The objective of this research is to develop effective tools to help power system operators and planners to address the challenges posed by the intermittency effects that are becoming more pronounced due to the increasing penetration of wind-based energy in the resource mix. The approach is to develop, through detailed statistical analysis, probabilistic models of the output of wind farms with different technology turbines located at various sites in a power system. These models will be used to evaluate system operational reliability and operational costs under different reserve policies for systems with various levels of wind penetration. The investigation will determine the robustness of the models as the penetration becomes deeper and will allow the construction of trade-off surfaces between different levels of reliability and operational costs.
Intellectual Merit
The proposed probabilistic models will be able to represent the impacts of wind speed variability and their imperfect forecasts on wind resource generation. The models will consider geographical correlation between wind sites, demand/wind speed temporal correlations, cannibalization effects among units in wind farms and smoothing effects resulting from the integration of several wind farms into a power system.
Broader Impacts
The project will allow the realization of one of the main aspects of the smart grid vision ?o the widespread deployment of renewable energy resources. The educational goals are to develop a vision for reshaping the power and energy systems curricula to provide appropriate training to the new generation of green energy engineers, who will provide the leadership and expertise to make sustainable energy a reality.
This research project was instrumental in the removal of some of the impediments to deeper wind resource penetration by developing tools for operators and planners to economically harness the wind energy while ensuring the reliability of the integrated system is maintained. Such tools are absolutely essential to allow the realization of one of the main aspects of the smart grid related to the widespread deployment of renewable resources. The tools provide operators and planners the ability to effectively manage the intermittency effects of wind-based generation and are sufficiently comprehensive to allow their adaptation to solar systems and other renewable sources, such as tidal power generation projects. Specifically, this project developed a set of effective tools to help power system operators and planners to address the challenges posed by the intermittency effects that are becoming more pronounced due to the increasing penetration of wind-based energy in the resource mix. These tools provide a comprehensive uncertainty analysis framework as they consider both set-theoretic and probabilistic modeling approaches to describe the uncertainty introduced by renewable-based generation. 1) Set-Theoretic Uncertainty Modeling: In this approach, we describe the amount of renewable-based generation by a set-theoretic model, i.e., uncertain variations in renewable-based generation are viewed as forecast error, which can be bounded (with some confidence level) around the nominal forecast. For this uncertainty model, and within the context of transmission systems, we have developed an analytically tractable method to assess whether in a particular timeframe, certain variables of interest, e.g., system frequency, remain within acceptable ranges for all possible realizations of renewable-based electric power generation. This method provides a computationally efficient approach (as an alternative to repeated simulation approaches) for quantifying the impact of renewable-based penetration on system dynamic performance. 2) Probabilistic Uncertainty Modeling: In this approach, the evolution of renewable-based power generation is characterized by a stochastic process that is assumed to be Markovian. Then, the evolution of the power system states, as described by the standard differential algebraic equation (DAE) model, is also described by a stochastic process. In general, it is not possible to analytically obtain the power system states’ joint probability density function (pdf); furthermore, attempting to obtain it numerically is computationally expensive. Thus, given the difficulty in obtaining this pdf, we have developed an analytical method to compute its moments. One key advantage of this method is that it only requires the computation of a finite number of moments, as needed for a particular application, which is generally much less information than one would need to completely determine the pdf. In addition to these analysis tools, we developed a general, stochastic simulation methodology that provides the capability to quantify the impacts of integrated renewable and storage resources on the power system economics, emissions and reliability variable effects over longer-periods with the various sources of uncertainty explicitly represented. We model the uncertainty in the demands, the available capacity of conventional generation resources, the charge/discharge operations of storage resources and the time-varying, intermittent renewable resources, with their temporal and spatial correlations, as discrete-time random processes. We deploy Monte Carlo simulation techniques to systematically sample these random processes to generate their realizations so as to construct the so-called sample paths. The approach emulates the side-by-side power system and transmission-constrained day-ahead market operations. We compute the metrics of interest used to assess the performance of the power system and associated markets. These metrics include the hourly expected locational marginal prices (LMPs), the revenues of the generators, total payments made by buyers in the DAMs, congestion rents, the system-wide CO2 emissions, as well as the reliability indices LOLP and EUE. We note that these metrics are computed by explicitly accounting for the deliverability of the electricity. The methodology is also able to capture the seasonal effects in demands and renewable outputs, the impacts of maintenance scheduling and the ramifications of new policy initiatives. The representative results we obtained from the extensive studies we performed using the stochastic simulation methodology effectively demonstrate the strong capabilities of the simulation approach. Some of the key findings of the numerical simulation work include the following: energy storage and wind resources tend to complement each other and the symbiotic effects reduce wholesale costs and improve system reliability; emission impacts with energy storage depend on the resource mix characteristics and the location of energy storage; and, storage seems to attenuate the "diminishing returns" associated with increased penetration of wind generation. The methodology provides heretofore unavailable implements to power system practitioners, investment analysts and policy analysts. As various power networks are integrating increasingly more wind resources with many projections reaching the 50 % penetration level, the operations and planning tools we constructed based on these models provide the capability to ensure the effective harnessing of wind generation, while ensuring the economic and reliable electricity supply to society.
