A New Method for Probalistic Load Flow Analysis

  Copyright: © RWTH Aachen

This research aims at a novel method for probabilistic load flow (PLF) analysis, with incorporation of uncertainty from varying load demand and renewable energy sources. Sto- chastic response surface method (SRSM), on which our pro- babilistic load flow analysis is based, achieves computatio- nal efficiency while maintaining accurate results, compared with numerical and analytical approaches.

Modern power systems include various sources of uncertain- ty, including variable load demand and generation, network topology changes, and possible loss of transmission lines as well as generators. It is vital for operators, engineers and scientists to account for these uncertainty sources. Neglec- ting the uncertain characteristics in planning and operation may lead to deviations from expected behavior, challenges in decision making, and possibly unexpected outages in the power systems. The new approach to probabilistic load flow in this work focuses on uncertainty linked to load demand and renewable energy sources.

Probabilistic load flow is a technology based on the tradi- tional deterministic load flow (DLF), where though, uncer- tainties hidden in power injection at load bus and generati- on bus are represented in terms of probability distribution function (PDF) or cumulative distribution function (CDF) and the calculation yields PDF or CDF of the voltage magni- tude at the bus and power flow of branch. State of the art methods to get PLF results include analytical and numerical approaches. Monte Carlo (MC) methods are typical nume- rical approaches, able to achieve great accuracy at the ex- pense of a heavy computational burden. Analytical methods can accelerate the calculation procedure, but at the expense of precision, due to simplifications and approximations introduced in the process. The method proposed here aims at a trade-off between computational burden and accuracy. For this purpose, the stochastic response surface method is investigated, and applied to the load flow problem. SRSM is an extension of the deterministic response surface me- thod, which is used in other engineering applications. SRSM approximates uncertainties in target variables through a series expansion of the given input random variables. The coefficients in this series expansion are attained by ade- quately selecting the input surfaces and utilizing them to obtain the outputs of interest PLF combined with SRSM method is developed to achieve better understanding of the relationship between uncertainty sources and power flow outputs. This is done by leveraging on the stochastic response surface, which is a statistically equivalent reduced model of the original.

This new method has potentially broad application to the future smart grid. Owing to the improved efficiency, SRSM based PLF can provide solutions for both operation and ex- pansion planning in presence of uncertainties, fast enough for on-line use. One example of application is the rapid as- sessment of the ability of the network to accommodate the integration of renewable sources or to determine the confi- dence of the security of system operation due to e.g. wind generation and photovoltaic (PV) sources. Furthermore, this proposed PLF can be extended to voltage stability assess- ment, optimal power flow, as well as to electricity market considerations. Finally, this method is in principle applicable to both transmission and distribution networks.