Energy efficiency financing is considered among the top priorities in the energy sector among several stakeholders. In this context, accurately estimating the energy savings achieved by energy efficiency actions before being approved and implemented is of major importance to ensure the optimal allocation of the available financial resources. This study aims to provide a machine-learningbased methodological framework for a priori predicting the energy savings of energy efficiency renovation actions. The proposed solution consists of three tree-based algorithms that exploit bagging and boosting as well as an additional ensembling level that further mitigates prediction uncertainty. The proposed models are empirically evaluated using a database of various, diverse energy efficiency renovation investments. Results indicate that the ensemble model outperforms the three individual models in terms of forecasting accuracy. Also, the generated predictions are relatively accurate for all the examined project categories, a finding that supports the robustness of the proposed approach.
Artificial Intelligence (AI) technologies are moving from customized deployments in specific domains towards generic solutions horizontally permeating vertical domains and industries. For instance, decisions on when to perform maintenance of roads or bridges or how to optimize public lighting in view of costs and safety in smart cities are increasingly informed by AI models. While various commercial solutions offer user friendly and easy to use AI as a Service (AIaaS), functionality-wise enabling the democratization of such ecosystems, open-source equivalent ecosys- tems are lagging behind. In this chapter, we discuss AIaaS functionality and corresponding technology stack and analyze possible realizations using open source user friendly technologies that are suitable for on-premise set-ups of small and medium sized users allowing full control over the data and technological platform without any third-party dependence or vendor lock-in.
The increase in heterogeneous data in the building energy domain creates a difficult challenge for data integration. Schema matching, which maps the raw data from the building energy domain to a generic data model, is the necessary step in data integration and provides a unique representation. Only a small amount of labeled data for schema matching exists and it is time-consuming and labor-intensive to manually label data. This paper applies semantic-similarity methods to the automatic schema-mapping process by combining knowledge from natural language processing, which reduces the manual effort in heterogeneous data integration. The active-learning method is applied to solve the lack-of-labeled-data problem in schema matching. The results of the schema matching with building-energy-domain data show the pre-trained language model provides a massive improvement in the accuracy of schema matching and the active-learning method greatly reduces the amount of labeled data required.
Data that are generated from buildings, either directly with smart meters or indirectly from external data sources, are continuously growing. As they keep growing, their terminology does not follow the same structure, as they are being generated from different data sources and thus the ability to analyze and extract insights from them is restricted from the lack of harmonization. In applications, such as intelligent energy systems, building automation and controls, energy prediction and smart building services, data quality is critical to provide innovative services using Artificial Intelligence based (i.e., Machine Learning, Deep Learning) techniques. To exploit these data, gain insights and create solutions with added value, a framework for common use should be applied. In this context, the aim is to create a solution that takes into consideration the different type of ontologies of the building and its attributes, and to create a data model, using graph model representations. Under unique data models, the different features of the building and their dependencies, will enable optimal querying when data are accessed, and intelligent reasoning between typologies. To do that, several different datasets are exploited, which incorporate building attributes and can be applicable to multiple Big Data applications, and the findings are presented as graphical representations of the graph models.
Energy efficiency (EE) projects are often fragmented, of high transaction costs, and fall below the minimum value that many private financial institutions are willing to consider. The finance community is lacking a tested, evidence-based platform, providing decision makers with support regarding the impacts of various investment criteria, risk aware assessment, and performance applied on a pool of EE investments. The capability offered by emerging near big data analytics to integrate cross-domain financial and energy consumption is key to building the necessary market confidence in EE projects and making them an attractive investment asset class. The availability of comparable, anonymized historical data pooled from major market segments, structured along major project characteristics, can encourage greater EE investment flow. The aim is to present data-driven applications based on machine learning methods that can attract and mobilize private funding on such projects, providing investors/financiers (e.g., commercial/green investment banks, institutional/insurance funds, etc.) and project developers (public/local authorities, energy providers, ESCOs, construction companies, etc.) with data and tools to identify sustainable investment pathways and decrease the EE investment risk. Extensive data processing is applied to elaborate and categorize financing instruments and risk mitigation strategies, and to identify best practices on private financing as a basis for benchmarking.
One of the main challenges of today’s societies is the avoidance of the climate change since the climate crisis in now more evident than ever. Buildings have a large share of total energy consumption and, thus, it is obvious that actions should be taken to reduce their needs. Taking into consideration that nowadays data related to building’s metrics are available in significantly higher rate than in the past, due to the advance of the related technologies, it is necessary to find ways to exploit them in order to draw useful inferences regarding their consumptions and how they can be reduced. For that reason, in this paper we present a Visualisation Engine, which offers a variety of visualisations over stored data. With the usage of the proposed Visualisation Engine, we envision to be able to conduct sufficient research over the data, to generate insightful information regarding their behaviour, and to assist the development of useful solutions towards the direction of more energy efficient buildings.
During the Buildings’ lifecycle, massive amounts of data, that contain information related to their energy consumption, are generated. Towards the creation of smart building networks, this produced information must be intercepted and harmonized according to building ontologies and schemas. The pattern recognition from building metadata is based on inferencing and intelligent querying, that can be achieved with the utilization of graph and property databases that deploy and host building information. This paper presents a Reasoning Engine Architecture implemented in the context of the H2020 project called MATRYCS that persists building semantic information. It will be leveraged to support real life applications by improving the inference operations.
The reduction of the environmental impact in the building sector is necessary in achieving global sustainability. In this context, Building Automation and Control systems provide the opportunity for efficient monitoring and control facilities’ subsystems, such as the heating and cooling system, the ventilation system, the hot water system, the lighting appliances among others, with the goal of improving thermal comfort as well as energy efficiency. This paper presents a Building Automation and Control system aiming at facilitating data-driven monitoring of complex, multi-storey facilities, by disagreggating total consumption of the different floors and rooms of the building and offering advanced insights and benchmarking indicators. The service is showcased with a use case application on a real building, where the benefits of the service for the energy manager are highlighted.
Sensors, smart meters and IoT devices are key parts of the Building Information Systems. The amount of data generated from these sources is vast and the need for storage, fusion with secondary datasets (such as weather data) and aggregations has arisen in order to enhance building automation control activities. These data are stored on various data-sources, relational and non-relational, using different data formats. The combination of data coming from multiple data-sources constitutes a hard task, since each database uses different query language and structure. However, by combining all the available data-sources, it would be beneficial to reduce the volume of data during the training process. This paper presents an architecture that combines data from multiple data-sources (Databases, Object Storages, Building Information Systems) and create pipelines for aggregating the overall data.
In this paper, utilisation of Deep Learning for training energy consumption predictive models is examined, as vast amount of data from Internet of Things devices are available nowadays. Thus, the feedforward Artificial Neural Network (ANN) is proposed for predicting the adjusted baseline energy consumption, using the hour of the day, the day of the week and weather data as training features. The proposed models incorporate both linear and non-linear relationships, in contrast to linear regression methods.
To validate the proposed method, an experimental application is implemented, applying the developed models on an educational institution in Latvia. The building has been renovated regarding its heating supply and ventilation system, as well as its enclosing structures insulation. The predictions from the ANN models are compared with the ones from the traditional degree days method, indicating that ANN achieves higher accuracy in energy savings estimation for electricity and diesel fuel consumption.
Accurately forecasting solar plants production is critical for balancing supply and demand and for scheduling distribution networks operation in the context of inclusive smart cities and energy communities. However, the problem becomes more demanding, when there is insufficient amount of data to adequately train forecasting models, due to plants being recently installed or because of lack of smart-meters. Transfer learning (TL) offers the capability of transferring knowledge from the source domain to different target domains to resolve related problems. This study uses the stacked Long Short-Term Memory (LSTM) model with three TL strategies to provide accurate solar plant production forecasts. TL is exploited both for weight initialization of the LSTM model and for feature extraction, using different freezing approaches. The presented TL strategies are compared to the conventional non-TL model, as well as to the smart persistence model, at forecasting the hourly production of 6 solar plants.
Energy communities can support the energy transition, by engaging citizens through collective energy actions and generate positive economic, social and environmental outcomes. Renewable Energy Sources (RES) are gaining increasing share in the electricity mix as the economy decarbonises, with Photovoltaic (PV) plants to becoming more efficient and affordable. By incorporating Artificial Intelligence (AI) techniques, innovative applications can be developed to provide added value to energy communities. In this context, the scope of this paper is to compare Machine Learning (ML) and Deep Learning (DL) algorithms for the prediction of short-term production in a solar plant under an energy cooperative operation. Three different cases are considered, based on the data used as inputs for forecasting purposes. Lagged inputs are used to assess the historical data needed, and the algorithms’ accuracy is tested for the next hour’s PV production forecast. The comparative analysis between the proposed algorithms demonstrates the most accurate algorithm in each case, depending on the available data. For the highest performing algorithm, its performance accuracy in further forecasting horizons (3 hours, 6 hours and 24 hours) is also tested.
The building elements’ thermal transmittance index (known as U value) is the most essential parameter for the estimation of its thermal losses, and subsequently its energy performance. Laboratory measurements of the U value are practically of interest only for pre-fabricated or standardised building elements; hence, the commonly acceptable practice is to use theoretical estimations, taking into account the thermal resistance indexes of the building elements’ layers. In this context, this paper aims at the assessment of the uncertainty budget of one of the already used in-situ building elements’ thermal transmittance measurement, as long as this is applied using a thermal camera instead of contact thermometers. A specially designed and easily constructed auxiliary measurement set-up is presented and is used in comparison with simultaneous thermocouple measurements, to verify the adequacy of the auxiliary set up response time. This study concludes with the analysis of the uncertainty that affects the proposed procedure, along with an assessment of the impact in the final result. The uncertainty of these measurements is strongly related with the respective decisions on building upgrade and renovation. Therefore, managing it with a simplified procedure will be rather profitable both for those seeking to perform an in-situ U value measurement and for those who are engaged into a decision process upon renovating or not an existing building.
As energy efficiency is becoming a subject of utter importance in today’s societies, the European Union and a vast number of organizations have put a lot of focus on it. As a result, huge amounts of data are generated at an unprecedented rate. After thorough analysis and exploration, these data could provide a variety of solutions and optimizations regarding the energy efficiency subject. However, all the potential solutions that could derive from the aforementioned procedures still remain untapped due to the fact that these data are yet fragmented and highly sophisticated. In this paper, we propose an architecture for a Reasoning Engine, a mechanism that provides intelligent querying, insights and search capabilities, by leveraging technologies that will be described below. The proposed architecture has been developed in the context of the H2020 project called MATRYCS. In this paper, the reasons that resulted from the need of efficient ways of querying and analyzing the large amounts of data are firstly explained. Subsequently, several use cases, where related technologies were used to address real-world challenges, are presented. The main focus, however, is put in the detailed presentation of our Reasoning Engine’s implementation steps. Lastly, the outcome of our work is demonstrated, showcasing the derived results and the optimizations that have been implemented.
Photovoltaic (PV) modules and solar plants are one of the main drivers towards zero-carbon future. Energy communities that are engaging citizens through collective energy actions can reinforce positive social norms and support the energy transition. Furthermore, by incorporating Artificial Intelligence (AI) techniques, innovative applications can be developed with huge potential, such as supply and demand management, energy efficiency actions, grid operations and maintenance actions. In this context, the scope of this paper is to present an approach for forecasting an energy cooperative’s solar plant short term production by using its infrastructure and monitoring system. More specifically, four Machine Learning (ML) and Deep Learning (DL) algorithms are proposed and trained in an operational solar plant producing high accuracy short-term forecasts up to 6 hours. The results can be used for scheduling supply of the energy communities and set the base for more complex applications that require accurate short-term predictions, such as predictive maintenance.
Despite the large number of technology-intensive organisations, their corporate know-how and underlying workforce skill are not mature enough for a successful rollout of Artificial Intelligence (AI) services in the near-term. However, things have started to change, owing to the increased adoption of data democratisation processes, and the capability offered by emerging technologies for data sharing while respecting privacy, protection, and security, as well as appropriate learning-based modelling capabilities for non-expert end-users. This is particularly evident in the energy sector. In this context, the aim of this paper is to analyse AI and data democratisation, in order to explore the strengths and challenges in terms of data access problems and data sharing, algorithmic bias, AI transparency, privacy and other regulatory constraints for AI-based decisions, as well as novel applications in different domains, giving particular emphasis on the energy sector. A data democratisation framework for intelligent energy management is presented. In doing so, it highlights the need for the democratisation of data and analytics in the energy sector, toward making data available for the right people at the right time, allowing them to make the right decisions, and eventually facilitating the adoption of decentralised, decarbonised, and democratised energy business models.
The building sector is undergoing a deep transformation to contribute to meeting the climate neutrality goals set by policymakers worldwide. This process entails the transition towards smart energy-aware buildings that have lower consumptions and better efficiency performance. Digitalization is a key part of this process. A huge amount of data is currently generated by sensors, smart meters and a multitude of other devices and data sources, and this trend is expected to exponentially increase in the near future. Exploiting these data for different use cases spanning multiple application scenarios is of utmost importance to capture their full value and build smart and innovative building services. In this context, this paper presents a high-level architecture for big data management in the building domain which aims to foster data sharing, interoperability and the seamless integration of advanced services based on data-driven techniques. This work focuses on the functional description of the architecture, underlining the requirements and specifications to be addressed as well as the design principles to be followed. Moreover, a concrete example of the instantiation of such an architecture, based on open source software technologies, is presented and discussed.
Current climate change threats and increasing CO2 emissions, especially from the building stock, represent a context where action is required. It is necessary to provide efficient manners to manage energy demand in buildings and contribute to a decarbonised future. By combining new technologies, such as artificial intelligence, Internet of things, blockchain, and the exploitation of big data towards solving real life problems, the way could be paved towards smart and energyaware buildings. In this context, the aim of this paper is to present a critical review and an in-detail definition of the big data value chain for the built environment in Europe, covering multiple needs and perspectives: “policy”, “technology” and “business”, in order to explore the main challenges and opportunities in this area.