This study presents a comprehensive benchmarking framework for time-series clustering, addressing the lack of standardized guidelines for selecting appropriate approach for clustering tasks. The framework derives and evaluates 15 clustering pipelines comprising multiple well-known clustering techniques and diverse distance metrics, with the constraint of using the same number of clusters for all pipelines to ensure comparability and consistency in the evaluation process. The procedure for standardizing clustering labels, generating ensemble outcomes, and incorporating a stability score is introduced to provide a comprehensive and rigorous evaluation of clustering pipelines. The challenge posed by arbitrary label assignments across different pipelines is resolved through a standardization process that aligns the clusters for meaningful comparison. The ensemble approach mitigates inconsistencies in clustering results and addresses the limitations of traditional clustering validity metrics, leading to more stable and reliable groupings. Additionally, the inclusion of a novel stability score adds a critical layer of evaluation, enabling the identification of the most consistent and accurate clustering outcomes. The results highlight limitations of traditional quality metrics, while showcasing the strong performance for various pipelines with more than 90% of similarity with ensemble results. This would aid in informed selection of the best pipelines for specific applications. The framework is further discussed in the context of optimizing clustering for demand response strategies in smart grids, highlighting its real-world relevance. To support transparency and reproducibility, the study provides open-source code for validating and applying it to various time-series datasets, offering a robust tool for benchmarking clustering across domains.

The growing complexity of urban energy systems, climate uncertainties, and geopolitical disruptions highlight the need for energy flexibility and smart management. Recent developments in smart buildings enable real-time adaptability and collective energy behavior through the deployment of Reinforcement Learning (RL), which optimizes energy use, integrates distributed resources, and enhances demand response. However, challenges in communication, system diversity, and user intervention must be addressed for scalable and secure multi-agent RL-based management. This study evaluates the application of CIRLEM, a previously developed and introduced Energy Management system that integrates Collective Intelligence (CI) with an online, value-based, modelfree RL algorithm. The experiment is carried out in Building Energy Living Lab in France, as one of the pilots of COLLECTiEF, an European funded Horizon 2020 project, equipped with an advanced building management system for one year. The control algorithm interacts with the building management system every 15 min, optimizing setpoints based on real-time monitoring of energy use and indoor environmental conditions. The results indicate an 18% reduction in overall energy use compared to the reference baseline, with heating and cooling demands decreasing by 5% and 32%, respectively. Additionally, peak power demand is curtailed up to 15% for heating and 50% for cooling. The performance of the control algorithm is in an excellent level for more than 50% of the time in 1-month analyses through achieving load reduction and shifting. This experimental study demonstrates that CIRLEM effectively enhances energy flexibility while maintaining thermal comfort, demonstrating its potential for broader implementation, paving the way decentralized energy management solutions in smart buildings and urban energy networks.

Featured Application The proposed service-oriented interoperability framework enables the integration of heterogeneous building simulation services within distributed co-simulation workflows for early-stage building design. The approach is particularly suited to situations where independently developed tools must interact without direct model coupling or a unified simulation environment. The framework is also designed to facilitate future integration of optimization and decision-support processes through autonomous service interactions.Abstract Interoperability remains a central issue in multi-performance building simulation, where heterogeneous domain-specific tools must be combined despite differences in modeling formalisms, numerical solvers, and execution schemes. Existing approaches, including data exchange standards and component-based frameworks such as the Functional Mock-up Interface (FMI), address specific levels of interoperability but often require model-level access, component wrapping, Functional Mock-up Unit (FMU) packaging, or framework-specific integration. This paper examines service-level interoperability, where domain-specific simulation tools are exposed as autonomous web services coordinated through an external orchestration mechanism. A structured, JSON-based Pivot DataSet (PDS) organizes data exchange between services, while coupling strategies are implemented at the orchestration level to manage interactions without accessing internal model structures. The approach is evaluated using a classroom case study from the Agence Nationale de la Recherche (ANR) COSIMPHI research project, focusing on communication overhead, synchronization constraints, and coupling behavior in distributed co-simulation. Under the investigated weak-coupling conditions, the waveform relaxation method (WRM) reduces synchronization iterations by 144 & times; over one day and by approximately 3319 & times; over one month compared with minute-by-minute sequential chaining. These results, obtained under weak thermal-acoustic coupling conditions, highlight the relevance of service-level interoperability and orchestration-level coupling for distributed building-performance simulation workflows involving independently developed domain tools. Their generalization to stronger coupling regimes, however, remains a direction for future work.

Indoor Environmental Quality directly affects public health, productivity, and well-being, while also playing a vital role in developing climate-neutral, energy-efficient, and resilient buildings. This paper presents a comprehensive dataset of indoor environmental parameters that affect thermal comfort, indoor air quality, and visual comfort, which was created under the European Union’s Horizon 2020 Project Collective Intelligence for Energy Flexibility. The dataset comprises high-resolution measurements of carbon dioxide, pollutants, volatile organic compounds, air temperature, relative humidity, and illuminance on a horizontal plane, collected over a two-year period at 1-minute intervals. Data were gathered from 14 pilot buildings across four European climates: Cyprus, France, Italy, and Norway, covering diverse building types such as schools, medical centres, sports arenas, residential complexes, universities, and elder care facilities, representing about 40 % of common European building categories. Sensors were installed in specific thermal zones within each building to monitor environmental conditions. All data is organized by building and zone and supplemented with standardized Brick metadata to ensure interoperability. This comprehensive dataset, with its broad geographic coverage, variety of building types, long-term high-frequency measurements, and multimodal data, provides a valuable resource for comparative IEQ research, cross-domain modelling, and integrated assessments of comfort, ventilation, and daylighting across different climates and operational settings and is available upon request under a non-disclosure agreement provided by the consortium.

Coordinated control of residential air-conditioning systems is a promising demand response scheme to reduce peak loads and lower energy bills at a district level. Existing schemes have achieved only limited real-world success due to infrastructural requirements and low consumer adoption. In this paper, a scalable method for the coordinated control of air-conditioners in a group of hundreds of buildings is proposed in which consumers are allowed to override control signals based on their local thermal comfort. The control is based on a novel Split-Input Actor-Critic Reinforcement Learning architecture with a neural network that suggests temperature setpoints for each consumer. It is trained on a grey-box model of the system developed using historical smart meter data. The approach notably does not require behind-the-meter inputs during deployment. In a test system, the controller is able to reduce peak loads by up to 20% without increasing the total energy consumption compared to the baseline case with local control only. The robustness of the controller to different reinforcement learning architecture, inputs, model accuracy and stochasticity is studied. Additionally, the learnt policies are visualized, providing insights about the decision-making of the controller.

Batteries for electric vehicles are becoming major assets against global warming. Their production is the most impactful life cycle phase and their sizing depends on the vehicle usage. However, these usages are neither representative nor even taken into account in most studies. The main reason is the lack of a representative mobility usage construction method. This paper aims to propose a method to build prospective, representative mobility usages. To do so, the method combines data from prospective scenarios in order to parametrize different typical trips. Results present the evolution of the travelled distance and times in function of the scenario chosen or the type of trips.

This paper surveys the primary computational hurdles of Energy Systems optimization coming from different sources: model-induced complexity, optimization algorithm requirements, and uncertainties handling (both aleatoric and epistemic). Techniques to reduce complexity such as time-series and spatial aggregation, model order reduction, and specialized optimization strategies are reviewed for their effectiveness in balancing computational feasibility and model fidelity. Furthermore, Various uncertainty-management frameworks, including scenario-based approaches, robust optimization, and distributionally robust methods, are reviewed and their limitations in scaling and data requirements are discussed. The potential of hybrid modeling emerges as a key avenue: by fusing mechanistic and machine learning elements, hybrid techniques for modelling and optimization can harness the strengths of both worlds while mitigating their respective drawbacks. The paper highlights several directions for further research to develop advanced methods to tackle the complexity of MES.

Demand-side management is becoming essential for achieving global energy objectives and enhance the integration of intermittent renewable generation. However, in residential demand response (DR) programs, signal overrides by consumers, often driven by thermal comfort concerns, can undermine the targeted objectives. Modeling override behavior is challenging due to limited access to behind-the-meter data. This study proposes an offline-to-online reinforcement learning (RL) approach to improve pre-existing DR policies using consumer feedback, such as direct override signals, along with onlyweather and smart-meter data. Using the Calibrated Q-Learning (CalQL) algorithm, we demonstrate the potential to enhance DR outcomes in a simulated air-conditioner-based system involving 100 residential consumers. The results show significant improvements in peak load reduction (up to 60%) and override minimization (up to 99%) while maintaining scalability and practicality for real-world deployment. We study two distinct initial policy types, and further analyze the main parameters influencing such learning paradigms to discuss their impact on DR outcomes.

The local authorities and utilities need a deeper understanding of the drivers of electricity consumption in order to not only maintain energy balance, operational planning and grid resilience but also to design effective flexibility mechanisms and demand-side management actions. Most existing models rely on calendar-derived variables and meteorological data and they exhibit limitations in discerning granular local variations, hence, consequently fail to capture details in local shifts. Capturing local trends and accurately modelling meteorological impacts during extreme weather events remains challenging at a local scale. Additionally, diffused and unexpected factors (e.g. sports events, strike etc.) may influence electricity consumption in ways that are not anticipated. Building on earlier successes in refining well-known effects, i.e., Christmas or sports events, this research work leverages Natural Language Processing (NLP) and zero-shot labeling using a pretrained, open-source, natural language inference (NLI) model to classify textual data of social media. Our method generates new variables in the form of time series based on the content of these posts. These context-specific variables reflect local characteristics. In this paper, we focus on evaluating the impact of adding a single variable to the baseline prediction model of three French cities, without examining their combined effects altogether. The results are promising. For instance, the results show that the information about the public demonstration in a city improves the prediction error by up to 9%, therefore, improving the performance of model.

Understanding and estimating power demand and energy consumption for customer billing, grid planning, incorporating flexibility in the grid and enabling ecological transition need the most accurate forecasts at local level (e.g. at the level of town, city, regions etc.). Most models used by ENEDIS (the French DSO) takes weather information and calendar data (i.e. weekends, public holidays and season) as inputs. However, significant errors are observed on certain days, which can be attributed to irregular human activities. This research contributes to build a better understanding of these events. We have been able to develop a process to optimise the correlation between the energy consumption and events related variables created using NLP. These variables are created using unstructured textual data obtained from social media using NLP. Not only this method could be used to enhance forecasting accuracy by “neutralizing” the special days and stopping it to interfere with other variable estimation, but it could also help us monitor the overall impact of some events on electricity consumption. The results show that certain events have significant impact on local energy consumption and incorporating these event variables in baseline model helps in reducing the prediction error up to 7%.

The ordinary differential equations used to model a dynamic system can evolve during the simulation in circumstances where unpredictable events occur, more specifically, in regard to the domain of power electronics, for example, static converters will exhibit natural switching. Optimal sizing, on top of developing such a model, is a significant challenge for designers, particularly due to the complexity of incorporating efficiently both time-domain and frequency-domain constraints and objectives. This paper presents a methodology and tool to address this issue, leveraging a ‘white-box’ modeling approach, with automatic gradient computation. An efficient optimizer is coupled with a differential equation solver, capable of leveraging automatic differentiation and symbolic derivation, leading to both faster and more accurate outcomes than alternative methods. Furthermore, the developed solver incorporates original functionalities that are crucial for optimization, such as the ability to automatically detect the steady state and extract time-domain and frequency-domain features from the simulations to be optimized or constrained. The methodology is demonstrated through its application in regard to the optimal design of an aircraft electrical power channel.

As part of the energy transition and the rise in energy prices, the number of collective self-consumption operations in France is steadily increasing. However, energy flow monitoring currently relies on historical ”day+1” data provided by Linky meters, which does not offer real time feedback to help participants adapt their energy consumption behaviors. This article introduces a new open-source infrastructure for real-time monitoring based on Linky meter data, enabling participants to make informed decisions and take timely actions. It includes a description of the xKy device, applied to a collective self-consumption operation involving nine participants, supported by the Energy Transition Observatory (OTE). The project encompasses the implementation of gateways in participants' homes and the development and operation of real-time monitoring website, aimed at increasing participants' self-consumption rate.

Real-time electricity pricing has the potential to provide incentives for retail consumers to offer flexibility services by altering their consumption patterns. However, such incentive schemes have met with limited success in the real world due to problems such as low consumer interest and the creation of rebound peaks after periods of high pricing. In this paper, a model of an individual consumer’s response to real-time prices, which captures these effects, is presented. A contract between a retail service provider and a consumer is proposed, and a method for personalized real-time price generation based on smart meter data, using reinforcement learning, is implemented. Initial results suggest that the approach can be used to achieve grid-level objectives while rewarding consumer flexibility.

Purpose This study aims to optimize electrical systems represented by ordinary differential equations and events, using their frequency spectrum is an important purpose for designers, especially to calculate harmonics. Design/methodology/approach This paper presents a methodology to achieve this, by using a gradient-based optimization algorithm. The paper proposes to use a time simulation of the electrical system, and then to compute its frequency spectrum in the optimization loop. Findings The paper shows how to proceed efficiently to compute the frequency spectrum of an electrical system to include it in an optimization loop. Derivatives of the frequency spectrum such as the optimization inputs can also be calculated. This is possible even if the sized system behavior cannot be defined a priori , e.g. when there are static converters or electrical devices with natural switching. Originality/value Using an efficient sequential quadratic programming optimizer, automatic differentiation is used to compute the model gradients. Frequency spectrum derivatives with respect to the optimization inputs are calculated by an analytical formula. The methodology uses a “white-box” approach so that automatic differentiation and the differential equations simulator can be used, unlike most state-of-the-art simulators.

Non-Intrusive Load Monitoring (NILM) is referred to as the task of decomposing the aggregated power load of a residential or commercial building into appliance-level consumption without the installation of dedicated smart meters. NILM, mostly considered as a supervised learning problem, is crucial for energy monitoring and management as this approach can detect load malfunction and prevent wastage of energy consumption. For most of regression based NILM research, not much work has been done in analyzing data against temporal parameters such as day of the week, week of the month, week of the year, and various quarters in a year. Additionally, various NILM research works focus on finding the best regression model for predicting appliance consumption but discards the fact that appliance usage varies on factors such as seasons, different working hours, and weekends. This paper aims to evaluate some regression algorithms used towards NILM research based on 8 different training and testing Methods which according to our knowledge covered major factors that affect the appliance usage. The dataset used for the evaluation of the regression models, were collected from a research lab at Grenoble INP, in Grenoble, France. The evaluation results show that instead of one algorithm, individual regression algorithms generated favorable outcomes for various appliances based on different train-test Methods. Furthermore, a novel Bayesian optimized Ensemble regressor model for predicting individual appliance consumption from aggregated load data is also proposed. Instead of just using the aggregated power information, the proposed model also uses temporal information of the dataset to estimate accurate consumption output of individual appliances. Extensive simulation results corroborate the merits of the proposed approach, which outperforms the benchmarking Methods. To ensure reproducibility of the results by the research community and a potential future improvement of the framework by other researchers, the complete source code is provided in the following repository: https://github.com/Mohammad-Kaosain-Akbar/NILM-Ensemble-Bayesian-Optimization

Tertiary buildings could be an important lever to meet the goals necessitated by the energy transition. The availability of high-quality datasets from this sector will be a crucial enabler in meeting these goals by developing and testing new energy management approaches in the buildings. In this paper, we present the thermal energy datasets available and published online for the PREDIS-MHI zone of the GreEn-ER building, a tertiary building with more than a thousand sensors used for research, teaching, and administrative activities in Grenoble. PREDIS-MHI platform is a net-zero sub-section that is energetically isolated from the rest of the building. Its data has been used in a wide range of applications from indoor temperature forecasting, thermal simulation calibration, and even occupant comfort experiments

Electric vehicles are considered by many as an emission-free or low-emission solution to meet the challenge of sustainable transportation. However, the operational input, electrical energy, has an associated cost, greenhouse gasses, which results in indirect emissions. Given this knowledge, we pose the following question: “Are zero-emission transportation targets achievable given our current energy mix?” The objective of this article is to assess the impact of a grid’s energy mix on the indirect emissions of an electric vehicle. The study considers real-world data, vehicle usage data from an electric vehicle, and carbon intensity data for India, the USA, France, the Netherlands, Brazil, Germany, and Poland. Linear programming-based optimization is used to compute the best charging scenario for each of the given grids and, consequently, the indirect emissions are compared to those of a high-efficiency 1.5 L diesel internal combustion engine for the vehicle: a 2019 Renault Clio dCi 85. The results indicate that for grids with low renewable energy penetration, such as those of Poland and India (Maharashtra), an electric vehicle, even when optimally charged, can be classified as neither a low- nor zero-emission alternative to normal thermal vehicles. Also, for grids with elevated levels of variation in their carbon intensity, there is significant potential to reduce the carbon footprint related to charging an electric vehicle. This article provides a real-world perspective of how an electric vehicle performs in the face of different energy mixes and serves as a precursor to the development of robust indicators for determining the carbon reductions related to the e-mobility transition.

The proliferation of sensors in buildings has given us access to more data than before. To shepherd this rise in data, many open data lifecycles have been proposed over the past decade. However, many of the proposed lifecycles do not reflect the necessary complexity in the built environment. In this paper, we present a new open data lifecycle model: Open Energy Data Lifecycle (OPENDAL). OPENDAL builds on the key themes in more popular lifecycles and looks to extend them by better accounting for the information flows between cycles and the interactions between stakeholders around the data. These elements are included in the lifecycle in a bid to increase the reuse of published datasets. In addition, we apply the lifecycle model to the datasets from the GreEn-ER building, a mixed-use education building in France. Different use cases of these datasets are highlighted and discussed as a way to incentivise the use of data by other individuals.