Approximate model predictive building control via machine learning

Drgoňa, Ján; Picard, Damien; Kvasnica, Michal; Helsen, Lieve

doi:10.1016/j.apenergy.2018.02.156

Cited by 210 publications

(101 citation statements)

References 57 publications

Supporting

Mentioning

101

Contrasting

Order By: Relevance

“…The ML based building load forecast models have been extensively reviewed in [8][9][10][11][12][13]. Some of these models are developed for specific application areas such as building performance measurement and verification [14][15][16][17], building control [18][19][20] and demand-side management [21,22], whereas a significant number of studies are application agnostic. Literature demonstrates the capability of supervised ML algorithms such as artificial neural networks (ANN) [23], support vector machines (SVM) [24], decision trees [25,26], Gaussian processes [27][28][29] and nearest neighbours [30], among others, in developing reliable building load forecast models.…”

Section: Literature Reviewmentioning

confidence: 99%

A Machine Learning Pipeline for Demand Response Capacity Scheduling

Krishnadas¹,

Kiprakis

2020

Energies

View full text Add to dashboard Cite

Demand response (DR) is an integral component of smart grid operations that offers the necessary flexibility to support its decarbonisation. In incentive-based DR programs, deviations from the scheduled DR capacity affect the grid’s energy balance and result in revenue losses for the DR participants. This issue aggravates with increasing DR delivery from participants such as large consumer buildings who have limited standard methods to follow for DR capacity scheduling. Load curtailment based DR capacity availability from such consumers can be forecasted reliably with the help of supervised machine learning (ML) models. This study demonstrates the development of data-driven ML based total and flexible load forecast models for a retail building. The ML model development tasks such as data pre-processing, training-testing dataset preparation, cross-validation, algorithm selection, hyperparameter optimisation, feature ranking, model selection and model evaluation are guided by deployment-centric design criteria such as reliability, computational efficiency and scalability. Based on the selected performance metrics, the day-ahead and week-ahead ML based load forecast models developed for the retail building are shown to outperform the timeseries persistence models used for benchmarking. Furthermore, the deployment of these models for DR capacity scheduling is proposed as an ML pipeline that can be realised with the help of ML workflows, computational resources as well as systems for monitoring and visualisation. The ML pipeline ensures faster, cost-effective and large-scale deployment of forecast models that support reliable DR capacity scheduling without affecting the grid’s energy balance. Minimisation of revenue losses encourages increased DR participation from large consumer buildings, ensuring further flexibility in the smart grid.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

A Machine Learning Pipeline for Demand Response Capacity Scheduling

Krishnadas¹,

Kiprakis

2020

Energies

View full text Add to dashboard Cite

show abstract

“…Finally, profile uncertainties are considered in an MPC framework with chance constraints and machine learning algorithms are employed in [9] for developing approximate MPC laws for household temperature control.…”

Section: Preprint Submitted To Journal Of Electric Power Systems Resementioning

confidence: 99%

Meshed DC microgrid hierarchical control: A differential flatness approach

Zafeiratou¹,

Prodan²,

Lefèvre³

et al. 2020

Electric Power Systems Research

View full text Add to dashboard Cite

In this paper, a meshed DC microgrid control architecture whose goal is to manage load balancing and efficient power distribution is introduced. A novel combination of port-Hamiltonian (PH) modeling with differential flatness and B-splines parametrization is introduced and shown to improve the microgrid's performance. A three layer supervision structure is considered: i) B-spline parametrized flat output provide continuous profiles for load balancing and price reduction (high level); ii) the profiles are tracked through a MPC implementation with stability guarantees (medium level); iii) explicit switching laws applied to the DC/DC converters ensure appropriate power injection. Each level functions at a different timescale (from slow to fast), and the control laws are chosen appropriately. The effectiveness of the proposed approach is evaluated by simulations over a DC microgrid composed by a collection of solar panels (PV), an energy storage system (ES), a utility grid (UG) and a consumers' demand.

show abstract

“…In order to reduce the computational complexity which often affects MPC implementations, in [10] a machine learning approach based on multivariate regression and dimensionality reduction algorithms has been proposed, and a case study involving a 6-zone building has been carried out.…”

Section: Introductionmentioning

confidence: 99%

An integrated model predictive control approach for optimal HVAC and energy storage operation in large-scale buildings

et al. 2019

View full text Add to dashboard Cite

This paper deals with the problem of cost-optimal operation of smart buildings that integrate a centralized HVAC system, photovoltaic generation and both thermal and electrical storage devices.Building participation in a Demand-Response program is also considered. The proposed solution is based on a specialized Model Predictive Control strategy to optimally manage the HVAC system and the storage devices under thermal comfort and technological constraints. The related optimization problems turn out to be computationally appealing, even for large-scale problem instances. Performance evaluation, also in the presence of uncertainties and disturbances, is carried out using a realistic simulation framework.

show abstract

Approximate model predictive building control via machine learning

Cited by 210 publications

References 57 publications

A Machine Learning Pipeline for Demand Response Capacity Scheduling

A Machine Learning Pipeline for Demand Response Capacity Scheduling

Meshed DC microgrid hierarchical control: A differential flatness approach

An integrated model predictive control approach for optimal HVAC and energy storage operation in large-scale buildings

Contact Info

Product

Resources

About