Adequacy of generation system with large-scale deployment of solar power and energy storage

摘要

Solar power and energy storage technologies are expected to have a large contribution in future electricity supply mix. This implies their significant impact on generation system adequacy (GSA) and creates the need for their inclusion into reliability studies. The most precise approach for performing such analyses, particularly in the presence of time-dependent power sources and complex operating policies, is the sequential Monte Carlo (SMC) technique.The general objective of this thesis was to develop new or adapt the existing models for creating chronological time series of power production/consumption from solar power and energy storage plants that would allow incorporating the given power sources into the SMC-based adequacy assessments. The final models had to be reasonably simplified to reduce the overall calculation time. The simplification also concerned the minimization of the required specific knowledge and input data.To achieve the goal of the thesis, the following four tasks were completed.First, a solar radiation model (SRM) was developed to create synthetic values of the hourly total, beam and diffuse radiation for single or multiple locations.SRM represents a set of single-site univariate stochastic algorithms and other auxiliary models which deploy the clearness index as a primary predictor. The SRM extension for multiple locations, i. e. incorporation of the spatial correlation of solar radiation (SCSR), was accomplished in two steps. In the first step a hypothesis was made that at long timescales simple characterizations of SCSR are possible. To prove the hypothesis, the author performed a regression analysis of the satellite-derived monthly and daily values of the clearness index for over 300,000 location pairs in 4 US regions; the given analysis is an original contribution of the thesis. In the second step, by applying the derived SCSR formulae and the existing methods of linear algebra, a general procedure was introduced for incorporating SCSR into stochastic algorithms. The performed validation studies showed that the proposed individual modifications and procedures are effective and SRM, overall, provides the synthetic solar radiation data of good quality.Second, simplified models of photovoltaic and concentrated solar (parabolic trough and central receiver based) power plants were developed for translating solar radiation to the corresponding power production. The proposed photovoltaic system model combines the Hay-Davies-Klucher-Reindl correlation for calculating solar radiation on a tilted surface and a reduced version of the 5-parameter model for converting the incident radiation to net power output. The new concentrated solar power models use a simplified simulation procedure consisting of three steps: (a) calculation of the solar field thermal output, (b) plant dispatch and conversion of useful thermal energy to gross power, and (c) estimation of parasitic losses and net production. The results from the validation studies, using the System Advisor Model as a reference, demonstrated the adequacy of the adopted approaches to model reductions.Third, a general dispatch model was developed for energy storage system (ESS) taking into account the inherent characteristics of the simulation-based adequacy evaluations. It represents another original contribution of the thesis.According to the proposed approach, energy storage is deployed to reduce the renewable energy curtailments and thermal generation capacity, which allows natural combination of its normal and emergency operations. The main novelties in this case are: (a) the use of detrended residual load during ESS control to be able to consider renewable power production and distinguish between different operating timescales, and (b) the correction of ESS dispatch based on the classical proportional integral controller technique to achieve a realistic representation of the storage level variations. The individual calculation steps of the proposed methodology were demonstrated through a simple adequacy analysis of a generation system comprising thermal power plants and hydroelectric ESS with annual and daily operating cycles.And finally, a case study of GSA in peninsular Spain was performed. The focus areas of the study were limited to: (a) adequacy of the existing Spanish generation system and the level of redundancy, (b) capacity credit of time-dependent power sources, and (c) sensitivity of the final results to certain modeling aspects. The Spanish power system was represented by system load, thermal generation, renewable energy sources (wind, solar and run-off-the-river hydro) and ESS (reservoir-based hydro). The excess of thermal generation and the capacity credits of the time-dependent power sources were estimated by using the effective load carrying capability as the metric. The study demonstrated how the new solar power and energy storage dispatch models could be applied to the Monte Carlo based GSA assessments, and provided additional knowledge on generation adequacy in Spain.