Model-Based Predictive Control for building energy management. I: Energy modeling and optimal control

Salakij, Saran; Yu, Na; Paolucci, Samuel; Antsaklis, Panos J.

doi:10.1016/j.enbuild.2016.09.044

Cited by 75 publications

(33 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Case study Exp. Tool MB/DD Control [4] Commercial Building Yes E+ None Yes [5] Commercial building Yes E+ None Yes [6] Commercial building Yes E+ MB No [7] 2 office buildings and Yes None MB No 1 residential building [8] 2 commercial buildings n/a E+ MB No [9] Residential area No None MB Yes [10] 2 residential buildings Yes E+ MB Yes [11] 3 residential buildings No E+ MB Yes [12] 6 commercial buildings Yes E+ MB Yes [13] Residential building Yes None MB Yes [14] Commercial building No E+ MB-DD No [15] 2 commercial buildings Yes E+ MB-DD No [16] Office building Yes None DD No [17] Office building Yes E+ DD No [18] Residential house Yes TRANSYS DD No [19] Residential building Yes None DD No [20] Office building No E+ DD No [21] Commercial building Yes + RI None DD Yes [22] Living lab (1 room) Yes + RI None DD Yes [23] Commercial building Yes + RI None DD Yes [24] Residential house Yes None DD Yes [25] 9 commercial buildings No E+ DD Yes [26] Commercial building No E+ DD Yes scheme. In [22], an approach based on reinforcement learning, called Model-Assisted Batch Reinforcement Learning, is considered to provide data-driven control for the demand response problem in HVAC systems.…”

Section: Refmentioning

confidence: 99%

Data-driven model predictive control using random forests for building energy optimization and climate control

et al. 2018

View full text Add to dashboard Cite

Model Predictive Control (MPC) is a model-based technique widely and successfully used over the past years to improve control systems performance. A key factor prohibiting the widespread adoption of MPC for complex systems such as buildings is related to the difficulties (cost, time and effort) associated with the identification of a predictive model of a building. To overcome this problem, we introduce a novel idea for predictive control based on historical building data leveraging machine learning algorithms like regression trees and random forests. We call this approach Data-driven model Predictive Control (DPC), and we apply it to three different case studies to demonstrate its performance, scalability, and robustness. In the first case study we consider a benchmark MPC controller using a bilinear building model, then we apply DPC to a data-set simulated from such bilinear model and derive a controller based only on the data. Our results demonstrate that DPC can provide comparable performance with respect to MPC applied to a perfectly known mathematical model. In the second case study, we apply DPC to a 6 story 22 zone building model in EnergyPlus, for which model-based control is not economical and practical due to extreme complexity, and address a Demand Response problem. Our results demonstrate scalability and efficiency of DPC showing that DPC provides the desired power curtailment with an average error of 3%. In the third case study, we implement and test DPC on real data from an off-grid house located in L' Aquila, Italy. We compare the total amount of energy saved with respect to the classical bang-bang controller, showing that we can perform an energy saving up to 49.2%. Our results demonstrate the robustness of our method to uncertainties both in real data acquisition and weather forecast.

show abstract

Section: Refmentioning

confidence: 99%

Data-driven model predictive control using random forests for building energy optimization and climate control

et al. 2018

View full text Add to dashboard Cite

show abstract

“…An MPC tool was developed on the basis of the multivariable control model proposed by 64 and the set-point optimization technique proposed by. 60 The algorithm requires the similarity index (SI) shown in the following equation:…”

Section: • Peak Load Reductionmentioning

confidence: 99%

Artificial intelligence implementation framework development for building energy saving

et al. 2020

View full text Add to dashboard Cite

In this study, artificial intelligence (AI) control tools were developed to construct an AI implementation framework for energy saving for buildings. Although numerous AI studies related to energy conservation have been conducted, most of them have reported computing algorithms and control effects for single objects. This is the first study to use a framework to integrate five-category AI control tools to execute three-level building energy conservation; the three levels consist of equipment-level control, facility-level control, and whole building energy saving. Energy-saving effects were tested in a real building. The complex three-floor building primarily with a total area of 9072 m 2 serves as an office space and a semiconductor production line. Seventy percent energy consumption comes from air conditioning system and motor power. Twenty percent is lighting system and the other 10% is plug power and office automation equipment. Before implementation, the yearly energy cost reached US$1004339. In 2018, an AI implementation framework was introduced to systematically deploy AI at the site. A total of 47.5%, 37%, and 36.9% of energy was saved at equipment, facility, and whole building levels; up to US$385203 was saved. These energy savings proved the feasibility of the implementation framework. Furthermore, unmet demands of AI studies were met, and an approach to fill the research gap is discussed. K E Y W O R D S artificial intelligence (AI), artificial intelligence implementation framework (AIif), building energy saving, equipment-level control, facility-level control List of Symbols, Abbreviations, and Notation: ω i , weighting coefficients; GS T , the sum of the global similarities between the selected m cases; MV i , the mean difference of the variable i; P j , proportion of the prediction; V 1 , variation of control parameter y; y ic , the neural outputs of variable i for the control; y ip , the neural outputs of variable i for past cases; AI

show abstract

“…Typically, real-time building operation data (equipment and systems), predictive weather data, and occupant data are fed to energy models that simulate and evaluate various control strategies across a future time horizon, identifying the control strategy with the best predicted energy and comfort outcomes. This type of model predictive control (MPC) is an advanced method of process control that has been in use in chemical plants and oil refineries since the 1980s, only recently being appropriated for power system balancing models and building controls (Morari and Lee 1999; Salakij et al 2016). Model predictive controllers rely on dynamic models of the process, most often linear empirical models obtained by system identification.…”

Section: Real-time Optimization Control and Fault Detection And Diamentioning

confidence: 99%

Building simulation: Ten challenges

2018

View full text Add to dashboard Cite

Model-Based Predictive Control for building energy management. I: Energy modeling and optimal control

Cited by 75 publications

References 28 publications

Data-driven model predictive control using random forests for building energy optimization and climate control

Data-driven model predictive control using random forests for building energy optimization and climate control

Artificial intelligence implementation framework development for building energy saving

Building simulation: Ten challenges

Contact Info

Product

Resources

About