ABSTRACT Wind energy is one of the main renewable energy sources in the current energy transition. Due to ever more and ever larger wind turbines (WT), the requirements for WT operation become more complex. Model predictive control (MPC) for WTs shows the potential to handle these requirements and conflicting control objectives in a single optimization‐based controller. Recent research has widely investigated MPC for WT in simulation, but mostly lacks experimental validation. This work aims to experimentally validate MPC on a full‐scale WT under real conditions. To this end, we combine an extended Kalman filter for nonlinear state estimation with robust linear time‐varying MPC. We evaluate the proposed control algorithm in terms of time‐domain performance and power curve in simulation. However, the main contribution of this work is the experimental validation on a 3MW WT in Northern Germany with a total duration of 3‐h continuous full access of the controller. We were able to demonstrate stable operation of the proposed MPC in the upper partial load regime, transition regime, and lower full load regime, at measured wind speeds between 4.76 and 13.06 m/s, inside and outside the wake shadow of another WT. The power curve determined in simulation shows comparable results to a reference feedback controller. The MPC formulation combines several control objectives in a single optimization problem, yet the tuning effort still remains complex. In future work, we plan to reduce the complexity of the control loop based on this experimentally validated MPC. We provide our experimental data at https://doi.org/10.5281/zenodo.14644908.

In safety- and precision-critical control scenarios for permanent magnet synchronous motors (PMSMs), the external spontaneous disturbance causes unexpected speed drop. The disturbance occurs without routine, so it cannot be modeled specifically. The large speed drop and slow response speed cause a reduced life of the machines driven by PMSMs. Therefore, it is crucial to implement a method that can lead the controller to learn the effects caused by disturbances. To this end, this paper proposes a novel approach based on the basic structure of a backpropagation neural network (BP) for adaptive real-time adjustment in motor control. Regarding the lack of explainability of BP in existing methods, the electric motor physics is embedded into the BP (BP-PHY) gradient update part to enlarge the range of stability. To overcome the shortage of a potentially unstable output of neural network (NN), the learning parameter of NN is tailored based on the stability theory and motor physics. Finally, the proposed methods are implemented into simulations and experiments. The recovery time after disturbance decreases to 51.3% and the speed drop decreases to 50.3% compared to the basic controller of the PMSM, while the control stability of the NN is ensured.

Neuromuscular training to strengthen leg muscles is an important part of the treatment of musculoskeletal disorders and chronic diseases and preventing age-related muscle loss. This study evaluates different individualization approaches and their real-time implementation for OpenSim musculoskeletal models to estimate the external knee adduction moment during a leg-press exercise. A robotic neuromuscular training platform was utilized to perform isometric and dynamic leg extension exercises. Data were collected for 13 subjects using a 3D motion capture system and force plate measurements from the robotic training platform. Functional joint parameters, determined through dynamic reference movements, were integrated into the OpenSim models, allowing a personalized representation of the hip, knee, and ankle joints. This integration was compared with a conventional scaling method. The results indicate that the incorporation of functional joint axes can significantly enhance the accuracy of biomechanical simulations. These methods provide a real-time and a more precise estimate of the external knee adduction moment compared to conventional scaling approaches and underscore the importance of individualized model parameters in biomechanical research.

This work explores the potential of connected, digitalized Wire Arc Additive Manufacturing (WAAM) within the framework of Industrie 4.0, analyzing it through distinct process layers: workpiece, assembly, and product. Each layer presents unique timeframes and stakeholder interactions, necessitating varied data infrastructure demands, including a consideration of data security and privacy challenges. The workpiece layer mostly covers the local production setup and is thus directly coupled with the product and process quality as well as maintaining a safe operation. In the assembly layer, ensuring interoperability among diverse stakeholders is crucial, requiring clear definitions of responsibilities and access rights to enhance data exchange. The product layer prioritizes the reliability and trustworthiness of information for informed decision-making, advocating for solutions that guarantee authenticity and verifiability while addressing privacy concerns through techniques like privacy-preserving computing. The paper identifies a critical gap in real-world applications of these concepts in additive manufacturing. It proposes a data-driven quality control approach to enhance process and product quality in arc welding, leveraging digital shadows to create effective interfaces within production networks. This approach has demonstrated potential reductions in welding fume emissions by 12–40%, alongside connected applications that minimize exposure and energy consumption.

The spark-ignited (SI) hydrogen combustion engine has the potential to noticeably reduce greenhouse gas emissions from passenger cars. To prevent nitrogen oxide emissions and to increase fuel efficiency and power output, complex air paths and operating strategies are utilized. This makes the engine control problem more complex, challenging the conventional engine calibration process. This work combines and extends the state-of-the-art in real-time combustion engine modeling and optimal control, presenting a novel control concept for the efficient operation of a hydrogen combustion engine. The extensive experimental validation with a 1.5 l three-cylinder hydrogen SI engine and a dynamically operated engine test bench with emission and in-cylinder pressure measurements provides a comprehensible comparison to conventional engine control. The results demonstrate that the proposed optimal control decreased the load tracking errors by a factor of up to 2.8 and increased the engine efficiency during lean operation by up to 10 percent while decreasing the calibration effort compared to conventional engine control.

To enhance energy production, the wind energy industry builds increasingly larger wind turbines (WT) in terms of hub height and rotor diameter. This enlargement results in higher structural loads, which emphasizes the importance of load-reducing WT control alongside the nominal control of power and rotational speed. Generally, it is not possible to directly measure these structural loads, so a controller needs to include their online estimation. Here we extend a baseline WT state estimator in the form of an Extended Kalman Filter (EKF) to include two unaccounted loads. These loads are a change of thrust in the main bearing and a yaw moment in the rotor hub. We incorporate these loads via data-driven local linear neuro-fuzzy models (LLNFM). These LLNFMs can represent nonlinear relationships while maintaining limited complexity. We use alaska/Wind to generate the underlying regression data and to perform the state estimation both in simulation. The regression data consists of nominal WT operation under turbulent conditions for different average wind speeds. We further superimpose the individual pitch angles to excite the WT. The evaluation is performed in a different scenario, using another turbulent wind field and a realistic noisy sensor configuration. While we can estimate the change of thrust in the one per revolution (1P) frequency range, we achieve more precise results for the yaw moment due to the incorporation of sensors with a strong correlation to it. The estimation of the change of thrust and the yaw moment appear to be sufficiently precise for load-reducing WT control in practically relevant frequency ranges.

This paper presents a nonlinear model predictive control strategy for automotive fuel cell system operation. The system-level air-path modeling approach, widely adopted in control-oriented studies, is extended by the critical aspect of membrane hydration. The control problem formulation reflects the task of dynamic power tracking and efficiency-optimized actuation of the peripheral components as well as adherence to system constraints to avoid harmful operation. This is combined with an analysis of the design parameters sampling time, integration scheme, and prediction horizon for efficient transcription of the optimal control problem. Closed-loop simulation results, conducted using a sampling rate of 8ms, the standard fourth-order explicit Runge-Kutta method with one integration step and a horizon length of 25, successfully meet the control objectives.

Fuel cell systems constitute an electrochemical energy conversion system increasingly used in stationary and mobile applications. Complying with operational limits in transient operation can be achieved by model-based predictive control algorithms. The key challenge arises from the identification of suitable models for embedded real-time optimization. This article presents a data-driven predictive control approach for the air path and power control of a fuel cell system. In particular, we use data-enabled predictive control (DeePC) based on a concise system representation using column subset selection (CSS). The impact of problem formulation, regularization, and different solvers for quadratic programs (QPs) on the turnaround time on embedded hardware is investigated. In addition, we provide an online update algorithm for the system representation to account for the operating regions not contained in the initial dataset. The proposed approach is validated on a high-fidelity fuel cell system simulation and hardware-in-the-loop (HiL) experiments. We demonstrate safe and fast closed-loop control using the column subset algorithms for a comprehensive dataset and reduction in closed-loop cost for unknown operating areas of up to 25%. The control algorithm and the update algorithm are shown to be real-time feasible on a single-core embedded hardware.

Modern vehicles are commonly sold with air conditioning systems. Especially in climates with traditionally high temperatures, these systems are not just for comfort but highly relevant also to safety. Various hardware setups for cooling and heating have been used to climatize the vehicle cabin. No matter which setup is used though, a cabin temperature controller needs to be implemented. Recent developments see an increase of multizone climate control methods. Even though various approaches have been presented, the task of temperature tracking and disturbance rejection remains challenging. In this article, it is shown using experiments that separating cooling circuit control from cabin temperature control does not affect the performance of the cabin control loop. For cabin temperature control, an approach is presented which allows for simultaneous air outlet temperature and mass flow management based on previous work. This is possible in multiple vehicle zones. The approach is based on a gray-box model of vent system and cabin. A mass flow management system based on real-time steady-state optimization grants simultaneous tracking and disturbance rejection. The weakness of this previous work, namely, the dependence on careful hand-tuning and mass flow independent tuning is overcome by applying an adaptive tuning approach. All results are experimentally validated using an experimental vehicle in real traffic for the two-zone control case. Consistent behavior over a wide range of mass flows and temperatures was reached. This correlates to a consistent thermal experience for passengers. Overshoot was effectively eliminated by applying the adaptive tuning. Concluding, an analysis of the move to practical implementation is shown. In this, a cabin temperature observer is presented that demonstrates the difficulties of implementing accurate three-zone control on a limited set of sensors. Based on the implemented cabin temperature observer, tracking can be achieved on a reduced sensor setup albeit at the cost of worse closed-loop controller performance.

Donor cell-specific tissue-engineered (TE) implants are a promising therapy for personalized treatment of cardiovascular diseases, but current development protocols lack a stable longitudinal assessment of tissue development at subcellular resolution. As a first step toward such an assessment approach, in this study we establish a generalized labeling and imaging protocol to obtain quantified maturation parameters of TE constructs in three dimensions (3D) without the need of histological slicing, thus leaving the tissue intact. Focusing on intracellular matrix (ICM) and extracellular matrix (ECM) networks, multiphoton laser scanning microscopy (MPLSM) was used to investigate TE patches of different conditioning durations of up to 21 days. We show here that with a straightforward labeling procedure of whole-mount samples (so without slicing into thin histological sections), followed by an easy-to-use multiphoton imaging process, we obtained high-quality images of the tissue in 3D at various time points during development. The stacks of images could then be further analyzed to visualize and quantify the volume of cell coverage as well as the volume fraction and network of structural proteins. We showed that collagen and alpha-smooth muscle actin (α-SMA) volume fractions increased as normalized to full tissue volume and proportional to the cell count, with a converging trend to the final density of (4.0% ± 0.6%) and (7.6% ± 0.7%), respectively. The image analysis of ICM and ECM revealed a developing and widely branched interconnected matrix. We are currently working on the second step, that is, to integrate MPLSM endoscopy into a dynamic bioreactor system to monitor the maturation of intact TE constructs over time, thus without the need to take them out.

In this paper, first the overall modeling approach for an optimized control of a hot-gas cycle for solar thermal power plants in the Modelica language is pointed out. The emphasis of the modeling work lies on the development of dynamic component models to be used in control systems. Depending on the control task, the discretization of the models has to be adapted. Main components of the hot-gas cycle are the solar thermal receiver and the storage system. The steam cycle is preliminarily only included as heat sink. Second, for control purposes a linear model-based controller (MPC) was implemented in Modelica based on an external state-of-the-art QP solver [2] linked to the Modelica model. The performance of the MPC is compared with a basic automation scheme based on classical PID controllers.

In-mould process data is essential for high quality injection moulding processes due to its high correlation to multiple quality criteria, such as part weight, dimensions and surface properties. To take advantage of the high correlation of cavity pressure and part quality, we developed a phase-unifying model-based cavity pressure control. This approach allows an approximate prediction and realisation of cavity pressure curves in real-time by calculating the optimal screw velocity adjustment based on a process model. The phase-unifying process control approach avoids discontinuities by eliminating the switch-over point and enables thereby a smooth transition from filling to packing. Besides the process control concept, the specification of a suitable process setting is the second key factor in achieving high part quality. A process adaption is always required, if process disturbances occur. For example, production interruptions can occur in the injection moulding process when parts are demoulded incorrectly. Furthermore, process disturbances can influence the injection moulding process such that the process parameters leave a predefined monitoring window and cause a machine interruption. During these interruptions, the thermal household of the mould and polymer melt changes. This causes rejection of parts during the subsequent start-up process due to the changed process conditions. The start-up of the process takes a long time until quality criteria can be again fulfilled. The study aims to shorten the start-up process as use-case for phase-unifying process control. For this purpose, injection moulded parts were produced for various interruption times. The part quality of the first five parts of start-up process was determined by weight measurements. Additionally, we carried out injection moulding trials for different cavity pressure references to quantify the correlation between cavity pressure characteristics and part weight as quality criteria. Quantitative correlations between cavity pressure characteristics and part quality were determined by calculating an ideal cavity pressure reference for inline process control for the start-up process. This allows high process stability under the occurrence of production interruptions with fast mould filling. The quality consistency of part weight was increased by 35 % for the first five produced parts. This leads to a highly reproducible process and fewer rejected parts during the production start-up. Part weight fluctuations are significantly lower compared to conventional process control, which shows the high potential of a combined inline and online phase-unifying process control for injection moulding.

Crane-based handling operations represent a vital part of today's maritime sector. To guarantee both safe and efficient payload handling with ship cranes, payload oscillations induced by the sea swell have to be damped. Approaches for payload stabilization usually consider either vertical position control (Active Heave Compensation, AHC) or sway reduction (Anti Sway Control, ASC). In this paper, a unified control scheme for spatial payload stabilization is proposed, utilizing the differential flatness of the crane system. The presented framework inverts the nonlinear payload dynamics by means of the flat mapping, thus facilitating controller design. Furthermore, the approach provides a systematic way to include the sea disturbance in the controller. Redundancy of the considered knuckle boom crane is exploited to track secondary control objectives. A related cost function weighting the crane's manipulability is derived and used in an optimization-based target selector. The controller design is evaluated in simulation for varying sea disturbances. The results suggest good AHC and ASC capabilities, where the payload's position error is reduced up to 60% for light to medium sea states.

In many medical applications data is a scarce resource and can often only be obtained with invasive surgery. This is for instance the case for physiological cardiovascular data that is necessary to improve the functionality of assistive heart devices. In this work we explore the viability of a GAN architecture to generate cardiovascular data towards enriching a data set obtained in animal testing on which training of future applications can be improved which potentially reduces the need for further animal testing. We evaluate the usefulness of our synthesized data using a downstream task.

Fuel cell systems are a viable alternative for energy conversion in stationary and mobile applications. Advanced control algorithms are the main levers to ensure safe operation in transients and increase the applicability of fuel cell systems in research and industry. This paper focuses on the control of the fuel cell air path and the net power output for a small-scale fuel cell system. For safe operation and durability even in transients, tight bounds on stoichiometry and compressor operation must be ensured at all times. To tackle this challenge, a data-based nonlinear model predictive controller is implemented and experimentally validated on a cathode path test bench with a real-time fuel cell stack simulation. Our results show accurate tracking, safe operation, and a reduction in settling time to new power reference set points of approximately 50% compared to a reference controller.

In this paper, techniques for reducing the number of optimization variables for quadratic programs arising in model predictive control are reviewed. Focusing on optimization with a first-order method, numerical properties such as the condition number of the Hessian are evaluated and suboptimality due to early termination of the optimization algorithm is investigated. The state elimination condensing approach and a recently proposed numerically robust approach based on the QR factorization are compared to the existing methods of prestabilizing the prediction and preconditioning for reducing the Hessian condition number for linear MPC. For the application example of controlling a fuel cell system, the worst-case turnaround time of the control algorithm can be reduced by more than 30% compared to standard state elimination algorithms. Thus the controller is shown to be real-time feasible using the QR factorization or additional preconditioning.

For a centralized path planning in the multi-agent path finding (MAPF), especially for road vehicles using traffic rules, a prioritized planning algorithm is one of the key methods that deal with real-time problems. The time used for planning is limited, especially when a number of agents are present, and a suboptimal solution has to be found. If possible, the previous solution should be reused for the replanning. Anytime Dynamic A* (AD*) can both replan the path when the environment is changed dynamically and is able to return a safe, but potentially suboptimal solution in case where a maximum cycle time is exceeded. A prioritized order selection is based on the path length to the goal. To model heterogeneous traffics with vehicles and pedestrians, a combination of the bicycle and pedestrian obstacle reciprocal collision avoidance algorithms (B-ORCA and PORCA) is used to realize a naturalistic interaction among the multi-agent system and the dynamic environment. The number of agents and of dynamic obstacles is randomly generated and the time used for planning is analyzed in view of the limited time available in the simulations. In this work, the path planning of up to three agents can be executed in real-time in MATLAB.

In autonomous applications for mobility and transport, a high-rate and highly accurate vehicle-state estimation is achieved by fusing measurements of global navigation satellite systems (GNSS) and inertial sensors. The state estimation and its protection-level generation often suffer from satellite-signal disturbances in urban environments and subsequent poor parametrization of the satellite observables. Thus, we propose an innovative scheme involving an extended H ∞ filter (EHF) for robust state estimation and zonotope for the protection-level generation. This scheme is shown as part of a tightly coupled navigation system based on an inertial navigation system and aided by the GPS/Galileo dual-constellation satellite navigation system. Specifically, GNSS pseudorange and deltarange observables are utilized. The experimental results of post-processing a real-world dataset show significant advantages of EHF against a conventional extended Kalman filter regarding the navigation accuracy and robustness under various GNSS measurement parametrizations and environmental circumstances. The zonotope-based protection-level calculation is proven valid, computationally affordable, and feasible for real-time implementations.

The validation and certification of wind turbines (WT) on nacelle test benches (NTB) is becoming increasingly important in the development process. While for certifiers the advantage lies in controlled test execution, for development departments it lies in testing as many system components as possible in a quasi-final prototype. However, the question arises which practical conditions must be fulfilled so that statements can also be made about WT control. In addition to the errors induced by a mechanical Hardware in the Loop (mHiL) system, the dynamic interactions between the WT controller and the NTB controller applying the mHiL concept are of interest. This analytical work based on simulations aims to systematically investigate how realistic the control behavior of a WT operated on a NTB is. For this purpose, the nominal behavior of a WT is compared with the operation on a NTB under realistic conditions and the resulting differences are subsequently reproduced in a synthetic load case. Finally, the differences are analyzed in terms of system theory. It is found that a frequency-dependent distorted behavior caused by operating the WT on a NTB is responsible for strong deviations compared to the WT operation in field. In the controller configuration studied, gain amplifications up to 5.17 dB are identified. The distortion is not exclusively caused by the mHiL closed loop behavior, but results from the interaction of all subsystems in both control loops. Therefore, its behavior is identified as a function of the system and controller parameters of both the WT and the NTB.