A5003062911- Power System Optimization and Stability
- Smart Grid Energy Management
- Optimal Power Flow Distribution
- Microgrid Control and Optimization
- Smart Grid Security and Resilience
- Power Systems Fault Detection
- Power System Reliability and Maintenance
- Anomaly Detection Techniques and Applications
- Network Security and Intrusion Detection
- Advanced Electrical Measurement Techniques
- Fish Ecology and Management Studies
- Water Quality Monitoring Technologies
- Machine Fault Diagnosis Techniques
- Computational Physics and Python Applications
- Time Series Analysis and Forecasting
- Fish Biology and Ecology Studies
- Machine Learning in Materials Science
- Computational Drug Discovery Methods
- Machine Learning and Algorithms
- HVDC Systems and Fault Protection
- Mobile Health and mHealth Applications
- X-ray Diffraction in Crystallography
- Electrical Fault Detection and Protection
- Advanced Adaptive Filtering Techniques
- Thermal Analysis in Power Transmission
Pacific Northwest National Laboratory
2018-2025
Battelle
2023-2024
University of Wyoming
2016
Load shedding has been one of the most widely used and effective emergency control approaches against voltage instability. With increased uncertainties rapidly changing operational conditions in power systems, existing methods have outstanding issues terms either speed, adaptiveness, or scalability. Deep reinforcement learning (DRL) was regarded adopted as a promising approach for fast adaptive grid stability recent years. However, DRL algorithms show two when being applied to system...
As power systems are undergoing a significant transformation with more uncertainties, less inertia and closer to operation limits, there is increasing risk of large outages. Thus, an imperative need enhance grid emergency control maintain system reliability security. Towards this end, great progress has been made in developing deep reinforcement learning (DRL) based solutions recent years. However, existing DRL-based have two main limitations: 1) they cannot handle well wide range...
Abstract Deep learning models have proven to be a powerful tool for the prediction of molecular properties applications including drug design and development energy storage materials. However, in order learn accurate robust structure–property mappings, these require large amounts data which can challenge collect given time resource-intensive nature experimental material characterization efforts. Additionally, such fail generalize new types structures that were not included model training...
We present a novel framework that we name "Learning Advance" for hypothesis generation and validation the discovery of chemical knowledge in context optimizing solubility amphiphile/water systems. The workflow begins with an initial hypothesis: incorporation common hydrotropic additives, such as sugars or urea, enhances limits. To test this assumption, employ grid search Latin hypercube sampling approach to design experimental combinations additive weight percentages. high-throughput robotic...
With the fast-paced advancement of artificial intelligence (AI), emerging applications large language models (LLMs) have demonstrated useful in material science and self-driving labs (SDLs). Lacking understanding relationship between probability distribution an LLM token output, especially a scientific setting, some concerns are maintaining rigor autonomy when integrating AI tools. In this correspondence, we propose to use with function calling as automation node. Such can be virtual or...
Many mHealth interventions for health behavior change are considered effective improving outcomes. However, there is a limited understanding of the role components in an intervention on its effectiveness. Insights into such as content and software features needed to design efficient interventions. In this study, we conducted exploratory analysis objective data from usage weight management app understand loss. We identified positive correlation between loss use intervention. also found...
Adult American eels (Anguilla rostrata) are vulnerable to hydropower turbine mortality during outmigration from growth habitat in inland waters the ocean where they spawn. Imaging sonar is a reliable and proven technology for monitoring of fish passage migration; however, there no efficient automated method eel detection. We designed deep learning model detection adult data. The employs convolution neural network (CNN) distinguish between 14 images non-eel objects. Prior image classification...
The paradigm shift in the electric power grid necessitates a revisit of existing control methods to ensure grid's security and resilience. In particular, increased uncertainties rapidly changing operational conditions systems have revealed outstanding issues terms either speed, adaptiveness, or scalability for systems. On other hand, availability massive real-time data can provide clearer picture what is happening grid. Recently, deep reinforcement learning (RL) has been regarded adopted as...
This paper presents a novel federated reinforcement learning (Fed-RL) methodology to enhance the cyber resiliency of networked microgrids. We formulate resilient (RL) training setup which (a) generates episodic trajectories injecting adversarial actions at primary control reference signals grid forming (GFM) inverters and (b) trains RL agents (or controllers) alleviate impact injected adversaries. To circumvent data-sharing issues concerns for proprietary privacy in multi-party-owned grids,...
Precisely detecting the fault location on transmission lines can significantly save labor effort and accelerate repairing restoration process. This paper presents a novel single-ended approach for using modern deep learning techniques. A mixed convolutional neural network with long short-term memory (LSTM) structure are trained to predict distance given voltage current measurements. Convolutional function, pooling layers, LSTM used preserve translation invariance capture temporal correlation...
Load shedding has been one of the most widely used and effective emergency control approaches against voltage instability. With increased uncertainties rapidly changing operational conditions in power systems, existing methods have outstanding issues terms either speed, adaptiveness, or scalability. Deep reinforcement learning (DRL) was regarded adopted as a promising approach for fast adaptive grid stability recent years. However, DRL algorithms show two when being applied to system...
In this work we propose a novel data-driven, real-time power system voltage control method based on the physics-informed guided meta evolutionary strategy (ES). The main objective is to quickly provide an adaptive mitigate fault-induced delayed recovery (FIDVR) problem. Reinforcement learning methods have been developed for same or similar challenging problems, but they suffer from training inefficiency and lack of robustness "corner unseen" scenarios. On other hand, extensive physical...
This paper describes an open-source library of transmission-level synchrophasor measurements, curated with the aim accelerating data-driven research and development in power systems domain. dataset contains measurements describing both disturbances ambient conditions, spans two years time, is sourced from electric utilities across United States. Comprised 1694 unique events, this largest repository real phasor measurement unit (PMU) data to date, will be invaluable for benchmarking new...
Deriving generation dispatch is essential for efficient and secure operation of electric power systems. This usually achieved by solving a security-constrained optimal flow (SCOPF) problem, which nature non-convex, nonlinear thus computationally intensive. The state-of-the-art optimization approaches are not able to solve this problem large-scale systems within the system time window (usually 5 minutes). In work, we developed supervised learning determine much shorter window. More...
An automatic system that utilizes data analytics and machine learning to identify adult American eel in obtained by imaging sonars is created this study. Wavelet transform has been applied de-noise the ARIS sonar a convolutional neural network model built classify eels non-eel objects. Because of unbalanced amounts laboratory field experiments, transfer strategy implemented fine-tune so it performs well for both data. The proposed can provide important information develop mitigation...
This letter presents a novel high impedance fault (HIF) detection approach using convolutional neural network (CNN). Compared to traditional artificial networks, CNN offers translation invariance and it can accurately detect HIFs in spite of variance noise the input data. A transfer learning method is used address common challenge system with little training Extensive studies have demonstrated accuracy effectiveness CNNbased for HIF detection.
This paper presents a novel parameter calibration approach for power system stability models using automatic data generation and advanced deep learning technology. A PMU-measurement-based event playback is used to identify potential inaccurate parameters automatically generate extensive simulation data, which are training convolutional neural network (CNN). The accurate will be predicted by the well-trained CNN model validated original PMU measurements. accuracy effectiveness of proposed...
Phasor measurement units can provide high resolution, synchronized measurements of a power system. These result in massive amounts data which is tall over time (30 samples per second) and wide many channels. Data mining techniques are promising tools for quick efficient examination system events with respect to multidimensional correlated data. In this paper, we discuss the implementation performance various classifying common event types associated location. Results presented using test...