This study introduces a novel method for optimising the size and control strategy of grid-connected, utility-scale photovoltaic (PV) systems with battery storage aimed at energy arbitrage and frequency containment reserve (FCR) services. By applying genetic algorithms (GA), the optimal configurations of PV generators, inverters/chargers, and batteries were determined, focusing on maximising the net present value (NPV). Both DC- and AC-coupled systems were explored. The performance of each configuration was simulated over a 25-year lifespan, considering varying pricing, solar resources, battery ageing, and PV degradation. Constraints included investment costs, capacity factors, and land use. A case study conducted in Wiesenthal, Germany, was followed by sensitivity analyses, revealing that a 75% reduction in battery costs is needed to make AC-coupled PV-plus-battery systems as profitable as PV-only systems. Further analysis shows that changes in electricity and FCR pricing as well as limits on FCR charging can significantly impact NPV. The study confirms that integrating arbitrage and FCR services can optimize system profitability.
Based on economic feasibility, renewable generators can use pumped hydro storage (PHS) to improve their profitability by performing energy arbitrage under real-time pricing (RTP) schemes. In this paper, we present a new method to optimise the size of and manage utility-scale wind–PV systems using PHS with energy arbitrage under RTP. PHS is used to supply load consumption and/or energy arbitrage. Further, both load-supply and power-generating systems are considered, and a genetic algorithm metaheuristic technique is used to perform the optimisation efficiently. Irradiance, wind speed, temperature, hourly electricity price, component characteristics, and financial data are used as data, and the system is simulated in 15 min time steps during the system lifetime for each combination of components and control variables. Uncertainty is considered for the meteorological data and electricity prices. The pump and turbine efficiencies and available head and penstock losses are considered as variables (not fixed values) to obtain accurate simulations. A sample application in Spain is demonstrated by performing a sensitivity analysis of different locations, electricity prices, and costs. PHS is not worth considering with the present cost of components. In load-supply systems in Zaragoza (Spain), we found that PHS would be worth considering if its cost was lower than 850 EUR/kW (considering all PHS components except reservoirs) +20 EUR/m3 for reservoirs (equivalent to 105 EUR/kWh with a 70 m head), whereas in Gran Canaria Island (with a considerably higher irradiation and wind speed), the required PHS cost is considerably lower (~350 EUR/kW + 10 EUR/m3). For power-generating systems, PHS required costs ranging from 400–700 EUR/kW + 15–20 EUR/m3 for obtaining the optimal PV–wind–PHS system with economic results similar to those of the optimal power-generating system without PHS. Thus, the renewable–PHS system with energy arbitrage under RTP could be profitable for many locations globally given the wide range of the PHS cost; however, each case is different and must be evaluated individually. The presented model can be used for optimising the renewable–PHS system in any location with any costs and RTP schemes.
Currently, the electrical supply in stand-alone systems is usually composed of renewable sources with fossil-fuel generators and battery storage. This study shows a novel model for the metaheuristic–stochastic optimization (minimization of the net present cost, and NPC) of sizing and energy management for stand-alone photovoltaic (PV)–wind–diesel systems with hybrid pumped hydro storage (PHS)–battery storage systems. The model is implemented in C++ programming language. To optimize operations—thus reducing PHS losses and increasing battery lifetimes—optimal energy management can optimize the power limits of using the PHS or battery to supply or store energy. The probabilistic approach considers the variability of wind speed, irradiation, temperature, load, and diesel fuel price inflation. The variable efficiencies of the components and losses and advanced models for battery degradation are considered. This methodology was applied to Graciosa Island (Portugal), showing that, compared with the current system, the optimal system (with a much higher renewable power and a hybrid PHS–battery storage) can reduce the NPC by half, reduce life cycle emissions to 14%, expand renewable penetration to 96%, and reduce the reserve capacity shortage to zero.
This study introduces a novel method for optimising the size and control strategy of grid-connected, utility-scale photovoltaic (PV) systems with battery storage aimed at energy arbitrage and frequency containment reserve (FCR) services. By applying genetic algorithms (GA), the optimal configurations of PV generators, inverters/chargers, and batteries are determined, focusing on maximising the net present value (NPV). Both DC- and AC-coupled systems are explored. The performance of each configuration is simulated over a 25-year lifespan, considering varying pricing, solar resources, battery ageing, and PV degradation. Constraints include investment costs, capacity factors, and land use. A case study conducted in Wiesenthal, Germany, is followed by sensitivity analyses, revealing that a 75% reduction in battery costs is needed to make AC-coupled PV-plus-battery systems as profitable as PV-only systems. Further analysis shows that changes in electricity and FCR pricing, as well as limits on FCR charging, can significantly impact NPV. The study confirms that integrating arbitrage and FCR services can optimize system profitability.
Based on economic feasibility, renewable generators can use pumped hydro storage (PHS) to improve their profitability by performing energy arbitrage under real time pricing (RTP) schemes. In this paper, we present a new method to optimise the size and manage utility-scale wind-PV systems using PHS with energy arbitrage under RTP. The storage in the PHS is used to supply load consumption and/or energy arbitrage. Further, both load supply and power-generating systems are considered, and a genetic algorithm metaheuristic technique is used to perform the optimisation efficiently. Irradiance, wind speed, temperature, hourly electricity price, component characteristics, and financial data are used as data, and the system is simulated in 15 min time steps during the system lifetime for each combination of components and control variables. Uncertainty is considered for the meteorological data and electricity prices. The pump and turbine efficiencies and available head and penstock losses are considered as variables (not fixed values) to obtain accurate simulations. A sample application in Spain is demonstrated by performing a sensitivity analysis of different locations, electricity prices, and costs. The PHS is not worth considering with the present cost of components.
Assessing the training process of artificial neural networks (ANNs) is vital for enhancing their performance and broadening their applicability. This paper employs the Monte Carlo simulation (MCS) technique, integrated with a stopping criterion, to construct the probability distribution of the learning error of an ANN designed for short-term forecasting. The training and validation processes were conducted multiple times, each time considering a unique random starting point, and the subsequent forecasting error was calculated one step ahead. From this, we ascertained the probability of having obtained all the local optima. Our extensive computational analysis involved training a shallow feedforward neural network (FFNN) using wind power and load demand data from the transmission systems of the Netherlands and Germany. Furthermore, the analysis was expanded to include wind speed prediction using a long short-term memory (LSTM) network at a site in Spain. The improvement gained from the FFNN, which has a high probability of being the global optimum, ranges from 0.7% to 8.6%, depending on the forecasting variable. This solution outperforms the persistent model by between 5.5% and 20.3%. For wind speed predictions using an LSTM, the improvement over an average-trained network stands at 9.5%, and is 6% superior to the persistent approach. These outcomes suggest that the advantages of exhaustive search vary based on the problem being analyzed and the type of network in use. The MCS method we implemented, which estimates the probability of identifying all local optima, can act as a foundational step for other techniques like Bayesian model selection, which assumes that the global optimum is encompassed within the available hypotheses.