A5021855388- Smart Grid Energy Management
- Meteorological Phenomena and Simulations
- Energy Efficiency and Management
- Model Reduction and Neural Networks
- Computational Physics and Python Applications
- Land Use and Ecosystem Services
- Building Energy and Comfort Optimization
- Energy Load and Power Forecasting
- Hydrological Forecasting Using AI
- Urban Design and Spatial Analysis
- Impact of Light on Environment and Health
- Hydrology and Watershed Management Studies
- Human Mobility and Location-Based Analysis
- Thermal Radiation and Cooling Technologies
- Remote-Sensing Image Classification
- Fluid Dynamics and Turbulent Flows
- Flood Risk Assessment and Management
- Seismology and Earthquake Studies
- Urban Heat Island Mitigation
- Maritime Ports and Logistics
- Urban Transport and Accessibility
- Scientific Computing and Data Management
- Urban and Freight Transport Logistics
- Big Data Technologies and Applications
- Fluid Dynamics and Vibration Analysis
Lawrence Berkeley National Laboratory
2018-2022
Littelfuse (United States)
2020-2021
National Energy Research Scientific Computing Center
2021
Gdańsk Medical University
2021
Massachusetts Institute of Technology
1944-2018
Stanford University
2011-2016
Stanford Medicine
2014-2015
Benjamin Franklin Institute of Technology
1944
Machine learning (ML) provides novel and powerful ways of accurately efficiently recognizing complex patterns, emulating nonlinear dynamics, predicting the spatio-temporal evolution weather climate processes. Off-the-shelf ML models, however, do not necessarily obey fundamental governing laws physical systems, nor they generalize well to scenarios on which have been trained. We survey systematic approaches incorporating physics domain knowledge into models distill these broad categories....
Abstract. Recently, deep learning (DL) has emerged as a revolutionary and versatile tool transforming industry applications generating new improved capabilities for scientific discovery model building. The adoption of DL in hydrology so far been gradual, but the field is now ripe breakthroughs. This paper suggests that DL-based methods can open up complementary avenue toward knowledge hydrologic sciences. In avenue, machine-learning algorithms present competing hypotheses are consistent with...
While deep learning has shown tremendous success in a wide range of domains, it remains grand challenge to incorporate physical principles systematic manner the design, training, and inference such models. In this paper, we aim predict turbulent flow by its highly nonlinear dynamics from spatiotemporal velocity fields large-scale fluid simulations relevance turbulence modeling climate modeling. We adopt hybrid approach marrying two well-established simulation techniques with learning....
Urban planning applications (energy audits, investment, etc.) require an understanding of built infrastructure and its environment, i.e., both low-level, physical features (amount vegetation, building area geometry etc.), as well higher-level concepts such land use classes (which encode expert socio-economic end uses). This kind data is expensive labor-intensive to obtain, which limits availability (particularly in developing countries). We analyze patterns urban neighborhoods using...
Scaling area-based conservation, including through initiatives ed or co-managed by Indigenous Peoples and local communities, is a flagship goal of the Kunming-Montreal Global Biodiversity Framework. Conservationists often aspire to scale initiatives, but this rarely achieved in practice. Identifying addressing “bottlenecks” – factors that limit initiative adoption could help shape more effective scaling strategies. Therefore, we integrate insights from 84 experts with existing evidence...
This work describes a methodology for informing targeted demand-response (DR) and marketing programs that focus on the temperature-sensitive part of residential electricity demand. Our uses data is becoming readily available at utility companies-hourly energy consumption readings collected from "smart" meters, as well hourly temperature readings. To decompose individual into thermal-sensitive base load (non-thermally-sensitive), we propose model response based thermal regimes, i.e.,...
Radiative particles are ubiquitous in nature and various technologies. Calculating radiative properties from known geometry designs can be computationally expensive, trying to invert the problem come up with specific desired is even more challenging. Here, we report a machine-learning (ML)-based method for both forward inverse dielectric metallic particles. Our decision-tree-based model able provide explicit design rules problems. Furthermore, use same trained problem, which greatly...
In this study we propose a new method to simulate hyper-realistic urban patterns using Generative Adversarial Networks trained with global land-use inventory. We generated synthetic "universe" that qualitatively reproduces the complex spatial organization observed in patterns, while being able quantitatively recover certain key high-level metrics.
For demand-side management programs concerned with heating, ventilation, and air conditioning (HVAC) energy consumption, smart meter data collected at the whole-premise level has recently been used to decompose usage into its HVAC non-thermal components, which are typically not separately monitored. In this paper, we study extent program design decisions based on models using whole-home consumption differ from made full knowledge of appliance-level end-use patterns. We develop a model...
This paper describes Steptacular, an online interactive incentive system for encouraging people to walk more. A trial offering Steptacular the employees of Accenture-USA was conducted over a 6 month period. Over 5,000 registered program and close 3,000 participants wore USB-enabled pedometers; from time they plugged their pedometer into computer upload hourly step counts website; website had range features encourage more walking. These included monetary rewards which were randomly redeemable...
As a way to match peaks in demand available supply real-time on the power grid, energy utility companies employ Demand-Response (DR) strategies. With recent deployment of advanced metering infrastructure collecting highly granular (sub-hourly) data consumption from millions users system operators may now understand how arises down individual level. In this paper we present an application dynamic model that describes residential users' thermally-sensitive using hourly electricity and weather...
Existing electricity market segmentation analysis techniques only make use of limited consumption statistics (usually averages and variances). In this paper we power demand distributions (PDDs) obtained from fine-grain smart meter data to perform based on distributional clustering. We apply approach mining 8 months readings about 1000 US Google employees.
Uncertainty in consumption is a key challenge at energy utility companies, which are faced with balancing highly stochastic demand increasingly volatile supply characterized by significant penetration rates of intermittent renewable sources. This paper proposes methodology to quantify uncertainty that highlights the dependence cost-of-service volatility demand. We use large and rich dataset time series provide evidence there substantial degree high-level structure statistics across users may...
Research on freight transportation has seen a tremendous increase in the last decades, yet it still lags behind that passenger travel, particularly at macro-level suitable for nation-wide policy analysis. A key challenge demand modeling is availability of data drivers - such as cost, time, and trip length which usually proprietary expensive. Moreover available to public heterogeneous published by number different bodies. In this study we integrate many publicly-available datasets these...
Abstract. Recently, deep learning (DL) has emerged as a revolutionary and versatile tool transforming industry applications generating new improved capabilities for scientific discovery model building. The adoption of DL in water science so far been gradual, but the related fields are now ripe breakthroughs. This paper proposes that DL-based methods can open up viable, complementary avenue toward knowledge hydrologic sciences. In avenue, machine-learning algorithms present competing...
Accurately forecasting urban development and its environmental climate impacts critically depends on realistic models of the spatial structure built environment, dependence key factors such as population economic development. Scenario simulation sensitivity analysis, i.e., predicting how changes in underlying at a given location affect urbanization outcomes other locations, is currently not achievable large scale with traditional growth models, which are either too simplistic, or depend...