arXiv subjects

Rafael E. Carrillo

Publications and source records attributed to Rafael E. Carrillo.

At least 19 recordsLinked to original sources

arXiv2025

SolarCrossFormer: Improving day-ahead Solar Irradiance Forecasting by Integrating Satellite Imagery and Ground Sensors

Accurate day-ahead forecasts of solar irradiance are required for the large-scale integration of solar photovoltaic (PV) systems into the power grid. However, current forecasting solutions lack the temporal and spatial resolution required by system operators. In this paper, we introduce SolarCrossFormer, a novel deep learning model for day-ahead irradiance forecasting, that combines satellite images and time series from a ground-based network of meteorological stations. SolarCrossFormer uses novel graph neural networks to exploit the inter- and intra-modal correlations of the input data and improve the accuracy and resolution of the forecasts. It generates probabilistic forecasts for any location in Switzerland with a 15-minute resolution for horizons up to 24 hours ahead. One of the key advantages of SolarCrossFormer its robustness in real life operations. It can incorporate new time-series data without retraining the model and, additionally, it can produce forecasts for locations without input data by using only their coordinates. Experimental results over a dataset of one year and 127 locations across Switzerland show that SolarCrossFormer yield a normalized mean absolute error of 6.1 % over the forecasting horizon. The results are competitive with those achieved by a commercial numerical weather prediction service.

cs.LG↗

arXiv2024

Risk-aware Scheduling and Dispatch of Flexibility Events in Buildings

Residential and commercial buildings, equipped with systems such as heat pumps (HPs), hot water tanks, or stationary energy storage, have a large potential to offer their consumption flexibility as grid services. In this work, we leverage this flexibility to react to consumption requests related to maximizing self-consumption and reducing peak loads. We employ a data-driven virtual storage modeling approach for flexibility prediction in the form of flexibility envelopes for individual buildings. The risk-awareness of this prediction is inherited by the proposed scheduling algorithm. A Mixed-integer Linear Program (MILP) is formulated to schedule the activation of a pool of buildings in order to best respond to an external aggregated consumption request. This aggregated request is then dispatched to the active individual buildings, based on the previously determined schedule. The effectiveness of the approach is demonstrated by coordinating up to 500 simulated buildings using the Energym Python library and observing about 1.5 times peak power reduction in comparison with a baseline approach while maintaining comfort more robustly. We demonstrate the scalability of the approach by solving problems with 2000 buildings in about 21 seconds, with solving times being approximately linear in the number of considered assets.

eess.SY↗

arXiv2022

Uncertainty-aware Flexibility Envelope Prediction in Buildings with Controller-agnostic Battery Models

Buildings are a promising source of flexibility for the application of demand response. In this work, we introduce a novel battery model formulation to capture the state evolution of a single building. Being fully data-driven, the battery model identification requires one dataset from a period of nominal controller operation, and one from a period with flexibility requests, without making any assumptions on the underlying controller structure. We consider parameter uncertainty in the model formulation and show how to use risk measures to encode risk preferences of the user in robust uncertainty sets. Finally, we demonstrate the uncertainty-aware prediction of flexibility envelopes for a building simulation model from the Python library Energym.

eess.SY↗

arXiv2022

DRT-based modelling framework for Li-ion cells

The correct assessment of battery states is essential to maximize battery pack performances while ensuring reliable and safe operation. This work introduces EIS2MOD, a novel modelling framework for Li-ion cells based on Distribution of Relaxation Time (DRT). A physically based Electric Circuit Model (ECM) is developed starting from Electrochemical Impedance Spectroscopy (EIS) and Open Circuit Voltage (OCV) measurements. DRT is applied to deconvolve the electrochemical phenomena from the EIS. The presented methodology is based on: i) DRT calculation from EIS, ii) DRT analysis for ECM configuration and iii) Model parameters extraction and fitting. The proposed framework is applied to large format Li-ion pouch cells, which are tested over the whole State of Charge (SoC) range and a wide temperature range (-10°C to 35°C). Different current profiles have been tested to validate the model, showing its high accuracy in reproducing the battery cell behavior (e.g. RMSE on the battery terminals voltage lower than 1.50% for driving cycle simulations at variable temperature and SoC). An additional advantage of EIS2MOD is its light computational load thus offering an attractive framework for battery management system implementation.

eess.SY↗

arXiv2021

Smart sensors network for accurate indirect heat accounting in apartment buildings

A new method for accurate indirect heat accounting in apartment buildings has been recently developed by the Centre Suisse d'Electronique et de Microtechnique (CSEM). It is based on a data driven approach aimed to the smart networking of any type of indirect heat allocation devices, which can provide, for each heat delivery point of an apartment building, measurements or estimations of the temperature difference between the heat transfer fluid and the indoor environment. The analysis of the data gathered from the devices installed on the heating bodies, together with the measurements of the overall building heat consumption provided by direct heat metering, allows the evaluation of the characteristic thermal model parameters of heating bodies at actual installation and working conditions. Thus overcoming the negative impact on accuracy of conventional indirect heat accounting due to off-design operation, in which these measurement systems normally operate. The method has been tested on conventional heat cost allocators (HCA), and on innovative smart radiator thermostatic valves developed by CSEM. The evaluations were carried out at the centralized heating system mock-up of the Istituto Nazionale di Ricerca Metrologica (INRIM), and also in a real building in Neuchatel, Switzerland. The method has proven to be an effective tool to improve the accuracy of indirect heat metering systems; compared to conventional HCA systems, the error on the individual heating bill is reduced by 20% to 50%.

eess.SY↗

arXiv2021

Spatio-temporal graph neural networks for multi-site PV power forecasting

Accurate forecasting of solar power generation with fine temporal and spatial resolution is vital for the operation of the power grid. However, state-of-the-art approaches that combine machine learning with numerical weather predictions (NWP) have coarse resolution. In this paper, we take a graph signal processing perspective and model multi-site photovoltaic (PV) production time series as signals on a graph to capture their spatio-temporal dependencies and achieve higher spatial and temporal resolution forecasts. We present two novel graph neural network models for deterministic multi-site PV forecasting dubbed the graph-convolutional long short term memory (GCLSTM) and the graph-convolutional transformer (GCTrafo) models. These methods rely solely on production data and exploit the intuition that PV systems provide a dense network of virtual weather stations. The proposed methods were evaluated in two data sets for an entire year: 1) production data from 304 real PV systems, and 2) simulated production of 1000 PV systems, both distributed over Switzerland. The proposed models outperform state-of-the-art multi-site forecasting methods for prediction horizons of six hours ahead. Furthermore, the proposed models outperform state-of-the-art single-site methods with NWP as inputs on horizons up to four hours ahead.

cs.LG↗

arXiv2020

State space models for building control: how deep should you go?

Power consumption in buildings show non-linear behaviors that linear models cannot capture whereas recurrent neural networks (RNNs) can. This ability makes RNNs attractive alternatives for the model-predictive control (MPC) of buildings. However RNN models lack mathematical regularity which makes their use challenging in optimization problems. This work therefore systematically investigates whether using RNNs for building control provides net gains in an MPC framework. It compares the representation power and control performance of two architectures: a fully non-linear RNN architecture and a linear state-space model with non-linear regressor. The comparison covers five instances of each architecture over two months of simulated operation in identical conditions. The error on the one-hour forecast of temperature is 69% lower with the RNN model than with the linear one. In control the linear state-space model outperforms by 10% on the objective function, shows 2.8 times higher average temperature violations, and needs a third of the computation time the RNN model requires. This work therefore demonstrates that in their current form RNNs do improve accuracy but on balance well-designed linear state-space models with non-linear regressors are best in most cases of MPC.

eess.SY↗

arXiv2020

A hybrid learning method for system identification and optimal control

We present a three-step method to perform system identification and optimal control of non-linear systems. Our approach is mainly data driven and does not require active excitation of the system to perform system identification. In particular, it is designed for systems for which only historical data under closed-loop control are available and where historical control commands exhibit low variability. In a first step, simple simulation models of the system are built and run under various conditions. In a second step, a neural network architecture is extensively trained on the simulation outputs to learn the system physics, and retrained with historical data from the real system with stopping rules. These constraints avoid overfitting that arise by fitting closed-loop controlled systems. By doing so, we obtain one (or many) system model(s), represented by this architecture, and whose behaviour can be chosen to match more or less the real system. Finally, state-of-the-art reinforcement learning with a variant of domain randomization and distributed learning is used for optimal control of the system. We first illustrate the model identification strategy with a simple example, the pendulum with external torque. We then apply our method to model and optimize the control of a large building facility located in Switzerland. Simulation results demonstrate that this approach generates stable functional controllers which outperform on comfort and energy benchmark rule-based controllers.

eess.SY↗

arXiv2017

Robust sparse image reconstruction of radio interferometric observations with purify

Next-generation radio interferometers, such as the Square Kilometre Array (SKA), will revolutionise our understanding of the universe through their unprecedented sensitivity and resolution. However, to realise these goals significant challenges in image and data processing need to be overcome. The standard methods in radio interferometry for reconstructing images, such as CLEAN, have served the community well over the last few decades and have survived largely because they are pragmatic. However, they produce reconstructed inter\-ferometric images that are limited in quality and scalability for big data. In this work we apply and evaluate alternative interferometric reconstruction methods that make use of state-of-the-art sparse image reconstruction algorithms motivated by compressive sensing, which have been implemented in the PURIFY software package. In particular, we implement and apply the proximal alternating direction method of multipliers (P-ADMM) algorithm presented in a recent article. First, we assess the impact of the interpolation kernel used to perform gridding and degridding on sparse image reconstruction. We find that the Kaiser-Bessel interpolation kernel performs as well as prolate spheroidal wave functions, while providing a computational saving and an analytic form. Second, we apply PURIFY to real interferometric observations from the Very Large Array (VLA) and the Australia Telescope Compact Array (ATCA) and find images recovered by PURIFY are higher quality than those recovered by CLEAN. Third, we discuss how PURIFY reconstructions exhibit additional advantages over those recovered by CLEAN. The latest version of PURIFY, with developments presented in this work, is made publicly available.

astro-ph.IM↗

arXiv2017

A Fourier dimensionality reduction model for big data interferometric imaging

Data dimensionality reduction in radio interferometry can provide savings of computational resources for image reconstruction through reduced memory footprints and lighter computations per iteration, which is important for the scalability of imaging methods to the big data setting of the next-generation telescopes. This article sheds new light on dimensionality reduction from the perspective of compressed sensing theory and studies its interplay with imaging algorithms designed in the context of convex optimization. We propose a post-gridding linear data embedding to the space spanned by the left singular vectors of the measurement operator, providing a dimensionality reduction below image size. This embedding preserves the null space of the measurement operator and hence also its sampling properties as per compressed sensing theory. We show that this can be approximated by first computing the dirty image and then applying a weighted subsampled discrete Fourier transform to obtain the final reduced data vector. This Fourier dimensionality reduction model ensures a fast implementation of the full measurement operator, essential for any iterative image reconstruction method. The proposed reduction also preserves the i.i.d. Gaussian properties of the original measurement noise. For convex optimization-based imaging algorithms, this is key to justify the use of the standard L2-norm as the data fidelity term. Our simulations confirm that this dimensionality reduction approach can be leveraged by convex optimization algorithms with no loss in imaging quality relative to reconstructing the image from the complete visibility data set. Reconstruction results in simulation settings with no direction dependent effects or calibration errors show promising performance of the proposed dimensionality reduction. MATLAB code implementing the proposed reduction method is available on GitHub.

astro-ph.IM↗

arXiv2017

PURIFYing real radio interferometric observations

Next-generation radio interferometers, such as the Square Kilometre Array (SKA), will revolutionise our understanding of the universe through their unprecedented sensitivity and resolution. However, standard methods in radio interferometry produce reconstructed interferometric images that are limited in quality and they are not scalable for big data. In this work we apply and evaluate alternative interferometric reconstruction methods that make use of state-of-the-art sparse image reconstruction algorithms motivated by compressive sensing, which have been implemented in the PURIFY software package. In particular, we implement and apply the proximal alternating direction method of multipliers (P-ADMM) algorithm presented in a recent article. We apply PURIFY to real interferometric observations. For all observations PURIFY outperforms the standard CLEAN, where in some cases PURIFY provides an improvement in dynamic range by over an order of magnitude. The latest version of PURIFY, which includes the developments presented in this work, is made publicly available.

astro-ph.IM↗

arXiv2016

A randomised primal-dual algorithm for distributed radio-interferometric imaging

Next generation radio telescopes, like the Square Kilometre Array, will acquire an unprecedented amount of data for radio astronomy. The development of fast, parallelisable or distributed algorithms for handling such large-scale data sets is of prime importance. Motivated by this, we investigate herein a convex optimisation algorithmic structure, based on primal-dual forward-backward iterations, for solving the radio interferometric imaging problem. It can encompass any convex prior of interest. It allows for the distributed processing of the measured data and introduces further flexibility by employing a probabilistic approach for the selection of the data blocks used at a given iteration. We study the reconstruction performance with respect to the data distribution and we propose the use of nonuniform probabilities for the randomised updates. Our simulations show the feasibility of the randomisation given a limited computing infrastructure as well as important computational advantages when compared to state-of-the-art algorithmic structures.

astro-ph.IM↗

arXiv2016

Scalable splitting algorithms for big-data interferometric imaging in the SKA era

In the context of next generation radio telescopes, like the Square Kilometre Array, the efficient processing of large-scale datasets is extremely important. Convex optimisation tasks under the compressive sensing framework have recently emerged and provide both enhanced image reconstruction quality and scalability to increasingly larger data sets. We focus herein mainly on scalability and propose two new convex optimisation algorithmic structures able to solve the convex optimisation tasks arising in radio-interferometric imaging. They rely on proximal splitting and forward-backward iterations and can be seen, by analogy with the CLEAN major-minor cycle, as running sophisticated CLEAN-like iterations in parallel in multiple data, prior, and image spaces. Both methods support any convex regularisation function, in particular the well studied l1 priors promoting image sparsity in an adequate domain. Tailored for big-data, they employ parallel and distributed computations to achieve scalability, in terms of memory and computational requirements. One of them also exploits randomisation, over data blocks at each iteration, offering further flexibility. We present simulation results showing the feasibility of the proposed methods as well as their advantages compared to state-of-the-art algorithmic solvers. Our Matlab code is available online on GitHub.

astro-ph.IM↗

arXiv2015

Why CLEAN when you can PURIFY? A new approach for next-generation radio-interferometric imaging

In recent works, sparse models and convex optimization techniques have been applied to radio-interferometric (RI) imaging showing the potential to outperform state-of-the-art imaging algorithms in the field. In this talk, I will review our latest contributions in RI imaging, which leverage the versatility of convex optimization to both handle realistic continuous visibilities and offer a highly parallelizable structure paving the way to high-dimensional data scalability. Firstly, I will review our recently proposed average sparsity approach, SARA, which relies on the observation that natural images exhibit strong average sparsity over multiple coherent bases. Secondly, I will discuss efficient implementations of SARA, and sparse regularization problems in general, for large-scale imaging problems in a new toolbox dubbed

astro-ph.IM↗

arXiv2014

PURIFY: a new algorithmic framework for next-generation radio-interferometric imaging

In recent works, compressed sensing (CS) and convex optimization techniques have been applied to radio-interferometric imaging showing the potential to outperform state-of-the-art imaging algorithms in the field. We review our latest contributions, which leverage the versatility of convex optimization to both handle realistic continuous visibilities and offer a highly parallelizable structure paving the way to significant acceleration of the reconstruction and high-dimensional data scalability. The new algorithmic structure promoted in a new software PURIFY (beta version) relies on the simultaneous-direction method of multipliers (SDMM). The performance of various sparsity priors is evaluated through simulations in the continuous visibility setting, confirming the superiority of our recent average sparsity approach SARA.

astro-ph.IM↗

arXiv2014

Exploiting Prior Knowledge in Compressed Sensing Wireless ECG Systems

Recent results in telecardiology show that compressed sensing (CS) is a promising tool to lower energy consumption in wireless body area networks for electrocardiogram (ECG) monitoring. However, the performance of current CS-based algorithms, in terms of compression rate and reconstruction quality of the ECG, still falls short of the performance attained by state-of-the-art wavelet based algorithms. In this paper, we propose to exploit the structure of the wavelet representation of the ECG signal to boost the performance of CS-based methods for compression and reconstruction of ECG signals. More precisely, we incorporate prior information about the wavelet dependencies across scales into the reconstruction algorithms and exploit the high fraction of common support of the wavelet coefficients of consecutive ECG segments. Experimental results utilizing the MIT-BIH Arrhythmia Database show that significant performance gains, in terms of compression rate and reconstruction quality, can be obtained by the proposed algorithms compared to current CS-based methods.

cs.CE↗

arXiv2014

Sparsity averaging for radio-interferometric imaging

We propose a novel regularization method for compressive imaging in the context of the compressed sensing (CS) theory with coherent and redundant dictionaries. Natural images are often complicated and several types of structures can be present at once. It is well known that piecewise smooth images exhibit gradient sparsity, and that images with extended structures are better encapsulated in wavelet frames. Therefore, we here conjecture that promoting average sparsity or compressibility over multiple frames rather than single frames is an extremely powerful regularization prior.

astro-ph.IM↗

arXiv2014

PURIFY: a new approach to radio-interferometric imaging

In a recent article series, the authors have promoted convex optimization algorithms for radio-interferometric imaging in the framework of compressed sensing, which leverages sparsity regularization priors for the associated inverse problem and defines a minimization problem for image reconstruction. This approach was shown, in theory and through simulations in a simple discrete visibility setting, to have the potential to outperform significantly CLEAN and its evolutions. In this work, we leverage the versatility of convex optimization in solving minimization problems to both handle realistic continuous visibilities and offer a highly parallelizable structure paving the way to significant acceleration of the reconstruction and high-dimensional data scalability. The new algorithmic structure promoted relies on the simultaneous-direction method of multipliers (SDMM), and contrasts with the current major-minor cycle structure of CLEAN and its evolutions, which in particular cannot handle the state-of-the-art minimization problems under consideration where neither the regularization term nor the data term are differentiable functions. We release a beta version of an SDMM-based imaging software written in C and dubbed PURIFY (http://basp-group.github.io/purify/) that handles various sparsity priors, including our recent average sparsity approach SARA. We evaluate the performance of different priors through simulations in the continuous visibility setting, confirming the superiority of SARA.

astro-ph.IM↗