SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Hedrick, E., K. Hedrick, D. Bhattacharyya, S.E. Zitney, and B.P. Omell, "Integrating Reinforcement Learning with Model Predictive Control for Adaptive Control of Energy Systems," Clearwater Clean Energy Conference, Clearwater, FL, August 1-4, 2022.
Model Reference Adaptive Control (MRAC) is based on the fundamental concept that the process under investigation is to be controlled to follow or “track” a reference system (model) characterized by a state/input/output model employing an adaptive optimization algorithm to adjust the controller parameters in real-time. The generic structure of the MRAC is shown in Fig. 1 consisting of the following primary components: Reference model, Process (system) model, controller and the adaption algorithm. The basic structure of the controller is specified by a linear construct with the corresponding real-time adaption algorithms given by a gradient-type (so-called MIT rule) or based on stability theory (Lyapunov, hyperstability). This approach to adaptive control is termed “direct”, since the controller (parameters) are adjusted based on the component models/algorithm in contrast to the “indirect” approach that adjusts the process model parameters applying real-time system identification techniques.
Model reference adaptive control (MRAC) has been studied for decades and successfully applied in multiple areas, including heating, ventilation, and air conditioning (HVAC) systems for buildings. MRAC is efficient in capturing the time-varying characteristics of buildings' indoor temperatures and outdoor weather environments. In this paper, the rate of convergence of MRAC is investigated, where a direct adaptive control with temperature set point reference tracking is used to regulate the indoor temperatures for buildings. Numerical results show that by controlling the HVAC systems of residential buildings using MRAC, the indoor temperatures converge Q-sublinearly to the desired temperature set points. In addition, the rate of convergence for MRAC is compared with a baseline adaptive model-free control method.
This paper presents a comprehensive digital twin (DT) framework for hydroelectric systems that enables adaptive control of turbine governors and excitation systems without requiring detailed manufacturer specifications. The proposed framework integrates neural network-based system identification with stabilizing adaptive control laws for the installed turbine controller and middle-branch adaptive tuning for the installed voltage regulator. Using real operational data from Unit C-8 at Rocky Reach Dam (1,349 MW capacity), highfidelity neural network models are developed to capture turbine and generator dynamics without requiring detailed manufacturer specifications. The DT enables safe controller synthesis and validation in simulation before deployment. For turbine control, the proposed method achieves a 79.9% mean square error (MSE) reduction compared with that of an optimal controller. For voltage regulation, the adaptive excitation controller achieves approximately 42.6% MSE reduction while preserving installed protection logic. The results demonstrate that DT technology provides a practical pathway for modernizing hydropower control systems with minimal operational disruption.
Automation control is a key strategy to improve the economic competitiveness of nuclear power plants. Not only does it help reduce operational costs, but it also extends the value proposition of these plants to nontraditional markets, including unattended operations in remote villages and space. However, the dynamics of the operating environments of nuclear reactors are subject to changes over time, and there are no widely adopted methods to ensure that the automation strategy will remain effective over the extended durations required for these applications. Adaptive control is a discipline that offers the possibility to accommodate such changes online. However, it relies on mathematical assumptions that must be respected to ensure robustness and reliability. In this work, we derive an adaptive control formulation for linear systems in which all states are observable and apply it to an instance of load-follow operation for Boiling Water Reactors. We assumed uncertainty in two factors: the temperature coefficient and the control rod worth, both of which are affected over time by the evolution of the nuclear reactor core environment. With an arbitrary penalty factor of 5, we found that the mean absolute and integral time absolute errors can be reduced by more than 90%, underscoring the strength of adaptive control. To extend the application to more challenges, different uncertainties and load-follow trajectories, as well as new formulations that include non-linearity and partial observability, are currently being developed.
Hedrick, E., K. Reynolds, D. Bhattacharyya, S.E. Zitney, and B.P. Omell, "Adaptive Control Using Reinforcement Learning Augmented Model Predictive Control," Advanced Manufacturing and Processing Conference 2022, Bethesda, MD, June 1-3 (2022).
Volt-VAR and Volt-Watt functionality in photovoltaic (PV) smart inverters provide mechanisms to ensure system voltage magnitudes and power factors remain within acceptable limits. However, these control functions can become unstable, introducing oscillations in system voltages when not appropriately configured or maliciously altered during a cyberattack. In the event that Volt-VAR and Volt-Watt control functions in a portion of PV smart inverters in a distribution grid are unstable, the proposed adaptation scheme utilizes the remaining and stably-behaving PV smart inverters and other Distributed Energy Resources to mitigate the effect of the instability. The adaptation mechanism is entirely decentralized, model-free, communication-free, and requires virtually no external configuration. Here we provide a derivation of the adaptive control approach and validate the algorithm in experiments on the IEEE 37 and 8500 node test feeders.
In this paper, we propose an adaptive control strategy for residential buildings' heating ventilation and air conditioning (HVAC) systems to support grid services. The proposed control strategy captures the time-varying characteristics of the building and weather environments. It assigns to each building a local controller to maintain the indoor temperature within a desired comfort level, while the load aggregator allocates the assigned reference power provided by the distribution system operator to support a specific grid service, such as demand peak reduction, balancing supply and demand, and consuming the solar photovoltaic (PV) power locally. Simulation results are presented to verify the performance of the proposed control strategy in terms of maintaining the indoor temperatures of several buildings within the desired comfort bands while tracking the reference power profile. Also, a comparison with a recently developed model-free control strategy is presented.
The primary objective of the technical collaboration is to validate and implement adaptive feed rate control, tool wear monitoring, and anomaly detection for the finish machining of additive components. Most additively manufactured components need finish machining to achieve surface finish and tolerance requirements. Distortions and surface roughness of additively manufactured components can result in high forces and tool breakage or excessive tool wear during post-additive finish machining, which can result in rework or part scrap. In this collaboration, methods, and models for feed rate control, tool wear monitoring, and anomaly detection via a digital twin will be evaluated and validated.
This article proposes an approach to address the current and the aggregated active power control challenge for large-scale inverter-based resources subjected to partially or completely loss of inverters and grid voltage variations. To address this problem, a distributed active power mechanism is proposed which generates desired inverters current. Then an adaptive mechanism distributes the voltage control input automatically between inverters in response to the partial or complete loss of inverters. An L2-gain-based controller is designed for each inverter to track the desired current and rejects the grid voltage disturbance. Here, simulation results show a significant robust tracking for collective active power and current.
Volt-VAR and Volt-Watt functionality in photovoltaic (PV) smart inverters provide mechanisms to ensure system voltage magnitudes and power factors remain within acceptable limits. However, these control functions can become unstable, introducing oscillations in system voltages when not appropriately configured or maliciously altered during a cyberattack. In the event that Volt-VAR and Volt-Watt control functions in a portion of PV smart inverters in a distribution grid are unstable, the proposed adaptation scheme utilizes the remaining and stably-behaving PV smart inverters and other Distributed Energy Resources to mitigate the effect of the instability in real-time. The adaptation mechanism is entirely decentralized, model-free, communication-free, and requires virtually no external configuration. This repository provides code to simulate the algorithm in experiments on the IEEE 37 node test feeder.
This software is intended to facilitate the sending of data from DeepLynx to some human machine interface (HMI). The HMI would read the data generated by this software and potentially make a physical change on a process which the HMI controls. This software enables a digital twin using DeepLynx to talk to an HMI, thereby making autonomous updates to the physical asset which is being twinned.
According to an embodiment, there is provided an onboard integrated computational system for an unmanned aircraft system (“Stabilis” autopilot). This is an integrated suite of hardware, software, and data-to-decisions services that are designed to meet the needs of business and research developers of UAS. Stabilis is designed to accelerate the development of any UAS platform and avionics system; it does so with hardware modularity and software adaptation. The Stabilis offers multiple technological advantages technological advantages including: Plug-and-adapt functionality; Data-to-decisions capability; and, On board parallelization capability.
Coupled Computational-Fluid-Dynamics (CFD) and Discrete-Element-Method (DEM) models provide an accurate description of multiphase physical systems where a solid granular particle phase exists in an underlying gaseous continuous medium. The time integration of the granular phase in these simulations is typically handled using an explicit scheme with a constant time-step among all particles that is invariant in time to resolve inter-particle collisions. A locally third-order accurate adaptive time integration technique for particles that employs an embedded locally second-order scheme for error determination is presented in this work. The particle time-step size is dynamically adapted based on solution error, thus leading to significant savings in computational time. The efficacy of our scheme is quantified using four test cases of varying complexity (binary collision, homogeneous cooling system, fluidized bed and hopper discharge). The adaptive time-stepping method exhibits improved performance (~ 2–3 times in most of the cases studied) compared to three commonly used non-adaptive time-step methods (first-order Euler-explicit, second-order Adams-Bashforth and third-order Runge-Kutta schemes), while maintaining the same level of accuracy and parallel scalability.
Trapped atomic ions find wide applications ranging from precision measurement to quantum information science and quantum computing. Beryllium ions are widely used due to the light mass and convenient atomic structure of beryllium; however, conventional ion loading from thermal ovens exerts undesirable gas loads for a prolonged duration. Here, we demonstrate a method to rapidly produce pure linear chains of beryllium ions with pulsed laser ablation, serving as a starting point for large-scale quantum information processing. Our method is fast compared to thermal ovens, reduces the gas load to only 10 -12 Torr (10 -10 Pa) level, yields a short recovery time of a few seconds, and also eliminates the need for a deep ultraviolet laser for photoionization. We also study the loading dynamics, which show non-Poissonian statistics in the presence of sympathetic cooling. In addition, we apply feedback control to obtain defect-free ion chains with desirable lengths.
Not provided.
Explore the source record for details and available documents.