Ten questions concerning model predictive control for energy efficient buildings

Building and Environment, Год журнала: 2016, Номер 105, С. 403 - 412

Опубликована: Июнь 19, 2016

Язык: Английский

---

Energy flexibility of residential buildings: A systematic review of characterization and quantification methods and applications

Advances in Applied Energy, Год журнала: 2021, Номер 3, С. 100054 - 100054

Опубликована: Июль 9, 2021

With building electric demand becoming increasingly dynamic, and a growing percentage of intermittent renewable power generation from solar photovoltaics wind turbines, the grid is facing increasing challenge to manage real time balance between supply demand. advancements in smart sensing metering, appliances, vehicles, energy storage technologies, side management residential buildings can help improve stability by optimizing flexible loads. This paper reviews recent studies on management, with focus characterization quantification flexibility covering various types loads, metrics, methods, applications. The reviewed showed four levels applications: level (45%), district or community (29%), system (19%), sector (7%). Shifting loads dominant type 60% applications, followed shedding (16%), modulating (6%). Depending technology application scope, operations have wide range performance, peak reductions 1%~65%, savings up 60%, operational cost reduction 1%~48%, greenhouse gas emission to29%. More than half (51%) employed control strategies achieve flexibility; among those 72% used optimal controls, while 28% rule-based controls. About 58% mathematical formulation quantify flexibility. Most were based simulation, less 15% had measurements experiments field tests. review reveals research opportunities address significant gaps existing literature: (1) establishing common definition performance metrics for that are agnostic, (2) developing an ontology standardize representation resources interoperability, (3) integrating occupant impacts into optimization flexibility, (4) requirements credits codes standards. Findings inform future development which essential reliable resilient grid.

Язык: Английский

---

Energy saving potential of natural ventilation in China: The impact of ambient air pollution

Applied Energy, Год журнала: 2016, Номер 179, С. 660 - 668

Опубликована: Июль 18, 2016

Язык: Английский

---

Optimal control of HVAC and window systems for natural ventilation through reinforcement learning

Energy and Buildings, Год журнала: 2018, Номер 169, С. 195 - 205

Опубликована: Март 28, 2018

Язык: Английский

---

A zone-level, building energy optimisation combining an artificial neural network, a genetic algorithm, and model predictive control

Energy, Год журнала: 2018, Номер 151, С. 729 - 739

Опубликована: Март 20, 2018

Buildings account for a substantial proportion of global energy consumption and greenhouse gas emissions. Given the growth in smart devices sensors there is an opportunity to develop new generation smarter, more context aware, building controllers. Therefore, this work, surrogate, zone-level artificial neural networks that take weather, occupancy indoor temperature as inputs, have been created. These are used evaluation engine by genetic algorithm with aim minimising consumption. Bespoke 24-h, heating set point schedules generated each zone small office Cardiff, UK. The optimisation strategy can be deployed two modes, day ahead or model predictive control which re-optimises every hour. Over February test week, shown reduce around 25% compared baseline strategy. When time use tariff introduced, altered minimise cost rather than successfully shifts load cheaper price periods reduces 27%

Язык: Английский

---

Data-driven predictive control for unlocking building energy flexibility: A review

Renewable and Sustainable Energy Reviews, Год журнала: 2020, Номер 135, С. 110120 - 110120

Опубликована: Авг. 9, 2020

Managing supply and demand in the electricity grid is becoming more challenging due to increasing penetration of variable renewable energy sources. As significant end-use consumers, through better integration, buildings are expected play an expanding role future smart grid. Predictive control allows harness available flexibility from building passive thermal mass. However, heterogeneous nature stock, developing computationally tractable control-oriented models, which adequately represent complex nonlinear thermal-dynamics individual buildings, proving be a major hurdle. Data-driven predictive control, coupled with "Internet Things", holds promise for scalable transferrable approach,with data-driven models replacing traditional physics-based models. This review examines recent work utilising side management application special focus on nexus model development date, previous reviews have not addressed. Further topics examined include practical requirements harnessing mass issue feature selection. Current research gaps outlined pathways suggested identify most promising techniques integration buildings.

Язык: Английский

---

Approximate model predictive building control via machine learning

Applied Energy, Год журнала: 2018, Номер 218, С. 199 - 216

Опубликована: Март 9, 2018

Many studies have proven that the building sector can significantly benefit from replacing current practice rule-based controllers (RBC) by more advanced control strategies like model predictive (MPC). However, optimization-based algorithms, MPC, impose increasing hardware and software requirements, together with complicated error handling capabilities required commissioning staff. In recent years, several introduced promising remedy for these problems using machine learning algorithms. The idea is based on devising simplified laws learned MPC. main advantage of proposed methods stems their easy implementation even low-level hardware. most reported were dealing only a limited complexity parametric space, single variable, which inevitably limits applicability to complex problems. this paper, we introduce versatile framework synthesis simple, yet well-performing mimic behavior controllers, also large scale multiple-input-multiple-output (MIMO) are common in sector. approach employs multivariate regression dimensionality reduction Particularly, deep time delay neural networks (TDNN) trees (RT) used derive dependency multiple real-valued inputs parameters. problem, as well cost, further reduced selecting significant features set This straightforward manual selection, principal component analysis (PCA) dynamic model. demonstrated case study employing temperature six-zone building, described linear 286 states 42 disturbances, resulting an MPC problem than thousand results show retain performance while decreasing cost.

Язык: Английский

---

Review of urban building energy modeling (UBEM) approaches, methods and tools using qualitative and quantitative analysis

Energy and Buildings, Год журнала: 2021, Номер 246, С. 111073 - 111073

Опубликована: Май 25, 2021

The world has witnessed a significant population shift to urban areas over the past few decades. Urban account for about two-thirds of world's total primary energy consumption, which building sector constitutes proportion approximately 40%. Stakeholders such as planners and policy makers face substantial challenges when targeting sustainable climate goals related buildings' sector, i.e. reduce use associated emissions. modeling is one possible solution that leverages limited resources estimate support appropriate formation. Over years, there have been only review studies on approaches. These lack an in-depth discussion future research opportunities data-driven, reduced-order, simulation-based methods. This paper proposes Strengths, Weaknesses, Opportunities, Threats (SWOT) analysis approaches, methods tools used modeling. Furthermore, this generalized framework based existing literature different aim study assist policymakers choosing develop implement planning projects available resources.

Язык: Английский

---

Optimization of building envelope design for nZEBs in Mediterranean climate: Performance analysis of residential case study

Applied Energy, Год журнала: 2016, Номер 183, С. 938 - 957

Опубликована: Сен. 23, 2016

Язык: Английский

---

Transfer learning with deep neural networks for model predictive control of HVAC and natural ventilation in smart buildings

Journal of Cleaner Production, Год журнала: 2019, Номер 254, С. 119866 - 119866

Опубликована: Дек. 28, 2019

Язык: Английский

---