检索结果-南通市图书馆

Identifying the validity domain of machine learning models in building energy systems

Energy and AI 2024年第1期15卷 328-341页

作者： Martin Rätz Patrick Henkel phillip stoffel Rita Streblow Dirk Müller RWTH Aachen University E.ON Energy Research CenterInstitute for Energy Efficient Buildings and Indoor ClimateMathieustraβe 10Aachen52074Germany

The building sector significantly contributes to climate *** improve its carbon footprint,applications like model predictive control and predictive maintenance rely on system ***,the high modeling effort hinders practical *** learning models can significantly reduce this modeling *** ensure a machine learning model’s reliability in all operating states,it is essential to know its validity *** states outside the validity domain might lead to extrapolation,resulting in unpredictable *** paper addresses the challenge of identifying extrapolation in data-driven building energy system models and aims to raise knowledge about *** that,a novel approach is proposed that calibrates novelty detection algorithms towards the machine learning *** novelty detection algorithms are identified through a literature review and a benchmark test with 15 candidates.A subset of five algorithms is then evaluated on building energy ***,on two-dimensional data,displaying the results with a novel visualization *** on more complex multi-dimensional use *** methodology performs well,and the validity domain could be *** visualization allows for a profound analysis and an improved understanding of the fundamental effects behind a machine learning model’s validity domain and the extrapolation regimes.

关键词： Extrapolation detection Validity domain Novelty detection Machine learning Artificial neural network Data-driven model predictive control Building energy systems

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Development of a Long-Term Operational Optimization Model for a Building Energy System Supplied by a Geothermal Field

引用

Journal of Thermal Science 2022年第5期31卷 1293-1301页

作者： KÜMPEL Alexander stoffel phillip MÜLLER Dirk RWTH Aachen University E.ON Energy Research CenterInstitute for Energy Efficient Buildings and Indoor ClimateAachen 52074Germany

In order to reduce energy consumption and CO_(2) emissions in the building sector, more and more renewable energy sources are integrated into energy systems. Especially geothermal fields combined with heat pumps are able to supply buildings with heat and cold at low carbon emissions. However, using geothermal fields as heat and cold source influences the ground temperature. Consequently, the ground temperature can change dramatically over a building’s lifetime, leading to less efficient operation of the energy system. Therefore, a sustainable operation is required to ensure the long-term efficiency of geothermal fields. In this paper, we develop an optimization model to derive operating strategies for an efficient long-term operation of a building energy system coupled to a geothermal field. The investigated energy system is the main building of the *** Energy Research Center in Aachen, Germany, which includes a heat pump, two boilers, a combined heat, and power unit, a glycol cooler, and a geothermal field with 41 probes. For each component, we develop energy-based sub-models, which are connected to form the overall system. The geothermal field is modeled by using a g-functions approach as well as a simplified resistance-capacitance approach. To achieve short computing times and realize an optimization horizon of several years, the optimization problem is formulated as mixed-integer linear programming(MILP). The developed model is optimized regarding two different objectives: the minimization of energy costs and the minimization of long-term temperature changes in the ground. Conclusions for an efficient and sustainable operation of the field, especially for the cooling supply, can be derived from the optimization results. It is shown that a state of equilibrium should be aimed to achieve an energy-efficient operation, in which the temperature of the field is close to the initial ground temperature.

关键词： geothermal energy building energy systems operational optimization long-term operation

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Cloud-Based Optimal Control of Individual Borehole Heat Exchangers in a Geothermal Field

引用

Journal of Thermal Science 2022年第5期31卷 1253-1265页

作者： stoffel phillip KÜMPEL Alexander MÜLLER Dirk RWTH Aachen University E.ON Energy Research CenterInstitute for Energy Efficient Buildings and Indoor Climate52074 AachenGermany

Integrating renewable energy sources is a crucial component in reducing CO_(2) emissions in the building sector. In particular, shallow geothermal energy is expected to play a significant role in the regenerative energy supply of buildings. An effective control strategy for the geothermal field is crucial to reduce the overall energy consumption. This paper analyzes the benefits of controlling an existing field’s individual borehole heat exchangers(BHE) using nonlinear model predictive control(NMPC) and moving horizon estimation. The considered geothermal field consists of 41 BHEs and is used for heating and cooling. Each BHE is equipped with temperature sensors for in-and outflow and has individually controllable valves, while a central hydraulic pump feeds all BHEs. The sensor measurements are accessed through a cloud platform, enabling also set point writing for the pump speed and the valve positions. To control the BHEs individually, we propose a two-stage process. In the calibration stage, a moving horizon estimator estimates the actual borehole and ground temperatures for each BHE. In the second stage, first, a nonlinear model predictive controller optimizes the number of active BHEs necessary to meet the buildings’ energy demand. With the estimated ground temperatures as a basis, it is determined which BHEs shall be(de)-activated. The active BHEs are fed with a fixed volume flow of 24 L/min to ensure turbulent heat transfer. To reduce the power usage of the pumps, an optimal control problem based on a simple hydraulic model of the geothermal field is used. The methodology is analyzed through simulations first and then validated experimentally. The results show that half or more of the BHEs could be deactivated most of the time, leading to 67% savings in electricity consumption by the hydraulic pump. The experimental validation confirms the high energy saving potential of the proposed methodology, reducing the consumption of electrical energy by 71%. Additiona

关键词： geothermal energy moving horizon estimation model predictive control operation optimization optimal control

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Safe operation of online learning data driven model predictive control of building energy systems

引用

Energy and AI 2023年第4期14卷 536-549页

作者： phillip stoffel Patrick Henkel Martin Ratz Alexander Kumpel Dirk Muller RWTH Aachen University E.ON Energy Research CenterInstitute for Energy Efficient Buildings and Indoor ClimateMathieustraße 10Aachen52074Germany

Model predictive control is a promising approach to reduce the CO 2 emissions in the building ***,the vast modeling effort hampers the widescale practical ***,data-driven process models,like artificial neural networks,are well-suited to automatize the ***,the underlying data set strongly determines the quality and reliability of artificial neural *** general,the validity domain of a machine learning model is limited to the data that was used to train *** based on system states outside that domain,so-called extrapolations,are unreliable and can negatively influence the control *** present a safe operation approach combined with online learning to deal with extrapolation in data-driven model predictive ***,the k-nearest neighbor algorithm is used to detect extrapolation to switch to a robust fallback *** continuously retraining the artificial neural networks during operation,we successively increase the validity domain of the artificial neural networks and the control *** apply the approach to control a building energy system provided by the BOPTEST *** compare controllers based on two data sets,one with extensive system excitation and one with baseline *** system is controlled to a fixed temperature set point in baseline ***,the artificial neural networks trained on this data set tend to extrapolate in other operating *** show that safe operation in combination with online learning significantly improves performance.

关键词： Data-driven model predictive control Online learning Novelty detection Artificial neural networks Building energy systems

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案：

请选择收藏分类：

通借通还

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

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：