Ruqi Ding

Independent metering control systems (IMCS) pose significant fault diagnosis challenges due to strong hydraulic– mechanical coupling, limited sensing, and the indirect mapping between internal faults and measurable signals. To address the limited fault coverage of existing IMCS diagnostic approaches, this paper presents a model-based fault diagnosis framework enhanced with fuzzy logic reasoning. First, a physical model is calibrated using time-domain similarity metrics to improve consistency between simulated and measured responses. Using this model, state- and input-based residuals are derived by exploiting IMCS analytical redundancy. These residuals capture the characteristic signatures of valve, displacement, and pressure sensor faults. To further enhance robustness, a trend-aware residual smoothing mechanism is introduced to suppress fluctuations caused by noise and disturbances while preserving fault-related signatures. The smoothed residuals are then evaluated by a fuzzy logic-based system, enabling interpretable fault classification and severity estimation. Experimental validation on a full-scale test platform demonstrates that the proposed method can accurately identify multiple fault types and severity levels, achieving an overall diagnostic accuracy of 96.67% without requiring additional sensors or hardware modifications. Future work will extend this framework from single-fault conditions to concurrent fault diagnosis.

To overcome the oscillations of the individual metering control system, a control scheme matrix including eight active damping methods is conducted and compared, with different signal feedback and different valve regulation. Through mathematical modeling and numerical simulation, the comparison result provides a guideline to determine a proper control scheme considering damping capacity, anti-disturbance ability, and availability for different applications. Considering the limitations of these damping methods, an inlet-outlet combined damping compensator is proposed, in which the inlet valve and the outlet valve are regulated simultaneously. Separate damping compensators for the inlet and outlet valves are designed, and the determination criterion of compensation parameters is presented. The benefit is that the additional dynamic behavior caused by pressure feedback is reduced to improve the anti-disturbance ability. A comparative simulation study using a 20-ton excavator was carried out under different working conditions. The results show that the proposed method has a similar oscillation damping with the traditional method under a given velocity command. More importantly, the proposed method reduces the velocity oscillation by additional 24.3%–28.5% and reduces the pressure oscillation by additional 5.2%–21.3% under sudden external disturbance.

To improve the performance of the magnetorheological (MR) semi-active suspension system, a fractional-order sliding mode control (FOSMC) strategy is proposed to address the nonlinear problems caused by the external disturbances and changes in parameters of system. Firstly, The MR damper designed by the research team is applied to the vehicle’s MR semi-active suspension, and a one-fourth semi-active suspension dynamic model is established. Mechanical performance tests of the MR damper are conducted through tensile experiments, confirming its satisfactory mechanical characteristics and suitability for meeting the output damping force requirements of the vehicle’s semi-active suspension. To accurately convert the computed control force into the MR damper’s control current, an adaptive neuro-fuzzy inference system (ANFIS) inverse model of the MR damper is constructed. Comparative data analysis demonstrates the inverse model’s high accuracy and its ability to effectively meet the controller’s requirements. The proposed FOSMC strategy effectively reduces the chattering phenomenon associated with sliding mode control, leading to improved vibration-damping performance of the suspension. Subsequently, a simulation model and experimental platform are utilized to evaluate the performance of the proposed controller. Experimental results indicate that the FOSMC-controlled suspension outperforms the passive suspension, achieving significant reductions in body acceleration, suspension dynamic travel, and tire dynamic displacement by 27.4%, 7.3%, and 17.39%, respectively. Similarly, the semi-active suspension controlled by sliding mode control (SMC) also shows improvements compared to the passive suspension, with reductions in body acceleration, suspension dynamic travel, and tire dynamic displacement by 20.9%, 4.3%, and 5.2%, respectively. In summary, the FOSMC-controlled MR semi-active suspension exhibits superior damping performance compared to both the passive suspension and the SMC-controlled semi-active suspension, demonstrating the effectiveness of the proposed FOSMC controller.

Accurate identification of dynamic parameters is essential for precise motion control and autonomous operation of heavy-duty hydraulic manipulators. However, due to their low-speed motion property, conventional approaches fail to simultaneously excite all parameters. To overcome this issue, a sequential parameter identification approach for heavy-duty hydraulic manipulators is proposed. All parameters are categorized based on their dynamic characteristic, and then distinct excitation trajectories have been designed to separately stimulate and identify each parameter. Dynamic parameters are fully excited, which is reflected in a reduced condition number of the observation matrix. Furthermore, an approach that ensures the physical feasibility of the identified parameters is constructed, which makes them more suitable for application in nonlinear control. The performance of the proposed method is evaluated with various identification methods, including traditional least squares, weighted least squares, and the method only considering physical feasibility. The results indicate a substantial decrease in torque prediction error compared to these methods. Specifically, the prediction accuracy of the joint torque using the proposed method has been improved by approximately 5.63% to 27.06%.

The increased flexibility and introduced coupling between links and cylinders pose challenges for modeling flexible hydraulic manipulators (FHMs). The traditional assumed mode method (AMM) encounters the problems of low computational accuracy resulting from a single basis function, as well as the low computational efficiency of multidegree-of-freedom calculations. In this article, a novel dynamic modeling method utilizing Lagrange’s dynamic equation with a multibasis function combination is proposed. To increase the modeling accuracy, flexible deflections are determined through a linear combination of several basis functions. All these basis functions are derived using the Rayleigh–Ritz method by considering geometric deformation boundary conditions and curvature changes in the links. To improve computational efficiency, the flexible link system is divided into two virtual rigid and flexible parts, which are analyzed separately to minimize the impact of coupling terms. Deflections in the dynamic equation affect only the flexible eigenmatrix and its derivative term such that the dynamic equations are simplified. The experiments on a 13-m FHM verify that, compared with the traditional AMM, the proposed modeling method can effectively increase accuracy while improving computational efficiency.

Flow soft sensing of hydraulic pump is a crucial way to avoid use of flow sensor, but it is difficult to achieve due to the high nonlinearity relationship with pressure, temperature, and rotational speed. Especially, the nonlinear flow relationship changes gradually with wear degradation. To cope with this issue, a flow soft sensing method is proposed combined with transfer learning and semi-empirical modelling in this study. A soft flow sensing model was firstly trained using sufficient dataset based on Back Propagation Neural Network (BPNN). An iterative model updating method is proposed with parameter-based transfer learning to cope with wear degradation after long-term operation, in which semi-empirical mathematical modelling is involved to provide an additional constraint and reduce the requirement of large sample size. In the updating process, the weights and thresholds of the initial soft-sensing model are fine-tuned to reduce the flow prediction error in forward propagation. The experimental results show that compared with the initial BPNN model and the updated model with transfer learning directly, the proposed method can reduce the average flow sensing error from 1.43 L/min and 3.25 L/min to 0.34 L/min, and the standard deviation (SD) from 0.57 L/min and 3.31 L/min to 0.35 L/min. Thus, the proposed method can achieve accurate soft sensing of the pump flow under wear degradation effectively.