• Energy Research
• Open Source close
• Embargo close
• engineering and technology close
• IT close
• IN close

The focus of this study is to analyse and compare the predictive capabilities of univariate and multivariate methods of forecasting the global horizontal irradiance (GHI) for an hour ahead. The forecasting problem is addressed using supervised machine learning methods. In order to simplify the model, a feature selection algorithm is used to identify the highly correlated features. The forecasting is performed by utilizing popular machine learning algorithms viz., random forest (RF), K-nearest neighbors regression (KNN), support vector machine (SVM) and artificial neural networks (ANN). The paper evaluates and contrasts the effectiveness of these models for this application. Additionally, the study examines how the forecasting models' performance varies throughout the year and across seasons.

Abstract As main contributors to greenhouse gas emissions, power and transportation are crucial sectors for energy system decarbonization. Their interaction is expected to increase significantly: plug-in electric vehicles add a new electric load, increasing grid demand and potentially requiring substantial grid upgrade; hydrogen production for fuel cell electric vehicles or for clean fuels synthesis could exploit the projected massive power overgeneration by intermittent and seasonally-dependent renewable sources via Power-to-Hydrogen. This work investigates the infrastructural needs involved with a broad diffusion of clean mobility, adopting a sector integration perspective at the national scale. The analysis combines a multi-node energy system balance simulation and a techno-economic assessment of the infrastructure to deliver energy vectors for mobility. The article explores the long-term case of Italy, considering a massive increase of renewable power generation capacity and investigating different mobility scenarios, where low-emission vehicles account for 50% of the stock. First, the model solves the energy balances, integrating the consumption related to mobility energy vectors and taking into account power grid constraints. Then, an optimal infrastructure is identified, composed of both a hydrogen delivery network and a widespread installation of charging points. Results show that the infrastructural requirements bring about investment costs in the range of 43–63 G€. Lower specific costs are associated with the exclusive presence of FCEVs, whereas the full reliance on BEVs leads to the most significant costs. Scenarios that combine FCEVs and BEVs lie in between, suggesting that the overall power + mobility system benefits from the presence of both drivetrain options.

The optical and photovoltaic properties of Si NCs / SiC multilayers (MLs) are investigated using a membrane-based solar cell structure. By removing the Si substrate in the active cell area, the MLs are studied without any bulk Si substrate contribution. The occurrence is confirmed by scanning electron microscopy and light-beam induced current mapping . Optical characterization combined with simulations allows us to determine the absorption within the ML absorber layer, isolated from the other cell stack layers. The results indicate that the absorption at wavelengths longer than 800 nm is only due to the SiC matrix. The measured short-circuit current is significantly lower than that theoretically obtained from absorption within the ML absorber, which is ascribed to losses that limit carrier extraction. The origin of these losses is discussed in terms of the material regions where recombination takes place. Our results indicate that carrier extraction is most efficient from the Si NCs themselves, whereas recombination is strongest in SiC and residual a-Si domains . Together with the observed onset of the external quantum efficiency (EQE) at 700-800 nm, this fact is an evidence of quantum confinement in Si NCs embedded in SiC on device level.

Abstract Latest trends in the photovoltaic sector see the use of innovative photovoltaic technologies with extended spectral responsivity ranging from 300 to 1200 nm for non-concentrating terrestrial applications, and to 1800 nm for concentrating PV and space applications. As a consequence, an update of the IEC 60904-9 standard is ongoing with a definition of new spectral ranges for the assessment of the spectral match. This poses new challenges to laboratories and research centers on whether or not they still are able to accurately measure the spectral mismatch of their sun simulator in the newly-defined spectral regions. Prior to that, there is a need to understand if the commercially available spectroradiometers are ready to extend their measurement range as prescribed by the forthcoming new standard. This paper analyses two options for an extension of the spectrum characterisation of solar simulators to 300–1200 nm and compares them in terms of spectral match of global normal irradiance (GNI) spectra acquired under natural sunlight by eight spectroradiometers during the 6th European Spectroradiometer Intercomparison. The acquired spectra are also compared in terms of an index of consistency of the spread of the measured spectra with the estimated measurement uncertainty, hereafter named as performance statistics E n . Results show that all investigated laboratories assure the equivalence of the spectral match classification well below the 25% limit corresponding to class-A simulators. When considering the more stringent class-A+ corresponding to a 12.5% limit, one of the two considered options that rearranges the 300–1200 nm spectral range into 6 bands appears to still assure the equivalence of the class A+ limits among considered instruments. The E n performance index analysis highlights some inconsistencies with the estimated measurement uncertainty or instrument drifts from the expected performance, and the need of further improvements in calibration, set up and measurement procedures.

Thermo-economically optimal design = optimal molecule + optimal process + optimal equipment.

This paper presents a stochastic bi-level approach for tariff design in the electricity market where the leader is represented by a retailer and the follower by a residential prosumager, i.e. a consumer equipped with an energy system consisting of photovoltaic panels and a battery storage device. Both players solve an optimization problem subject to uncertainty in market prices, weather-related variables and electricity demand. To account for the retailer’s attitude towards risk, the upper level problem includes a safety measure to maximize. The model allows to determine a dynamic pricing scheme with time-variant rates delivering the average profit that can be gained in a given percentage of unfavorable realizations of the uncertain parameters and the optimal load pattern that minimizes the expected prosumager's electricity bill. The stochastic bi-level problem is reformulated as a single level model and different approaches to deal with the lower level problem and the non linearity in the objective function are analyzed and empirically investigated. A large number of numerical experiments have been carried out on real test cases. The results have shown the efficacy of the proposed approach as a tool to define pricing schemes that reflect the retailer's risk attitude underlining the importance of explicitly dealing with uncertainty.

Industrial heat and cooling applications are an essential fraction of the overall energy demand, mainly produced by fossil fuels. Solar thermal energy production can satisfy such a need by adopting the High Vacuum Flat Plate Collectors (HVFPCs) and increasing their efficiency. The absorptance and emittance of Selective Solar Absorbers (SSAs) determine the thermal efficiency of HVFPCs. Being the absorptance already maximized, the thermal emittance of the absorber should be minimized to increase further the operating temperature of the collector and its efficiency. This research aims to reduce the thermal emittance of commercially available Selective Solar Absorber by depositing a thin silver film on the aluminium substrate. So, in this work, the thermal stability of a silver coating has been investigated, and a diffusion barrier layer has been adopted to stabilize the coating performance up to 360 degrees C. The low-emissive layer of Ag and a diffusion barrier of CrOx guarantees a decrease of 11% in thermal emittance at 200 degrees C of commercially available SSA deposited on aluminium. Further emittance reduction can be obtained by depositing a thin Ag film on both sides of the aluminium substrate before the SSA deposition, proving to be a promising way to enhance the efficiency of HVFPCs.

Abstract The analysis of a large amount of data on the failures of absorption refrigeration machines of the ammonia-water type was carried out. The data derive from the records of a 10-year maintenance contract under which every machine was installed. The records were computerized for the period 1989–1992, and they refer to sales starting from 1980 (i.e. to 73 778 machines). The failures are classified, identifying the most serious (generator and solution pump). The rate of early failures is rather high, but the failures are generally minor and easily repaired. The production quality was not constant, with a very bad year in 1984. The failure rate is evaluated for the first 10 years of life of the equipment. In general terms, one machine in two does not require any intervention during that period.