Learning and Adapting Model Predictive Controllers with Reinforcement Learning for Time-Varying Systems
Presented by NETL at 2019 AIChE Annual Meeting, 11/19
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.
Presented by NETL at 2019 AIChE Annual Meeting, 11/19
Backdoor data poisoning attacks have recently been demonstrated in computer vision research as a potential safety risk for machine learning (ML) systems. Traditional data poisoning attacks manipulate training data to induce unreliability of an ML model, whereas backdoor data poisoning attacks maintain system performance unless the MLmodel is presented with an input containing an embedded“trigger” that provides a predetermined response advantageous to the adversary. Our work builds upon prior back-door data-poisoning research for ML image classifiers and systematically assesses different experimental conditions including types of trigger patterns, persistence of trigger patterns during retraining, poisoning strategies, architectures (ResNet-50, NasNet, NasNet-Mobile), datasets (Flowers, CIFAR-10), and potential defensive regularization techniques (Contrastive Loss, Logit Squeezing, Manifold Mixup,Soft-Nearest-Neighbors Loss). Experiments yield four key findings. First, the success rate of backdoor poisoning at-tacks varies widely, depending on several factors, including model architecture, trigger pattern and regularization technique. Second, we find that poisoned models are hard to detect through performance inspection alone. Third, regularization typically reduces backdoor success rate, although it can have no effect or even slightly increase it, depending on the form of regularization. Finally, backdoors inserted through data poisoning can be rendered ineffective after just a few epochs of additional training on a small set of clean data without affecting the model’s performance.
In recent years there has been a considerable drive towards data-driven analysis, discovery and control of dynamical systems. To this end, operator theoretic methods, namely, Koopman operator methods have gained a lot of interest. In general, the Koopman operator is obtained as a solution to a least-squares problem, and as such, the Koopman operator can be expressed as a closed-form solution that involves the computation of Moore-Penrose inverse of a matrix. For high dimensional systems and also if the size of the obtained data-set is large, the computation of the Moore-Penrose inverse becomes computationally challenging. In this paper, we provide an algorithm for computing the Koopman operator for high dimensional systems in a time-efficient manner. We further demonstrate the efficacy of the proposed approach on two different systems, namely a network of coupled oscillators (with state-space dimension up to 2500) and IEEE 68 bus system (with state-space dimension 204 and up to 24,000 time-points).
PPL Electric Utilities’ Keystone Solar Future Project (KSFP) is part of the U.S. Department of Energy’s Office of Energy Efficiency and Renewables, Energy Enabling Extreme Real-time Grid Integration of Solar Energy (ENERGISE) grant. The three-year project, beginning in October, 2017, funded an innovative pilot program to develop and evaluate a technology platform and framework for managing high levels of distributed energy resources (DER), and improving future DER interconnection. The KSFP aimed to bridge gaps between existing and future technology, including Distributed Energy Resource Management Systems (DERMS), by monitoring, managing, and optimizing a system with high penetration of solar and other DER. Specifically, the project worked to make Demand Management Systems (DMS) applications, devices, and real-time communications technology DER-aware to allow for advanced applications for utility operations. Here, by installing and implementing a DERMS, project designers transformed how DERs interact with PPL’s distribution system, offering learnings and results worth sharing industry-wide.
In this paper, we present an intelligent system that has the capabilities of automatically selecting topology classes and optimizing circuit parameters of DC-DC power converters for a given design specification. The system, Machine-Learning-enhanced Automated Circuit Configuration and Evaluation of Power Converters (ML-ACCEPT), uses a hybrid of machine learning technologies, decision tree inference, reinforcement learning and deep neural networks. The system gives high accurate recommendations of design topology classes and computationally efficient results in optimizing power efficiencies in power converter design.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Abstract not provided.
Abstract not provided.
NREL has nearly a decade of experience in liquid cooling or HPC. The technical approach will be presented along with success stories on energy, cost and emission savings over two generations of HPC systems. Lessons learned will also be shared.
A method for active learning using sparse training data can include training a machine learning model using less than ten first training data points to generate a candidate machine learning model. The method can include performing a Monte Carlo process to sample one or more first outputs of the candidate machine learning model. The method can include testing the one or more first outputs to determine if each of the one or more first outputs satisfy a respective convergence condition. The method can include, responsive to at least one first output not satisfying the respective convergence condition, training the candidate machine learning model using at least one second training data point corresponding to the at least one first output. The method can include, responsive to the one or more first outputs each satisfying the respective convergence condition, outputting the candidate machine learning model.
One unifying thread of any fuel material considered for light water reactor (LWR) service is a need to exhibit satisfactory material response under a range of basic stimuli that must be designed for in this environment. The focus of the current work is twofold. First, the general conditions of relevance to LWR fuel service are identified. These are then translated into specific tests or test conditions that must be considered by experimental programs that intend to explore the behavior of a new fuel or fuel material system. Lessons learned and challenges specific to the current generation of LWR fuel replacement will be emphasized. (authors)
The commercial nuclear power industry is facing a potential shortage of certified nondestructive evaluation (NDE) analysts to meet future in-service inspection demands. Automated data analysis (ADA) currently supports human inspectors in tasks such as eddy current evaluations for steam generator examinations. Machine learning (ML) systems are nearing the capability to pass performance demonstration tests for ultrasonic testing (UT) inspections of reactor pressure vessel upper head penetrations in nuclear power plants (NPPs). Current research and development is focused on assisted analysis (AA) of ADA versus fully automated examinations. This presentation will cover assessment of ML flaw detection on dissimilar metal weld (DMW) piping joints.
Focus area: Primary focal area #2; secondary focal area #3: Learning about parameters and processes of land surface hydrologic and biogeochemical models in Earth System models by integrating machine learning, physics, and big data. Science challenges: How do we maximally leverage big-data observations to improve hydrobiogeochemical process description and parameterization so that such modules more realistically capture hydrologic and vegetation responses and feedbacks under the future climate? For example, how can we leverage physics, limited observations of vegetation and streamflow to better estimate evapotranspiration, and, relatedly, net primary productivity, especially for drought areas? Vegetation plays a critical role in regional and global water cycles; however, existing vegetation models have failed to predict vegetation response to droughts (McDowell & Xu, 2017) , arctic greening (Keenan & Riley, 2018) , and critical transitions between forest and savanna (Hirota et al., 2011) . These studies suggest that when we build process-based models (PBM) parameterized from regional and global plant traits, we tend to poorly describe plant adaptation and local-scale competition processes. The models and their associated parameters assigned for different regions in the world are not capturing essential heterogeneity in vegetation responses at finer spatial scales. Many parameters of the land surface models control hydrology and vegetation dynamics at the same time. The heterogeneity in vegetation response is a function of (i) plant type, (ii) plant size, (iii) competition and succession, (iv) environmental controls, and (v) local variations due to the unique ecological community that are very difficult to describe (e.g., the size of gaps resulting from fire that facilitated the coexistence of pioneering species). In the demographic models, only factors (i) and (iv) were captured, and plant types were generally described only by leaf phenology and climate zones. With current demographic models, we generally consider more traits to define plant types (i) and calibrate these traits to consider factors (ii), (iii) and (iv); however, it is substantially challenging to scale to regional and global simulations due to trait variations across space (Ali et al., 2016). Moreover, it has been noted that hillslope processes, including ridge-to-valley flow and sunny vs. shady slopes are primary organizers of water, energy, and vegetation (Clark et al., 2015; Fan et al., 2019) . Although gradual improvements in the hydrologic model component in earth system models may reduce this error (at a remarkably slow pace), the long-term, gradual impact of hydrology on plant traits are not well captured. Recent work showed that the hydrologic controls exerted by groundwater and lateral flow are primary regulators of rooting depth (Fan et al., 2017) . Such hydrologic controls have seldom been reflected in vegetation model parameterizations.
Abstract This paper reports on advances in the state‐of‐the‐art deep learning disruption prediction models based on the Fusion Recurrent Neural Network (FRNN) originally introduced in a 2019 NATURE publication [ https://doi.org/10.1038/s41586‐019‐1116‐4 ]. In particular, the predictor now features not only the “disruption score,” as an indicator of the probability of an imminent disruption, but also a “sensitivity score” in real time to indicate the underlying reasons for the imminent disruption. This adds valuable physics interpretability for the deep learning model and can provide helpful guidance for control actuators now implemented into a modern plasma control system (PCS). The advance is a significant step forward in moving from modern deep learning disruption prediction to real‐time control and brings novel AI‐enabled capabilities relevant for application to the future burning plasma ITER system. Our analyses use large amounts of data from JET and DIII‐D vetted in the earlier NATURE publication. In addition to “when” a shot is predicted to disrupt, this paper addresses reasons “why” by carrying out sensitivity studies. FRNN is accordingly extended to use more channels of information, including measured DIII‐D signals such as (i) the “n1rms” signal that is correlated with the n = 1 modes with finite frequency, including neoclassical tearing mode and sawtooth dynamics; (ii) the bolometer data indicative of plasma impurity control; and (iii) “q‐min”—the minimum value of the safety factor relevant to the key physics of kink modes. The additional channels and interpretability features expand the ability of the deep learning FRNN software to provide information about disruption subcategories as well as more precise and direct guidance for the actuators in a PCS.
A Machine Learning-based predictive model is developed for optimal dispatching energy storage system integrated with Electric Vehicle battery charger. The model is effectively trained using transfer learning algorithm and successfully validated via measurement data to achieve high fidelity and accuracy. The study findings provide possibilities for future development of the presented approach to implement computational heavy predictive algorithms with multi-step predictions in emerging modular bidirectional charger topologies to realize vehicle-to-grid (V2G) technology. The online tuning can be explored using the new measurements obtained during the operation, henceforth, enhancing the controller performance for integrating multiple DERs into distribution power networks that shows huge interest for the utilities and regulatory agencies. Multiple operation scenatios have been evaluated and simulation results are discussed to verify intended performance of the proposed control solution.
This paper presents a secure reinforcement learning (RL) based control method for unknown linear time-invariant cyber-physical systems (CPSs) that are subjected to compositional attacks such as eavesdropping and covert attack. We consider the attack scenario where the attacker learns about the dynamic model during the exploration phase of the learning conducted by the designer to learn a linear quadratic regulator (LQR), and thereafter, use such information to conduct a covert attack on the dynamic system, which we refer to as doubly learning-based control and attack (DLCA) framework. We propose a dynamic camouflaging based attack-resilient reinforcement learning (ARRL) algorithm which can learn the desired optimal controller for the dynamic system, and at the same time, can inject sufficient misinformation in the estimation of system dynamics by the attacker. The algorithm is accompanied by theoretical guarantees and extensive numerical experiments on a consensus multi-agent system and on a benchmark power grid model.