检索结果-内蒙古大学图书馆

Physics Informed Piecewise linear Neural Networks for Process Optimization

COMPUTERS & CHEMICAL ENGINEERING 2023年第1期174卷

作者： Koksal, Ece Serenat Aydin, Erdal Koc Univ Dept Chem & Biol Engn TR-34450 Istanbul Turkiye Koc Univ TUPRAS Energy Ctr KUTEM TR-34450 Istanbul Turkiye

Constructing first-principles models is usually a challenging and time-consuming task due to the complexity of real-life processes. On the other hand, data-driven modeling, particularly a neural network model, often suffers from overfitting and lack of useful and high-quality data. At the same time, embedding trained machine learning models directly into the optimization problems has become an effective and state-of-the-art approach for sur-rogate optimization, whose performance can be improved by physics-informed machine learning. This study proposes using piecewise linear neural network models with physics-informed knowledge for optimization problems with neural network models embedded. In addition to using widely accepted and naturally piecewise linear rectified linear unit (ReLU) activation functions, this study also suggests piecewise linear approximations for the hyperbolic tangent activation function to widen the domain. Optimization of three case studies, a blending process, an industrial distillation column, and a crude oil column are investigated. Physics-informed trained neural network-based optimal results are closer to global optimality for all cases. Finally, associated CPU times for the optimization problems are much shorter than the standard optimization results.

关键词： Machine learning artificial neural networks physics-informed neural networks piecewise linear approximation mixed integer linear programming

来源：评论

学校读者我要写书评

暂无评论

Optimal design of PV and hybrid storage based microgrids for healthcare and government facilities connected to highly intermittent utility grids

引用

APPLIED ENERGY 2023年 335卷

作者： Leon, L. M. Romero-Quete, D. Merchan, N. Cortes, C. A. Univ Nacl Colombia Elect & Elect Engn Dept Carrera 45 Num 26-85 Bogota 111321 Colombia

This paper proposes an improved methodology for the optimal sizing of small-scale microgrids conformed by photovoltaic (PV) generation systems and hybrid energy storage systems (batteries + supercapacitors). The methodology is based on a mixed-integer linear programming (MILP) formulation and includes a new way to take advantage of the aging model of the batteries to determine when these storage assets should be replaced. Moreover, the proposed sizing tool offers the possibility of including supercapacitors (SCs) in a microgrid's design to reduce battery degradation. The tool was put to test by sizing the microgrids for a hospital and a government building in Puerto Carrello, a non-interconnected area in Colombia. A comparison between the proposed methodology and another without considering battery degradation, along with a sensitivity analysis of the storage-acquisition cost parameter in this tool, revealed its usefulness for this kind of microgrid design.

关键词： Critical loads Generation unavailability mixed integer linear programming Non-interconnected areas Optimal sizing Small-scale microgrid

来源：评论

学校读者我要写书评

暂无评论

Style: A new optimization model for Synthesis of uTility sYstems with steam LEvel placement

引用

COMPUTERS & CHEMICAL ENGINEERING 2023年 170卷

作者： Jimenez-Romero, Julia Azapagic, Adisa Smith, Robin Univ Manchester Ctr Proc Integrat Dept Chem Engn Manchester M13 9PL Lancs England Univ Manchester Dept Chem Engn Sustainable Ind Syst Grp Manchester M13 9PL Lancs England

Moving industrial production to a more sustainable basis requires a step change in the efficiency of site steam and cogeneration systems, coupled with a switch away from fossil fuels to renewable energy sources. This paper presents the first step in the development of an optimization framework capable of identifying a roadmap to move utility systems to a more sustainable basis. Previous approaches to the design and optimization of steam and cogeneration systems have simplified the problem to the extent that many important practical issues have been neglected, restricting the scope of the options included. The use of grossly simplified models has been prompted by the mathematical difficulties of optimizing such complex energy systems. To overcome limitations in previous work and provide a sound basis for future developments, this paper proposes a new superstructure-based optimization model for the optimization of utility systems, accounting for optimum steam level placement. The latter is important for improving systems efficiency and reducing energy consumption. The optimization problem involves the selection of more realistic operating conditions of the steam mains (superheat conditions and pressure). The model accounts for water preheating, as well as superheating and de-superheating for process steam generation and use. Hot oil circuits are also included as hot utility option to overcome potential steam temperature and/or pressure limitations at high temperatures, and with it provide more flexibility in the framework. The general problem requires making several continuous and discrete decisions, where non-linearities and non-convexities from underlying physics and binary decisions exacerbate the complex nature of the problem, yielding a nonconvex mixed integer Non-linear programming (MINLP) formulation. However, MINLP formulations can become computationally intractable to solve. Thus, to guarantee tractability and fast conversion, the STYLE model develops a novel

关键词： Superstructure mixed integer linear programming Indirect heat recovery Industrial energy systems Steam systems

来源：评论

学校读者我要写书评

暂无评论

Bi-objective optimization for a two-depot automated storage/retrieval system

引用

ANNALS OF OPERATIONS RESEARCH 2021年第1-2期296卷 243-262页

作者： Man, Xiaoyi Zheng, Feifeng Chu, Feng Liu, Ming Xu, Yinfeng Donghua Univ Glorious Sun Sch Business & Management Shanghai Peoples R China Fuzhou Univ Sch Econ & Management Fuzhou Peoples R China Univ Paris Saclay Univ Evry F-91025 Evry France Tongji Univ Sch Econ & Management Shanghai Peoples R China

Operation management of automated storage and retrieval system (AS/RS) has a great impact on system performance and is a hot research topic. Most existing works for AS/RS operation management address determining storage and retrieval (S/R) machine sequence and minimizing its travel time. In the paper, we study a new bi-objective S/R machine sequencing problem with task release time and due date in a two-depot AS/RS. The objective is to minimize the total travel time of the S/R machine and the total tardiness simultaneously. For the problem, a new bi-objective mixed integer linear programming model is established. Based on problem property analysis, an exact epsilon-constraint method, a non-dominated sorting genetic algorithm II and a promising heuristic are devised for the problem. Especially, the computational results show the efficiency and effectiveness of the proposed algorithms.

关键词： Container terminal Automated storage retrieval system Two-depot Bi-objective optimization mixed integer linear programming Algorithms

来源：评论

学校读者我要写书评

暂无评论

Impact of multi-annual renewable energy variability on the optimal sizing of off-grid systems

引用

RENEWABLE & SUSTAINABLE ENERGY REVIEWS 2023年第1期183卷

作者： Javed, Muhammad Shahzad Jurasz, Jakub Guezgouz, Mohammad Canales, Fausto A. Ruggles, Tyler H. Ma, Tao Shanghai Jiao Tong Univ Engn Res Ctr Solar Energy & Refrigerat Sch Mech Engn Shanghai 200240 Peoples R China Wroclaw Univ Sci & Technol Fac Environm Engn PL-50370 Wroclaw Poland Malardalen Univ Future Energy Ctr Sch Business Soc & Engn S-72123 Vasteras Sweden Univ Costa Dept Civil & Environm Calle 58 55-66 Barranquilla 080002 Colombia Carnegie Inst Sci Dept Global Ecol Stanford CA 94305 USA

It remains a significant technical and economic challenge to fully power large-scale grids with intermittent renewable energy (RE). Meanwhile, due to the rapid decrease in the cost of RE power generation technologies in recent years, the number of real-world implementations and studies dedicated to the optimal capacity sizing of renewable off-grid systems has increased. However, a common approach in the literature is to rely on typical single-year meteorological and demand data. A negative effect of this assumption is that it does not consider the RE inter-annual variability, which might cause blackouts or oversizing the system and large curtailments. This study employs 43 years of hourly solar, wind, and demand data, coupled with different microgrid configurations, to evaluate the impact of diverse simulation periods on the total system cost, optimal RE mix, and system reliability. Our findings indicate that extended simulation periods considerably increased renewable energy systems (RES) reliability and that the resulting configurations can be up to 94% more robust than those obtained using a single year of data. Additionally, the optimal energy storage requirements increased when considering longer simulation periods, indicating that short simulation periods could underestimate energy storage capacities in off-grid systems. The overestimations or underestimations resulting from optimizations based on single-year data directly affect the long-term sustainability, reliability, and cost-effectiveness of the RES.

关键词： Variable renewable energy Pumped hydro storage Battery storage mixed integer linear programming Long-term modeling horizons Off-grid

来源：评论

学校读者我要写书评

暂无评论

Electric fleet charging management considering battery degradation and nonlinear charging profile

引用

ENERGY 2023年 283卷

作者： Shi, Junzhe Zeng, Teng Moura, Scott Univ Calif Berkeley Dept Civil & Environm Engn Berkeley CA 94720 USA Univ Calif Berkeley Berkeley CA 94720 USA

The populations of commercial electric vehicles (EVs) and electric robots (ERs) have been growing rapidly in recent years. Yet, the availabilities and incoordination of the charging infrastructure still constrain the operations of all EVs/ERs, resulting in wasted waiting time and, thus, decreased total profits. Coordinating these electric machines as a fleet and identifying the optimal operation and charging schedules to maximize total profit is essential. On the other hand, the charging process usually consists of two charging stages, constant current (CC) and constant voltage (CV), which lead to a nonlinear charging profile. Other factors, such as the high charging current, may significantly accelerate battery degradation and lead to capacity fade. However, the high nonlinearities make the battery charging profile and the degradation model computationally difficult to be integrated into optimization problems. In this study, we propose an innovative fleet management strategy that maximizes the operation revenue and minimizes the cost of electricity and battery degradation while addressing the aforementioned nonlinear charging profile. By proposing two linearization methods to replace the nonlinear parts, we formulated a mixed-integer linear Program (MILP). Furthermore, stemming from the numerical case study, two managerial insights, the impact of the battery SOH on fleet management and the selection of fast charging vs. normal charging modes, are outlined.

关键词： Fleet management Li-ion battery Battery degradation modeling Optimal charging Nonlinear charging profile mixed integer linear programming

来源：评论

学校读者我要写书评

暂无评论

An Adaptive Large Neighbourhood Search algorithm for a real-world Home Care Scheduling Problem with time windows and dynamic breaks

引用

COMPUTERS & OPERATIONS RESEARCH 2023年第1期159卷

作者： Mendez-Fernandez, Isabel Lorenzo-Freire, Silvia Manuel Gonzalez-Rueda, Angel Univ A Coruna Dept Math MODES Res Grp Campus Elvina La Coruna Spain Univ A Coruna Fac Comp Sci Ctr Informat & Commun Technol Res CIT Campus Elvina La Coruna Spain CITMAga Santiago De Compostela 15782 Spain

This paper presents a Home Care Scheduling Problem (from now on HCSP) based on a real case of a care company for elderly and dependent people located in the North of Spain. The problem incorporates many of the common features addressed in the HCSP literature, such as soft and hard time windows for the services, available working time of the caregivers or affinity levels between users and caregivers. However, it also includes other novel characteristics that increase the difficulty of the problem significantly, since the breaks between services will play a key role in the quality of the solutions. To evaluate the solutions, the users welfare will be prioritized over the cost associated with the *** problem has been formulated as a mixed integer linear programming (MILP) one but, due to the complexity of the model, it is not possible to solve it for real size instances. Therefore, we have designed a method that combines the Adaptive Large Neighbourhood Search (ALNS) methodology with a heuristic approach necessary to evaluate the objective functions. To analyse the behaviour of the algorithm, a set of computational experiments are carried out under different configurations. First, the MILP formulation and the algorithm have been compared over some standard instances from the literature. Finally, the performance of the algorithm is evaluated over a real case study based on the timetables of the company during some weeks from 2016 to 2017.

关键词： Home care Scheduling Adaptive large neighbourhood search mixed integer linear programming

来源：评论

学校读者我要写书评

暂无评论

The bike routeing problem with energy constraints

引用

INTERNATIONAL JOURNAL OF SYSTEMS SCIENCE-OPERATIONS & LOGISTICS 2024年第1期11卷

作者： Ancele, Yannis Pham, Quang Anh Ha, Minh Hoang Matellini, Dante Ben Nguyen, Trung Thanh Liverpool John Moores Univ Offshore & Marine LOOM Res Inst Liverpool Logist Liverpool England Natl Econ Univ ORLab SLSCM Fac Data Sci & Artificial Intelligence Hanoi Vietnam

As climate change becomes more crucial, transporting products in urban areas by bicycle gains popularity. More companies start using bicycles as an alternative transportation mode and face challenges to efficiently satisfy the needs of their customers and employees. While designing the bike routes for pick up and delivery, it is required to take into account the energy needed by cyclists to move. The energy consumed in a bike route has to be kept under a certain threshold for cyclists to be able to pedal during the whole work shift. This leads to a new variant of the vehicle routeing problem called the bike routeing problem which aims at tackling constraints arising for bicycle deliveries. We propose a novel mixed integer linear programming model to determine the bike routes for delivering goods in urban areas. An Evolutionary Local Search algorithm is developed to efficiently solve the problem using new split and local search procedures. Experimental results obtained on random and real instances show the accuracy and stability of the proposed algorithms, as well as the relevance of the new problem.

关键词： Vehicle routeing problem bike routeing problem energy constraint mixed integer linear programming split procedure evolutionary local search

来源：评论

学校读者我要写书评

暂无评论

Three-phase artificial intelligence-geographic information systems-based biomass network design approach: A case study in Denizli

引用

APPLIED ENERGY 2023年第1期343卷

作者： Senocak, Ahmet Alp Goren, Hacer Guner Pamukkale Univ Fac Engn Dept Ind Engn TR-20160 Denizli Turkiye

Renewable energy sources are of great importance in protecting the environment by reducing greenhouse gas emissions and global warming. Recently, more efficient conversion technologies and different types of resources have been introduced to generate energy in a sustainable way. Biomass-to-energy systems need to be designed efficiently due to high volumes of raw material flows and transportation costs. This study proposes a novel in-tegrated approach for the solution of biomass supply chain network design using artificial intelligence, geographic information systems, multi criteria decision making and mathematical modelling. First, the five-year forecasts of biomass raw materials including animal waste, agricultural residues, and municipal solid waste were made using support vector regression. Alternative biogas facility locations were determined spatially based on various criteria using geographic data and a decision-making technique. Then, considering annual net present value streams, the bioenergy system has been configured using a mixed integer linear programming model. The proposed methodology was applied on a real case in the city of Denizli, Turkey. The results showed that nine conversion facilities could be opened with 2000 kWh capacity each, and approximately 83.2% of net income could came from electricity sales, with the remainder from fertilizer sales. Furthermore, a sensitivity analysis was carried out to see the variations in model parameters such as biomass purchase cost, transportation cost, fertilizer sales prices and discount rate. It was found that the most dominant factor affecting net present value was fer-tilizer sales price, which was followed by discount rate.

关键词： Biomass supply chain Network design Support vector regression Geographic information systems mixed integer linear programming

来源：评论

学校读者我要写书评

暂无评论

An optimized trading strategy for an energy storage systems aggregator in an ancillary service market

引用

JOURNAL OF ENERGY STORAGE 2023年 72卷

作者： Lokhande, Smita Bichpuriya, Yogesh Kulkarni, Ankur A. Sarangan, Venkatesh Tata Consultancy Serv Ltd TCS Res Mumbai India Tata Consultancy Serv Ltd TCS Res Pune India Indian Inst Technol Syst & Control Engn Mumbai India Tata Consultancy Serv Ltd TCS Res Chennai India

With increasing proportion of intermittent renewable energy sources in the electric power system, the system operator (SO) is facing a challenge of maintaining the demand-supply balance which is more dynamic and uncertain than before. To manage the balance, the SO procures different services from balancing service providers like balancing mechanism units and distributed energy resources aggregators. In this work, we propose a model for an aggregator of energy storage systems (ESS). The distributed small size ESS can be grouped and utilized by an aggregator for trading of multiple services with different specifications and bidding rules in ancillary service markets. ESS, being flexible and having quick response time, can contribute in both directions for all the services and assist in maintaining the real time demand-supply balance of the system. The trading strategy proposed in this work captures the regulations and the bidding characteristics of all the services individually. This proposed solution is a mixed integer linear programming (MILP) problem which becomes computationally complex as the size of ESS dataset of the aggregator increases. Hence, we also propose an alternate approximation method which is scalable, comparatively easy to solve and takes less computation time while giving comparable optimal schedules. These proposed methods have also been compared with an intuitive baseline method. We have demonstrated the efficacy of the proposed methods with real world data of France balancing market.

关键词： Energy storage systems Aggregator Optimization Electricity market trading mixed integer linear programming

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：