Search

Distributed generation from wind and solar acts on regional electric demand as a reduced consumption, giving rise to a “load shadowing effect”. The net load becomes much more difficult to predict due to its dependence on the meteorological conditions. As a consequence, the growing penetration of variable generation increases the imbalance between demand and scheduled supply (net load forecast) and the reserve margins (net load uncertainty). The aim of this work is to quantify the benefit of the use of advanced probabilistic approaches rather than a traditional time-series method to assess the day-ahead reserves. For this purpose, several methods for load and net load uncertainty assessment have been developed and applied to a real case study considering also future solar penetration scenarios. The results show that, when forecasting only the load both traditional and probabilistic methods exhibit similar accuracy. Instead, in the case of net load prediction, i.e. when solar power is present, the probabilistic forecast can effectively limit the reserve margin needed to arrange the imbalance between residual demand and supply. The developed probabilistic approach provides a notable reduction of the Following Reserve which increases with the solar penetration: from 32.5% to 68.3% at 7% and 45% of penetration.

Solar photovoltaic energy is widespread worldwide and particularly in Europe, which became in 2016 the first region in the world to pass the 100 GW of installed capacity. As with all the renewable energy sources, for an effective management of solar power, it is essential to have reliable and accurate information about weather/climate conditions that affect the production of electricity. Operations in the solar energy industry are normally based on daily (or intra-daily) forecasts. Nevertheless, information about the incoming months can be relevant to support and inform operational and maintenance activities. This paper discusses a methodology to assess whether a seasonal climate forecast can provide a useful prediction for a specific sector, in this paper the European solar power industry. After evaluating the quality of the forecasts in providing probabilistic information for solar radiation, we describe how to assess their potential usefulness for a generic user by proposing an approach that takes into account not only their accuracy but also other potentially relevant factors. This approach is called index of opportunity and is then illustrated by presenting an example for the European solar power sector. The index of opportunity provides indications about where and when seasonal climate forecasts can benefit the decision-making in the photovoltaic sector. Even more importantly, it suggests an approach on how to evaluate their usefulness for the user's decision-making. This approach has the advantage of not limiting the definition of the usefulness only to the quality of the forecasts but rather considering, in an explicit way, all the factors that must be combined with the forecast's quality to define what is useful or not for the user.

The operation and economic profitability of modern energy systems is constrained by the availability of renewable energy and water resources. Lower water availability due to climate change, higher demand and increased water consumption for non-energy and energy needs may cause problems in Africa. In most African power systems, hydropower is a dominant renewable energy resource, and interconnection capacities are usually limited or unreliable. This paper describes a new modelling framework for analysing the water-energy nexus in the African Power Pools. This framework includes soft linking between two models: the LISFLOOD model is used to generate hydrological inputs and the Dispa-SET model is used for mid-term hydrothermal coordination and optimal unit commitment and power dispatch over the whole African continent. The results show a good agreement between the model outputs and the historical values, despite data-related limitations. Furthermore, the simulations provide hourly time series of electricity generation at the plant level in a robust way. It appears that some African power pools heavily rely on the availability of freshwater resources, while others are less dependent. In the long term, the dependence of the power system on water resources is likely to increase to meet the increasing electricity demand in Africa.

Hydro power can provide a source of dispatchable low-carbon electricity and a storage solution in a climate-dependent energy mix with high shares of wind and solar production. Therefore, understanding the effect climate has on hydro power generation is critical to ensure a stable energy supply, particularly at a continental scale. Here, we introduce a framework using climate data to model hydro power generation at the country level based on a machine learning method, the random forest model, to produce a publicly accessible hydro power dataset from 1979 to present for twelve European countries. In addition to producing a consistent European hydro power generation dataset covering the past 40 years, the specific novelty of this approach is to focus on the lagged effect of climate variability on hydro power. Specifically, multiple lagged values of temperature and precipitation are used. Overall, the model shows promising results, with the correlation values ranging between 0.85 and 0.98 for run-of-river and between 0.73 and 0.90 for reservoir-based generation. Compared to the more standard optimal lag approach the normalised mean absolute error reduces by an average of 10.23% and 5.99%, respectively. The model was also implemented over six Italian bidding zones to also test its skill at the sub-country scale. The model performance is only slightly degraded at the bidding zone level, but this also depends on the actual installed capacity, with higher capacities displaying higher performance. The framework and results presented could provide a useful reference for applications such as pan-European (continental) hydro power planning and for system adequacy and extreme events assessments.

The increased penetration of photovoltaic power introduces new challenges for the stability of the electrical grid, both at the local and national level. Many different effects are caused by high solar power injection into the electric grid. Among them, the increased risk of imbalance between the actual and scheduled power transmission is of particular relevance. The consequence is the need to exchange larger amounts of dispatchable power on the balancing energy market. The aim of this work is to analyze and quantify the effects of PV penetration in a target region and to evaluate the energy and economic benefits of using day-ahead PV forecast for power transmission scheduling. For this purpose, we developed several data-driven methods for transmission scheduling that include day-ahead PV power forecasts. We compared the resulting operational imbalances from these new models against two reference models currently used by the local grid operators. In the case of no PV generation in the target area, the more accurate reference model leads to an imbalance of 3.6% of the peak power transmission while more accurate data-driven method reduces the imbalance to 3.2%. When the distributed PV capacity is not zero, the imbalance of the reference model grows from 5.15% (at the current penetration of 7%) to 9.8% (at the maximum planned regional penetration of 45%). When we apply the new scheduling model, imbalances are reduced to respectively 3.5% and 5.8% at 7% and 45% of penetration. Since in Italy the costs of imbalances resulting from distributed PV are borne by ratepayers, these costs are estimated to be respectively 2.3% and 15% of the average electricity bill at 7% and 45% penetration if the reference scheduling is used. When applying the new model these costs are respectively reduced to 1.2% and 8.5%. © 2018 Elsevier Ltd

Accurate and robust short-term load forecasting plays a significant role in electric power operations. This paper proposes a variant of genetic programming, improved by incorporating semantic awareness in algorithm, to address a short term load forecasting problem. The objective is to automatically generate models that could effectively and reliably predict energy consumption. The presented results, obtained considering a particularly interesting case of the South Italy area, show that the proposed approach outperforms state of the art methods. Hence, the proposed approach reveals appropriate for the problem of forecasting electricity consumption. This study, besides providing an important contribution to the energy load forecasting, confirms the suitability of genetic programming improved with semantic methods in addressing complex real-life applications. © 2014 Elsevier B.V.

Air temperature is an effective predictor for electricity demand, especially during hot periods where the need of electric air conditioning can be high. This paper presents for the first time an assessment of the use of seasonal climate forecasts of temperature for medium-term electricity demand prediction. The retrospective seasonal climate forecasts provided by ECWMF (European Centre for Medium-Range Weather Forecasts) are used to forecast the June-July Italian electricity demand for the period 1990-2007.We find a relationship between summer (June-July) average temperature patterns over Europe and Italian electricity demand using both a linear and non-linear regression approach. With the aim to evaluate the potential usefulness of the information contained into the climate ensemble forecast, the analysis is extended considering a probabilistic approach.Results show that, especially in the Center-South of Italy, seasonal forecasts of temperature issued in May lead to a significant correlation coefficient of electricity demand greater than 0.6 for the summer period. The average correlation obtained from seasonal forecasts is 0.53 for the temperature predicted in May and 0.19 for the predictions issued in April for the linear model, while the non-linear approach leads to the coefficients of 0.62 and 0.36 respectively. For the probabilistic approach, seasonal forecasts exhibit a positive and significant skill-score in predicting the demand above/below the upper/lower tercile in many regions.This work is a significant progress in understanding the relationship between temperature and electricity demand. It is shown that much of the predictable electricity demand anomaly over Italy is connected with so-called heat-waves (i.e. long lasting positive temperature anomalies) over Europe. © 2014 Elsevier Ltd.

This study fills in knowledge gaps for the European air-conditioning (AC) market, which is fundamentally important to raising awareness about primary energy utilization. In contrast to space heating (SH) and domestic hot water (DHW) preparation, the European Union (EU) AC market is barely explored in scientific literature. While the focus of previous research has been on the residential sector, a shortfall of data for the services (wholesale and retail, offices, education, health, hotels and bars) exists. In this paper, data describing the actual space cooling (SC) market in Europe (quantity of SC units, equivalent full-load hours, installed capacities, seasonal energy efficiency values as well as cooled floor area per AC type and/or sector) is collected and explored using a bottom-up approach. Results indicate that SC is responsible for a significant portion of EU electricity consumption in households (nearly 5%) and even more in the service sector (~13%). Energy consumption for SC in the EU28 appears to be more than 140 TWh/y. The quantification of the European AC consumption shows a significant difference between the service and residential sectors: about 115 versus 25 TWh/y respectively. The SC market in Europe is characterized by a high potential for growth, especially in households., We would like to convey our deepest appreciation to the Horizon 2020 Hotmaps Project (Grant Agreement No. 723677), which provided the funding to carry out the present investigation. We are thankful to Pierre Iachetti (Eurac Research), who edited the entire text of the present investigation. Finally, our gratitude goes to Amy Segata (Eurac Research) for designing the graphs of this study.

Photovoltaic (PV) power forecasting has the potential to mitigate some of effects of resource variability caused by high solar power penetration into the electricity grid. Two main methods are currently used for PV power generation forecast: (i) a deterministic approach that uses physics-based models requiring detailed PV plant information and (ii) a data-driven approach based on statistical or stochastic machine learning techniques needing historical power measurements. The main goal of this work is to analyze the accuracy of these different approaches. Deterministic and stochastic models for dayahead PV generation forecast were developed, and a detailed error analysis was performed. Four years of site measurements were used to train and test the models. Numerical weather prediction (NWP) data generated by the weather research and forecasting (WRF) model were used as input. Additionally, a new parameter, the clear sky performance index, is defined. This index is equivalent to the clear sky index for PV power generation forecast, and it is here used in conjunction to the stochastic and persistence models. The stochastic model not only was able to correct NWP bias errors but it also provided a better irradiance transposition on the PV plane. The deterministic and stochastic models yield day-ahead forecast skills with respect to persistence of 35% and 39%, respectively. Copyright © 2017 by ASME.

The growing photovoltaic generation results in a stochastic variability of the electric demand that could compromise the stability of the grid, increase the amount of energy reserve and the energy imbalance cost. On regional scale, the estimation of the solar power generation from the real time environmental conditions and the solar power forecast is essential for Distribution System Operators, Transmission System Operator, energy traders, and Aggregators. In this context, a new upscaling method was developed and used for estimation and forecast of the photovoltaic distributed generation in a small area of Italy with high photovoltaic penetration. It was based on spatial clustering of the PV fleet and neural networks models that input satellite or numerical weather prediction data (centered on cluster centroids) to estimate or predict the regional solar generation. Two different approaches were investigated. The simplest and more accurate approach requires a low computational effort and very few input information should be provided by users. The power estimation model provided a RMSE of 3% of installed capacity. Intra-day forecast (from 1 to 4 h) obtained a RMSE of 5%–7% and a skill score with respect to the smart persistence from −8% to 33.6%. The one and two days ahead forecast achieved a RMSE of 7% and 7.5% and a skill score of 39.2% and 45.7%. The smoothing effect on cluster scale was also studied. It reduces the RMSE of power estimation of 33% and the RMSE of day-ahead forecast of 12% with respect to the mean single cluster value. Furthermore, a method to estimate the forecast error was also developed. It was based on an ensemble neural network model coupled with a probabilistic correction. It can provide a highly reliable computation of the prediction intervals. © 2017 Elsevier Ltd

Load forecasting is a critical task for all the operations of power systems. Especially during hot seasons, the influence of weather on energy demand may be strong, principally due to the use of air conditioning and refrigeration. This paper investigates the application of Genetic Programming on day-ahead load forecasting, comparing it with Neural Networks, Neural Networks Ensembles and Model Trees. All the experimentations have been performed on real data collected from the Italian electric grid during the summer period. Results show the suitability of Genetic Programming in providing good solutions to this problem. The advantage of using Genetic Programming, with respect to the other methods, is its ability to produce solutions that explain data in an intuitively meaningful way and that could be easily interpreted by a human being. This fact allows the practitioner to gain a better understanding of the problem under exam and to analyze the interactions between the features that characterize it.

Photovoltaic (PV) power production increased drastically in Europe throughout the last years. About the 6% of electricity in Italy comes from PV and for an efficient management of the power grid an accurate and reliable forecasting of production would be needed. Starting from a dataset of electricity production of 65 Italian solar plants for the years 2011-2012 we investigate the possibility to forecast daily production from one to ten days of lead time without using on site measurements. Our study is divided in two parts: an assessment of the predictability of meteorological variables using weather forecasts and an analysis on the application of data-driven modelling in predicting solar power production. We calibrate a SVM model using available observations and then we force the same model with the predicted variables from weather forecasts with a lead time from one to ten days. As expected, solar power production is strongly influenced by cloudiness and clear sky, in fact we observe that while during summer we obtain a general error under the 10% (slightly lower in south Italy), during winter the error is abundantly above the 20%., Submitted to Renewable Energy

This paper proposes a new approach for short-term load forecasting based on neural networks ensembling methods. A comparison between traditional statistical linear seasonal model and ANN-based models has been performed on the real-world building load data, considering the utilisation of external data such as the day of the week and building occupancy data. The selected models have been compared to the prediction of hourly demand for the electric power up to 24 hours for a testing week. Both neural networks ensembles achieved lower average and maximum errors than other models. Experiments showed how the introduction of external data had helped the forecasting.

Detection of turning points is a critical task for financial forecasting applications. This chapter proposes a comparison between two different classification approaches on such a problem. Nature-Inspired methodologies are attracting growing interest due to their ability to cope with complex tasks like classification, forecasting, and anomaly detection problems. A swarm intelligence algorithm, namely Particle Swarm Optimization (PSO), and an artificial immune system algorithm, namely Negative Selection (NS), have been applied to the task of detecting turning points, modeled as an Anomaly Detection (AD) problem. Particular attention has also been given to the choice of the features considered as inputs to the classifiers, due to the significant impact they may have on the overall accuracy of the approach. In this work, starting from a set of eight input features, feature selection has been carried out by means of a greedy hill climbing algorithm, in order to analyze the incidence of feature reduction on the global accuracy of the approach. The performances obtained from the two approaches have also been compared to other traditional machine learning techniques implemented by WEKA and both methods have been found to give interesting results with respect to traditional techniques.

This paper presents an application of two nature-inspired algorithms to the financial problem concerning the detection of turning points. Nature-Inspired methods are receiving a growing interest due to their ability to cope with complex tasks like classification, forecasting and anomaly detection problems. A swarm intelligence algorithm, Particle Swarm Optimization (PSO), and an artificial immune system one, the Negative Selection (NS), are applied to the problem of detection of turning points, modeled as an Anomaly Detection (AD) problem, and their performances are compared. Both methods are found to give interesting results with respect to an unpredictable behavior.

The detection of anomalies and faults is a fundamental task for different fields, especially in real cases like LAN networks and the Internet. We present an experimental study of anomaly detection on a simulated Internet backbone network based on neural networks, particle swarms, and artificial immune systems.

This paper presents a hybrid approach based on soft computing techniques in order to estimate ambient temperature for those places where such datum is not available. Indeed, we combine the Back-Propagation (BP) algorithm and the Simple Genetic Algorithm (GA) in order to effectively train neural networks in such a way that the BP algorithm initialises a few individuals of the GA's population. Experiments have been performed over all the available Italian places and results have shown a remarkable improvement in accuracy compared to the single and traditional methods.

This paper presents a new hybrid approach based both on traditional and soft computing techniques to provide ambient temperature for those places where such a datum is not available. Indeed, we combine neural networks with the nearest neighbouring algorithm; we use a fuzzy logic decision maker and later compare the results of each single technique to the hybrid one. Experiments have been performed on several Italian places; results have shown a remarkable improvement in accuracy compared to single methods.

Complex networks like the scale-free model proposed by Barabasi-Albert are observed in many biological systems and the application of this topology to artificial neural network leads to interesting considerations. In this paper, we present a preliminary study on how to evolve neural networks with complex topologies. This approach is utilized in the problem of modeling a chemical process with the presence of unknown inputs (disturbance). The evolutionary algorithm we use considers an initial population of individuals with differents scale-free networks in the genotype and at the end of the algorithm we observe and analyze the topology of networks with the best performances. Experimentation on modeling a complex chemical process shows that performances of networks with complex topology are similar to the feed-forward ones but the analysis of the topology of the most performing networks leads to the conclusion that the distribution of input node information affects the network performance (modeling capability).