Control Engineering Practice最新文献

For a reliable development of energy-maximising optimal control strategies in wave energy, wave excitation force estimators are fundamental components in the control loop, since the optimality of the solution depends on their precision. In the control deployment phase, it is necessary to test the precision of wave estimates to assess the eventual margin of improvement for optimal energy extraction. Nonetheless, current strategies for wave estimator validation are based on the possibility of repeating with high precision the wave signal, and blocking the device in a predefined position, assumptions which are possible only in certain test facilities. This study proposes a novel estimator validation strategy, termed closed-loop validation, based on the analysis of the system motion only, which is suitable for testing excitation force estimators under realistic (open-sea) conditions. For this purpose, two excitation force observers, based on a random walk and a harmonic oscillator wave description, are developed, on the basis of the energy conversion device identified model. In particular, the device is a WaveStar prototype at the Aalborg University facility tank, in Denmark. The observers (and the identified model) are validated under irregular sea states, with both the classical and the proposed scheme, proving the effectiveness of the latter in demonstrating the excitation force estimator performance without requiring a direct force measurement.

Aiming at the complex nonlinear and uncertain multi-variable coupling problems of heavy-duty gas turbine, a data-driven heavy-duty gas turbine modeling method combined with physical constraint is proposed. Firstly, by combining the advantages of MTGNN that focuses on local features and Transformer that extracts global information, an improved network is constructed to model heavy-duty gas turbine with full operating conditions and wide load variation. Secondly, the physical constraint related to the mechanism of heavy-duty gas turbine is added to eliminate the unreasonable output in the proposed heavy-duty gas turbine model, which improves the reliability and interpretability of the proposed modeling approach. Then, the Lyapunov function is utilized to prove the convergence of the proposed modeling approach. Finally, the actual datasets of heavy-duty gas turbine are used to carry out relevant emulation and comparison experiments. The results show that the proposed modeling approach is superior to other typical network models, which verifies the accuracy and practicability of the proposed modeling approach.

This paper introduces an adaptive time-delay estimation error compensation (ATDE2C) scheme that strategically reduces the difficult-to-predict tracking errors of time-delay control (TDC) for given reference trajectories. Specially, these tracking errors mainly come from the so-called TDE errors caused by one-sample delayed measurements associated with a time delay estimation (TDE) technique in TDC. ATDE2C enhances robustness of TDC against such inevitable TDE errors using a disturbance observer (DOB)-based approach, thereby enlarging the stable operational region and offering a more robust performance compared to the basic TDC. The ATDE2C scheme enhances tracking performance of robot manipulators by adaptively feeding back the difference between actual and estimated TDE errors using a disturbance observer. This approach is particularly effective in managing hard nonlinearities, such as Coulomb friction, encountered in high-degree-of-freedom robots during rapid joint movements. It minimizes frictional effects and prevents overshoots, ensuring precise control. Moreover, it allows for intuitive tuning of the controller through parameters focused on convergence speed, making it practical for application compared to existing controllers. Extensive testing, including simulations and experiments, confirms improved performance of ATDE2C in real-world applications.

Multi-phase machines are being proposed in research and industry applications due to their inherent advantages. Finite Control Set Model Predictive Controllers (FCS MPC) constitute a flexible tool to deal with the increased number of phases. However, FCS MPC is computationally demanding, reducing its applicability with fast power converters. Region-based methods for FCS MPC have been recently proposed to reduce the control computation time. However, these methods are still hindered by the use of a fixed application time for the selected voltage. The paper presents a region-based method that uses variable application times. The method achieves fast computation of the voltage vector and its application time. Moreover, the application time can be found with fine resolution unlike previous approaches and without iterations. The proposal is experimentally validated using a laboratory test bench based on a symmetric five-phase induction machine showing improved control results and a reduction in harmonic content.

Reusable launch vehicles (RLVs) offer advantages such as low cost and high efficiency, yet they still face challenges related to real-time precise control and reliability enhancement. To address these issues, this paper presents a fixed-time attitude control method designed for RLVs, incorporating a reliability-based control allocation scheme. Initially, a fixed-time radial basis function neural network (RBFNN) disturbance observer is introduced to estimate disturbances rapidly and accurately. Subsequently, a control method integrating non-singular terminal sliding mode control (NTSMC) with a saturation compensator is proposed to achieve real-time and precise angle tracking. Furthermore, a reliability-based control allocation scheme is employed to enhance the reliability of the attitude control system. Finally, simulation analysis is conducted to validate the effectiveness of the proposed control scheme.

The polyester fiber production process is a typical example of a large-scale system. The monitoring and prediction of key process variables play a crucial role in the stability of the polyester fiber production processes. Especially during the esterification stage, there are multiple types of strong nonlinear dependencies among many sensor variables, which has always been a challenge in non-diffusive system modeling and control. Therefore, an Edge-Specific Message Graph Convolution Network (ESMGCN) is proposed to achieve separate modeling of specific dependencies between each pair of sensors individually, and to describe more accurately non-diffusive polyester fiber production systems. In addition, to address the problem of different degrees of dependencies between different sensor community clusters in large-scale systems, a community graph structure generation method is proposed to construct the graph structure within and between sensor communities, respectively. Finally, a multivariate time series prediction model for the polyester fiber polymerization process is constructed. Experiments are conducted in a real large-scale industrial system, and the results verify the model’s effectiveness.

In this paper, the issue of fixed-time tracking control for the uncertain wearable exoskeleton with prescribed performance is addressed. To heighten the transient and steady-state performance of the exoskeleton system, a variable exponent coefficient-based fixed-time control approach is formulated. Benefiting from model-free control, the dynamic model of the exoskeleton is replaced by an ultra-local model that relies solely on input–output (IO) data. The tracking control problem of the exoskeleton system boils down to the boundedness problem of auxiliary variables via designing a nonlinear conversion function. The tracking errors converge to the origin in a fixed time and never violate predefined constraints. The mechanical design of the wearable exoskeleton is carried out by Solidworks, and the actual control system is meticulously constructed. The simulation and experimental outcomes confirm the efficacy and superiority of the presented control methodology.

This paper provides a thorough evaluation of well-known control algorithms, including proportional–integral (PI), model predictive control (MPC), and H-infinity ($H_{\infty }$) controllers, by implementing them in a full nonlinear wind turbine model under normal wind conditions in below and above-rated wind speeds. The simulation results show that all the controllers perform satisfactorily. This study extends MPC to include a feedforward (FF) loop (FF-MPC) that uses the wind speed information provided by a light detection and ranging (LiDAR) sensor, which measures the upcoming wind (in advance), to improve the overall control performance. The FF-MPC was tested under both normal and anomalous (i.e. gusty) wind conditions. The results were compared with those of the standard feedback MPC (FB-MPC). The results show that the incorporation of the FF loop into the standard FB controller can improve the control performance, which can result in improved reliability and lifespan of the turbine. Furthermore, MPC was augmented with an FF loop over PI and $H_{\infty }$ controllers owing to its versatility in handling constraints, nonlinearities, and multiple objectives, along with its inherent capability to incorporate preview wind data. All the controllers are tested using a high-fidelity aeroelastic model (i.e. Bladed by DNV). The use of a Bladed model is common in wind turbine controller design before the application to the real-life wind turbine, and Bladed also allows more realistic simulation when incorporating a LiDAR.

The amount of water in the soil has an impact on how well plants absorb nutrients. As a result, both above and below-the-surface water dynamics have an effect on crop growth. The quality of the soil and its mechanical qualities can be significantly altered by both an excess or a lack of water, which could make food production unsustainable. This study deals with the development of a precision farming approach to irrigation that considers the topographical features of the arable land and incorporates its morphological properties. The prediction of the water movement in the topsoil layer is an essential element of this strategy, which uses an agent-based model to describe the soil dynamics and the impact of irrigation and exploits a model predictive control (MPC) to optimize water usage. As a case study, the municipality of Samacá in the department of Boyacá, Colombia, is considered.

To meet the stringent nitrogen oxides emissions regulations, the recent marine two-stroke engines are equipped with exhaust gas recirculation (EGR) systems. Crossing the boundaries of exhaust control areas requires switching between the engine operating modes associated with the activation or deactivation the EGR system, which imposes challenges to the engine operation. This study aims to develop and numerically test effective control strategies for improving the mode switching of marine two-stroke engines with EGR systems. The investigated engine dynamic model is developed by integrating a zero-dimensional thermodynamic model developed in GT-POWER and the control model developed in MATLAB/Simulink. This model is first validated against the experimentally measured engine performance parameters, and subsequently employed to simulate the engine dynamic operation with mode switching. The simulation results are assessed to quantify the impact of several control strategies on the engine performance. The derived perrformannce parameters time variations demonstrate that the deactivation of the EGR branch and the small engine–turbocharger result in exceeding the manufacturer limits at high loads and increasing the fuel consumption. The proposed load-based control strategies, which employ load thresholds and time delays for the control of the EGR and turbocharging systems activation/deactivation, are proved effective, as the engine performance parameters are retained within their limits and fuel consumption penalties are minimised. This study provides insights for developing the control of the marine engines after-treatment systems, hence contributing towards enhancing shipping sustainability.