A stochastic approach to hydroelectric power generation planning in an electricity market

Abstract

The paper discusses the planning of hydroelectric power generation. A stochastic optimization procedure is offered to solve the complex task of planning the operation of three hydroelectric power plants. The proposed stochastic optimization algorithm is based on time average revenue maximization, taking into account the random nature of the future energy prices and river water inflows. Random variables are predicted by using an algorithm based on artificial neural networks. For computing within the operational planning, the task is divided into three parts. First, middle-term planning is used to solve the water resources distribution task. The second and third parts are related to day-ahead operational and unit commitment planning; in those cases, nonlinear programming tools are used. A case study presents the results of electric power generation in the case of optimum water resource distribution in the storage basins of hydroelectric power plant's cascade. The paper proves the workability of the developed algorithm for maximizing the income value and is intended to enable and support improved planning and decision-making for electric power producers.

Estimating the Costs of Operating Reserve Provision by Poundage Hydroelectric Power Plants

“…In all the OMs, a unit commitment problem has to be solved. The task formulation for the three cascaded poundage HPPs in the RF mode has been presented in previous articles [8], [9]. In the present paper, the reformulation of the optimization task (Fig.…”

“…The operation of HPPs is subject to numerous limitations due to environmental and safety concerns. The constraints imposed by bank erosion, reservoir capacity, integrity of dam facilities and various other factors [8], [9] are summarized in Table I.…”

On the use of dynamic reliability for an accurate modelling of renewable power plants

“…Principle component analysis has been used in [20] to evaluate the wind power generation with respect to the geographic properties of the installation site. In [21] it is stated that the optimization procedures of hydroelectric power plants require the use of techniques able to account for the non-linear behavior of these systems, such as statistical inference methods [22], evolutionary computing algorithms [23] or machine learning techniques [24]. All these approaches are based on data-driven statistical learning methods and they do not model the underlying physical process, i.e.…”

Simplified Model for Evaluation of Hydropower Plant Conversion into Pumped Storage Hydropower Plant

“…The paper [6] presents a stochastic approach to solving a combined optimization problem for short-term scheduling and long-term investment planning of storage power plants for a 30-year long planning horizon. In [7], the stochastic optimization algorithm to solve the complex task of planning the operation of three hydroelectric power plants was proposed based on time average revenue maximisation and considering the random nature of the electricity prices and river water inflows.…”