检索结果-内蒙古大学图书馆

57th Hawaii International Conference on System Sciences (HICSS)

作者： Chevalier, Samuel Murzakhanov, Ilgiz Chatzivasileiadis, Spyros Univ Vermont Burlington VT 05405 USA Tech Univ Denmark Lyngby Denmark

ISBN: (纸本)9780998133171

Computational tools for rigorously verifying the performance of large-scale machine learning (ML) models have progressed significantly in recent years. The most successful solvers employ highly specialized, GPU-accelerated branch and bound routines. Such tools are crucial for the successful deployment of machine learning applications in safety-critical systems, such as power systems. Despite their successes, however, barriers prevent out-of-the-box application of these routines to power system problems. This paper addresses this issue in three key ways. First, we reformulate several key power system verification problems into the canonical format utilized by modern verification solvers. Second, we enable the simultaneous verification of multiple verification problems (e.g., checking for the violation of all constraints simultaneously, and not by solving individual verification problems). To achieve this, we introduce an exact transformation that converts a set of potential violations into a series of ReLU-based neural network layers. This allows verifiers to interpret these layers directly, and determine the "worst-case" violation in a single shot. Third, power system ML models often must be verified to satisfy power flow constraints. We propose a dualization procedure which encodes linear equality and inequality constraints (such as power balance constraints and line flow constraints) directly into the verification problem in a manner which is mathematically consistent with the specialized verification tools. To demonstrate these innovations, we verify problems associated with data-driven security constrained DC-OPF solvers. We build and test our first set of innovations using the alpha, beta-CROWN solver, and we benchmark against Gurobi 10.0. Our contributions achieve a speedup that can exceed 100x and allow higher degrees of verification flexibility.

关键词： Branch and bound data driven modeling DC-OPF neural network verification machine learning

来源：评论

学校读者我要写书评

暂无评论

data-driven dynamic state estimation in power systems via sparse regression unscented Kalman filter

引用

SUSTAINABLE ENERGY GRIDS & NETWORKS 2024年 40卷

作者： Jamalinia, Elham Khazaei, Javad Blum, Rick S. Lehigh Univ Dept Elect & Comp Engn Bethlehem PA 18015 USA

This paper proposes a novel data-driven modeling and dynamic state-estimation approach for nonlinear power and energy systems, highlighting the critical role of a known dynamic model for accurate state estimation in the face of uncertainty and complex models. The proposed framework consists of a two-phase approach: data-driven model identification and state-estimation. During the model identification phase, which spans a relatively short time interval, state feedback is collected to identify the dynamics of the nonlinear systems in the power grid using a novel density-guided sparse identification algorithm. Unlike conventional sparse regression, which relies on a large library of linear and nonlinear functions to fit data, our proposed algorithm iteratively updates a relatively small initial library by adding higher-order nonlinear functions if the coefficients of the current functions are dense. Following the identification of the model's dynamics, the estimation phase addresses the challenge of incomplete state measurements. By implementing an unscented Kalman filter, the state variables of the system are dynamically estimated by measuring the noisy output. Finally, simulation results on an IEEE 30-bus system are presented to illustrate the effectiveness of the density-guided sparse regression unscented Kalman filter compared to a physics-based unscented Kalman filter with model uncertainty. This study contributes to the fields of data-driven modeling techniques, machine learning for power systems, and computational intelligence in smart grids. It emphasizes the use of advanced sparse regression and unscented Kalman filter methods for state estimation, enhancing the robustness and accuracy of monitoring and control in electrical and energy systems.

关键词： data driven modeling Dynamic state estimation Power and energy systems Unscented Kalman filter

来源：评论

学校读者我要写书评

暂无评论

Assessing the impact of borehole field data on AI-based deep learning models for heating and cooling prediction

引用

GEOTHERMICS 2024年 117卷

作者： Ahmed, Naveed Assadi, Mohsen Ahmed, Abdelazim Abbas Banihabib, Reyhaneh Zhang, Qian Univ Stavanger Dept Energy & Petr Engn N-4021 Stavanger Norway Natl Univ Sci & Technol NUST US Pakistan Ctr Adv Studies Energy USPCASE H-12 Sect Islamabad 44000 Pakistan

Sensor readings play a critical role in the prediction performance of AI-based data-driven fault-detection and diagnostics for borehole heat exchangers in heating and cooling systems. The accuracy and algorithm selection of individual deep neural networks are well-investigated research topics but the impact of quality of sensors data on algorithm selection and its further impact on prediction performance has not been investigated and quantified for geothermal heating/cooling systems. The objective of the current work is to investigate the impact of data quality, available as measured time-series data from the borehole field, on the prediction accuracy of the deep neural networks-based models. In the current work four different data sets (D-1, D-2, D-3, D-4) are used for datadriven modeling using four different deep machine learning approaches, i.e., LSTM, BD-LSTM, GRU and CNN. The developed AI models have been trained, validated, and tested using the real-field data sets with same number of independent input parameters i.e., inlet temperature and mass flow rate to predict the fluid outlet temperature. The validated AI models are used to predict the performance of borehole heat exchanger to meet the heating and cooling loads. The prediction results show that the BD-LSTM outperforms other deep neural networks in terms of accuracy. With the data sets D-1, D-2, D-3 and D-4 it was observed that BD-LSTM had lowest MAPE of 2.99 %, 0.22 %, 6.39 % and 2.21 % respectively. Several experiments show that the BD-LSTM and LSTM models are better than the other used approaches in terms of prediction accuracy, but larger number of trainable parameters results in more computation time than CNN and GRU respectively. More data points with higher sampling frequency, results in more accurate prediction as depicted by the results from D-2 and D-4. Although, the results showed that the data collection interval is crucial, but different test runs indicate that avoiding the loss of time-seri

关键词： Ground source heat pump Borehole coupled heat pump Heating and cooling Prediction model data driven modeling Artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

A data-driven epidemic model with social structure for understanding the COVID-19 infection on a heavily affected Italian province

引用

MATHEMATICAL MODELS & METHODS IN APPLIED SCIENCES 2021年第12期31卷 2533-2570页

作者： Zanella, Mattia Bardelli, Chiara Dimarco, Giacomo Deandrea, Silvia Perotti, Pietro Azzi, Mara Figini, Silvia Toscani, Giuseppe Univ Pavia Dept Math Pavia Italy Univ Pavia Dept Polit & Social Sci Pavia Italy Univ Ferrara Dept Math & Informat Ferrara Italy Hlth Protect Agcy ATS Viale Indipendenza 3 I-27100 Pavia Italy Univ Pavia Inst Appl Math & Informat Dept Math Enrico Magenes CNR Pavia Italy

In this work, using a detailed dataset furnished by National Health Authorities concerning the Province of Pavia (Lombardy, Italy), we propose to determine the essential features of the ongoing COVID-19 pandemic in terms of contact dynamics. Our contribution is devoted to provide a possible planning of the needs of medical infrastructures in the Pavia Province and to suggest different scenarios about the vaccination campaign which possibly help in reducing the fatalities and/or reducing the number of infected in the population. The proposed research combines a new mathematical description of the spread of an infectious diseases which takes into account both age and average daily social contacts with a detailed analysis of the dataset of all traced infected individuals in the Province of Pavia. These information are used to develop a data-driven model in which calibration and feeding of the model are extensively used. The epidemiological evolution is obtained by relying on an approach based on statistical mechanics. This leads to study the evolution over time of a system of probability distributions characterizing the age and social contacts of the population. One of the main outcomes shows that, as expected, the spread of the disease is closely related to the mean number of contacts of individuals. The model permits to forecast thanks to an uncertainty quantification approach and in the short time horizon, the average number and the confidence bands of expected hospitalized classified by age and to test different options for an effective vaccination campaign with age-decreasing priority.

关键词： Epidemic models disease control social contacts data analysis data driven modeling nonlinear incidence rate uncertainty quantification vaccination campaign healthcare system

来源：评论

学校读者我要写书评

暂无评论

data driven surrogate model-based optimization of the process parameters in electric discharge machining of D2 steel using Cu-SiC composite tool for the machined surface roughness and the tool wear

引用

REVISTA DE METALURGIA 2023年第2期59卷

作者： Somani, Nalin Walia, Arminder Singh Gupta, Nitin Kumar Panda, Jyoti Prakash Das, Anshuman Das, Sudhansu Ranjan DIT Univ Dept Mech Engn Dehra Dun India Thapar Polytech Coll Patiala Punjab India VSSUT Burla Dept Prod Engn Burla Odisha India

Electrical discharge machining (EDM) is mainly utilized for the die manufacturing and also used to machine the hard materials. Pure Copper, Copper based alloys, brass, graphite, steel are the conventional electrode materials for EDM process. While machining with the conventional electrode materials, tool wear becomes the main bottleneck which led to increased machining cost. In the present work, the composite tool tip comprises 80% Copper and 20% silicon carbide was used for the machining of hardened D2 steel. The powder metallurgy route was used to fabricate the composite tool tip. Electrode wear rate and surface roughness were assessed with respect to the different process parameters like input current, gap voltage, pulse on time, pulse off time and dielectric flushing pressure. During the analysis it was found that Input current (Ip), Pulse on time (Ton) and Pulse off time (Toff) were the significant parameters which were affecting the tool wear rate (TWR) while the Ip, Ton and flushing pressure affected more the surface roughness (SR). SEM micrograph reveals that increase in Ip were used to develop machine learning based surrogate models. Three machine learning (ML) models are random forest, polynomial regression and gradient boosted tree. The predictive capability of ML based surrogate models was assessed by contrasting the R2 and mean square error (MSE) of prediction of responses. The best surrogate model was used to develop a complex objective function for use in firefly algorithm-based optimization of input machining parameters for minimization of the output responses. leads to increase in the wear rate of the tool. The data obtained from experiments

关键词： data driven modeling Electric discharge machining Firefly algorithm Machine learning Surface roughness Tool wear

来源：评论

学校读者我要写书评

暂无评论

Reservoir interwell connectivity estimation from small datasets using a probabilistic data driven approach and uncertainty quantification

GEOENERGY SCIENCE AND ENGINEERING

引用

GEOENERGY SCIENCE AND ENGINEERING 2023年 230卷

作者： Thiam, Moustapha Nakhaee, Ali Univ Tehran Inst Petr Engn Coll Engn Sch Chem Engn Tehran Iran

Over the past decade, various artificial intelligence types have made significant progress in petroleum reservoir modeling, from machine learning to deep learning. These data-driven proxies mostly use the available data as features and targets to find patterns between them or make a function approximation. This study mainly sum-marizes the estimation of reservoir inter-well connectivity by combining physics and data-driven. In addition, polynomial chaos expansion (PCE) has been used to quantify the uncertainty of inter-well connectivity. The contribution of this work is dedicated to a probabilistic data-driven, namely Bayesian inference applied to an artificial neural network and a reduced physics called capacitance resistance modeling (CRM). The main ad-vantages of using such a probabilistic model are the quantification of uncertainty, which is well suited for decision-making in small datasets. As well as prior information, BBNs are often more appropriately calibrated than a classical neural network since its predicted uncertainty is more similar to the observed errors, and it is not necessary to use separate cross-validation. We implement a Markov chain Monte Carlo sampling algorithm;this approach evaluates the CRM parameters over time under waterflooding - dynamic parameters determination. Additionally, we have conducted a statistical inference to quantify the uncertainties associated with BNN modeling. Maximum a posteriori, a Bayesian optimization method, has also been used to compare estimated parameters by history-matching reservoirs' liquid production rates. An evaluation of the effectiveness of the proposed method was conducted on a synthetic and real reservoir consisting of a small and medium dataset. The results show that, when a small dataset is used, BBNs match the cumulative liquid production rate more accu-rately than MAP inference and the deterministic CRM, where an error of 5.87%, 10.94%, and 9.38% is achieved, respectively. In the case of the medium

关键词： Waterflooding Interwell connectivity coeffi cients data driven modeling Reduced physics modeling Polynomial chaos expansion History matching Artificial neural network

来源：评论

学校读者我要写书评

暂无评论

DNN modeling of partial differential equations with incomplete data

引用

JOURNAL OF COMPUTATIONAL PHYSICS 2023年 493卷

作者： Churchill, Victor Chen, Yuan Xu, Zhongshu Xiu, Dongbin Trinity Coll Dept Math Hartford CT 06106 USA Ohio State Univ Dept Math Columbus OH 43210 USA

We present a computational technique for modeling the evolution of partial differential equations (PDEs) with incomplete data. It is a significant extension of the recent work of data driven learning of PDEs, in the sense that we consider two forms of partial data: data are observed only on a subset of the domain, and data are observed only on a subset of the state variables. Both cases resemble more realistic data collection scenarios in real -world applications. Leveraging the recent work on modeling partially-observed dynamical systems, we present a deep neural network (DNN) structure that is suitable for PDE modeling with such kinds of incomplete data. In addition to the mathematical motivation for the DNN structure, we present an extensive set of numerical examples in both one -and two-dimensions to demonstrate the effectiveness of the proposed DNN modeling. In one example, the method can accurately predict the solution when data are only available in less than half (40%) of the domain.(c) 2023 Elsevier Inc. All rights reserved.

关键词： data driven modeling Deep neural networks Reduced PDE systems Incomplete data

来源：评论

学校读者我要写书评

暂无评论

A novel data-characteristic-driven modeling methodology for nuclear energy consumption forecasting

引用

APPLIED ENERGY 2014年 128卷 1-14页

作者： Tang, Ling Yu, Lean He, Kaijian Beijing Univ Chem Technol Sch Econ & Management Beijing 100029 Peoples R China

Due to the unique features of nuclear energy market, this paper tries to propose a novel data-characteristic-driven modeling methodology based on the principle of "data-characteristic-driven modeling", aiming at formulating appropriate forecasting model closely in terms of sample data's own data characteristics. In the novel data-characteristic-driven modeling methodology, two steps are mainly involved, i.e., data analysis and forecasting modeling. First, the sample data of nuclear energy consumption are thoroughly investigated in order to capture the main inner rules and hidden patterns driving the data dynamics, in terms of data characteristics. Second, the corresponding forecasting model is accordingly formulated and designed based on these data characteristics. For illustration and verification purposes, the proposed methodology is implemented to predict the nuclear energy consumption of USA and China. The empirical results demonstrate that the novel methodology with the principle of "data-characteristic-driven modeling" strikingly improves prediction performance, since the models elaborately built based on data characteristics statistically outperform all other benchmark models without consideration of data characteristics. This further confirms that the proposed methodology is a very promising tool in both analyzing and forecasting nuclear energy consumption. (C) 2014 Elsevier Ltd. All rights reserved.

关键词： data characteristics Nuclear energy consumption Time series forecasting data driven modeling

来源：评论

学校读者我要写书评

暂无评论

Investigating the impact of borehole field data's input parameters on the forecasting accuracy of multivariate hybrid deep learning models for heating and cooling

引用

ENERGY AND BUILDINGS 2023年 301卷

作者： Ahmed, Naveed Assadi, Mohsen Zhang, Qian Univ Stavanger Dept Energy & Petr Engn N-4036 Stavanger Norway Natl Univ Sci & Technol NUST US Pakistan Ctr Adv Studies Energy USPCASE H-12 Sect Islamabad 44000 Pakistan

Heat pump systems coupled with Borehole heat exchanger (BHEx) are crucial for the efficient and cost-effective heating and cooling of buildings. Accurate performance forecasting of BHEx can help to reduce energy consumption, optimize system design, improve system reliability and ultimately contribute to the sustainability of the built environment. The performance of deep neural networks, including hybrid deep learning algorithms, is dependent on the number and true identification of input parameters used to train the model. While research on the accuracy and selection of individual deep neural networks is well-investigated, the impact of selected input parameters on hybrid algorithm selection and its subsequent influence on multi-step forecasting performance has not been quantified specifically for geothermal heating/cooling systems. The focus of the current work is to investigate and quantify this impact on multivariate hybrid deep learning algorithms for forecasting borehole heat exchanger outlet temperature. In the current work the developed hybrid AI models (LSTM-CNN, GRU-CNN, CNN-LSTM, CNN-GRU) have been trained, validated, and tested using the measured borehole-field data with variable number of input parameters (3,4,5 and 6). The validated AI models are used to forecast 24 h ahead performance of borehole heat exchanger to meet the heating and cooling loads. The most essential input parameters for the data-driven AI models are determined via an importance-level analysis by using Explainable AI techniques of Shapley Additive Explanations (SHAP). The findings show that increasing the number of input parameters has the potential to enhance the multi-step forecasting accuracy of the examined deep learning models. However, it is crucial to note that there is a point where including additional input parameters may not lead to further improvements. Therefore, it is essential to identify the optimal number of input parameters that can maximize the accuracy of the hybr

关键词： Borehole coupled heat pump Heating and cooling Ground source heat pump Prediction model Hybrid forecasting models data driven modeling

来源：评论

学校读者我要写书评

暂无评论

Next Generation Public Supply Water Withdrawal Estimation for the Conterminous United States Using Machine Learning and Operational Frameworks

引用

WATER RESOURCES RESEARCH 2024年第7期60卷

作者： Alzraiee, Ayman Niswonger, Richard Luukkonen, Carol Larsen, Josh Martin, Donald Herbert, Deidre Buchwald, Cheryl Dieter, Cheryl Miller, Lisa Stewart, Jana Houston, Natalie Paulinski, Scott Valseth, Kristen US Geol Survey USGS Calif Water Sci Ctr Sacramento CA 95819 USA USGS Water Mission Area Menlo Pk CA USA USGS Upper Midwest Water Sci Ctr Lansing Lansing MI USA USGS Maryland Delaware DC Water Sci Ctr Catonsville MD USA USGS Colorado Water Sci Ctr San Diego CA USA USGS Oklahoma Texas Water Sci Ctr Austin TX USA

Estimation of human water withdrawals is more important now than ever due to uncertain water supplies, population growth, and climate change. Fourteen percent of the total water withdrawal in the United States is used for public supply, typically including deliveries to domestic, commercial, and occasionally including industrial, irrigation, and thermoelectric water withdrawal. Stewards of water resources in the USA require estimates of water withdrawals to manage and plan for future demands and sustainable water supplies. This study compiled the most comprehensive conterminous United States water withdrawal data set to date and developed a machine learning framework for estimating public supply withdrawals and associated uncertainty for the period 2000-2020. The modeling approach provides service area resolution estimates to allow for annual and monthly water withdrawal estimation while incorporating a complex array of driving factors that include hydroclimatic, demographic, socioeconomic, geographic, and land use factors. Model results reveal highly variable and lognormally distributed per-capita water withdrawal, spanning from 30 to 650 gallons per capita per day (GPCD), across community, regional, and national scales, with pronounced seasonal variations. Analysis of estimated withdrawal trends indicates that the national annual average withdrawal experienced a decline at a rate of 0.58 GPCD/year during the period from 2000 to 2020. Model interpretation reveals a complex interplay between public supply withdrawal and key predictors, including population size, warm-season precipitation, counts of large buildings and houses, and areas of urban and commercial land use. The developed models can forecast future public supply driven by various climate, demographic, and socioeconomic scenarios. Consistent, reproducible estimates of monthly water withdrawals were generated at the spatial resolution of USA water service areas Spatial and temporal variability of public sup

关键词： national water use public supply machine learning uncertainty water withdrawal data driven modeling

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：