimize the rigid suction cup structure, leading to the design of a semi-rigid suction cup with distributed rigid elements. The adhesion force experiments on various simulated substrates reveal that the rigid suction cup exhibit forces of 205.49 N, 307.56 N and 360.25 N on simulated substrates with curvature radii of 100 mm, 200 mm and 400 mm, respectively. In contrast, the semi-rigid suction cup exhibit slight fluctuations in adhesion force across different curved surfaces, yet maintain a stable overall value around 420 N. This solution achieves a significant enhancement in adhesion performance on complex curved surfaces, effectively reducing the risk of detachment during robot operation, which establishes a reliable foundation for expanding the application of wall-climbing robots in areas such as silos.

The openly-reported researches on mechatronic safety valves in aerospace applications show that the failure of solenoid valves will cause functional incapacitation of the safety valves. In view of this situation, we propose a valve terminal controlled mechatronic safety valve with the advantage of fault redundancy function, as well as the control strategy. The principle verification experimental investigations are carried out. The valve terminal is composed of three normally open two-position three-port solenoid valves. It is shown that by applying the valve terminal to control the switch of the passage between its back-pressure chamber and the pressure vessel or the atmosphere, combining with the proposed strategy, the main valve can operate normally under normal working mode and failure modes that DCF1 or DCF2 does not operate when power is on or off. The principle verification experimental results show that the control function of the solenoid valves by the controller, the action function of the solenoid valves and the opening pressure of the safety valve all meet the design requirements. The research results provide theoretical guidance and engineering application foundation for the subsequent application of the mechatronic safety valve in the pressurize transportation system of launch vehicle.

The problem of energy efficiency and thermal stability of underground coal mine explosion-proof vehicles under long-distance and large slope conditions has seriously restricted its unmanned development. Based on the chassis by-wire architecture, this thesis explores the thermal management strategy of unmanned development explosion-proof vehicles. A transient thermodynamic modeling method of multi physical field coupling is proposed, and the series hydraulic hybrid power system architecture based on closed loop is constructed. The architecture integrates the accumulator cooperative control and bidirectional energy conversion mechanism to improve the braking energy recovery efficiency and peak power output capacity. At the same time, a dynamic temperature boundary model based on oil flow direction identification is established to effectively overcome the time lag problem of the traditional model in transient response. Through the hardware in the loop experiment, the nonlinear relationship between the hydraulic oil temperature rise process and the system efficiency is revealed. The final test results show that the temperature prediction error of all key components such as hydraulic pump, motor and accumulator does not exceed ±3.5 ℃, and the model shows high accuracy in the prediction of system temperature level and change trend. This thesis provides a theoretical model and solution for the thermal management optimization of underground trackless auxiliary transportation equipment.

We design a purely soft-structured gripper and analyze the bending performance of its finger section. The gripper features a series of gas-driven elliptical cavities. A stepped cavity structure incorporates three parallel air channels for the fingertip, middle finger segment and finger root, respectively. The palm adopts an arc-triangle shape. The gripper is made of hyperelastic material Ecoflex 00-30 silicone rubber and polydimethylsiloxane, and its model uses the Yeoh constitutive model derived from uniaxial tensile theory. Building on this, we develop models for a single airbag, a single cavity group's bending and the bending deformation prediction of the entire multi-channel fully flexible gripper. Inputting the finger's structural parameters into the mathematical model yields the geometric relationship of its bending curve. The comparison of these analytical results with simulation data verifies the model's accuracy and practicality. Finally, analyzing the finger's bending performance allows us to determine optimal gripper performance parameters. Key factors include cavity structure type/number, finger width, cavity gap length and the bottom strain-limiting layer. This provides valuable research data and a reference for soft actuator development.

Addressing the issues of inadequate resistance to eccentric loads, low synchronization accuracy, limited control points and complex operation in derrick rectification equipment, an electro-hydraulic rectification system for mega-derricks is designed. This system employs a centralized control and multi-point drive architecture, utilizing a servo motor to power a radial piston pump as the oil source, using high-speed on/off valves for control and 12 high-pressure hydraulic cylinders are employed to synchronously lift the derrick legs. A dynamic master-slave multi-point synchronous control method with selection capability is proposed, which replaces closed-loop control with switch control, enabling high-precision synchronous control under heavy eccentric load conditions, regardless of the number of control points. We develop an electro-hydraulic rectification equipment for mega-derricks and conduct bench tests and field operations. The synchronous position error among multiple hydraulic cylinders is less than 0.3 mm, facilitating one-click alignment and significantly enhancing the efficiency, safety and automation level of large-scale derrick alignment tasks.

The application of traditional pneumatic mesh actuators is limited due to insufficient output force, with their structural design being a key constraining factor. To address this issue, we propose a stepped structure with non-uniform chamber heights, which enhances the terminal output force by suppressing radial expansion. A mathematical model of the actuator is established based on the principle of virtual work, revealing the correlation between structural parameters and bending performance. Combined with Hertzian contact theory, the contact force of the chamber wall is analyzed, and the theoretical relationship between the bending angle and contact pressure is established. To maximize the output force under the maximum input pressure 50 kPa, a three-factor and four-level orthogonal experimental design is adopted to optimize the parameter combination. The accuracy of the theoretical and finite element models is verified through bending performance analysis and experiments. Finally, the output forces of the stepped actuator and the traditional equal-height actuator are compared. The experimental results show that the stepped structure effectively suppresses radial expansion, with an output force of 3.5 N at 50 kPa, which is a 40% improvement compared to that of the traditional structure.

To diagnose internal leakage in hydraulic cylinders, a method based on a two-set-valued identification algorithm is proposed. This method detects faults by analyzing pressure signal variations under both normal and leakage conditions. Fifteen time-domain features are extracted from pressure signals in the rod chamber and piston chamber. Principal component analysis is applied to reduce these features to three principal components per chamber. A mathematical model integrating the six principal components from both chambers is established. The two-set-valued identification algorithm is employed to estimate model parameters for establishing internal leakage diagnosis algorithm. Experimental validation using AMESim simulation data confirms the method's feasibility and accuracy in diagnosing internal leakage, providing a theoretical basis for hydraulic cylinder fault detection.

Due to the fact that electro-hydraulic servo systems are often affected by internal and external disturbances, and internal parameters change due to mechanical structure wear, the traditional control strategies are no longer sufficient to meet the control performance requirements. Therefore, a filter-based adaptive asymptotic tracking control method with guaranteed performance is proposed. Firstly, a mathematical model of the electro-hydraulic servo system is established, which is transformed into a strict feedback form space state expression by defining state variables. Then, the controller design and stability analysis are carried out. A novel error transformation is designed and combined with a barrier function to achieve the prescribe performance constraints on the tracking error. In the controller design, a single-parameter adaptive method is adopted to estimate the unknown parameters. Meanwhile, to avoid the continuous differentiation of the virtual controller, a nonlinear filter is introduced. Finally, by using the Lyapunov stability theorem and Barbalat's lemma, it is proved that the system can achieve asymptotic tracking, and the tracking error is limited within the prescribe performance function range. All signals of the closed-loop system are semi-globally uniformly bounded. The effectiveness of the control strategy is verified through simulation and compared with traditional PID and backstepping control strategies. The research results show that the rise time is increased by 82.8% and 80.3% respectively, the adjustment time is decreased by 88.9% and 91.4% respectively, and there is no obvious overshoot with a relatively small tracking error.

This thesis proposes design approach of a biomimetic soft telescopic in-pipe robot that addresses the poor environmental adaptability and the insufficient active steering capability in current systems. The robot integrates flexible air chambers and pneumatic artificial muscles to construct a support-extension composite motion structure inspired by earthworm locomotion, adopting a multi-muscle coordinated actuation strategy and a continuous multi-segment locomotion method. The design includes both forward and inverse kinematic models, with system behavior verified through MATLAB simulations. An experimental platform enables performance tests in inclined and curved pipelines. The robot achieves an average crawling speed of 3.25 mm/s in a 30° inclined pipe and performs active steering in a 135° curved pipe, demonstrating strong adaptability and effective motion performance.